New submissions for Tue, 12 Mar 24

[DongZhouGu]

Keyword: human object interaction

Test-time Distribution Learning Adapter for Cross-modal Visual Reasoning

• Authors: Yi Zhang, Ce Zhang
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06059
• Pdf link: https://arxiv.org/pdf/2403.06059
• Abstract Vision-Language Pre-Trained (VLP) models, such as CLIP, have demonstrated remarkable effectiveness in learning generic visual representations. Several approaches aim to efficiently adapt VLP models to downstream tasks with limited supervision, aiming to leverage the acquired knowledge from VLP models. However, these methods suffer from either introducing biased representations or requiring high computational complexity, which hinders their effectiveness in fine-tuning the CLIP model. Moreover, when a model is trained on data specific to a particular domain, its ability to generalize to uncharted domains diminishes. In this work, we propose Test-Time Distribution LearNing Adapter (TT-DNA) which directly works during the testing period. Specifically, we estimate Gaussian distributions to model visual features of the few-shot support images to capture the knowledge from the support set. The cosine similarity between query image and the feature distribution of support images is used as the prediction of visual adapter. Subsequently, the visual adapter's prediction merges with the original CLIP prediction via a residual connection, resulting in the final prediction. Our extensive experimental results on visual reasoning for human object interaction demonstrate that our proposed TT-DNA outperforms existing state-of-the-art methods by large margins.

Keyword: visual relation detection

There is no result

Keyword: object detection

Not just Birds and Cars: Generic, Scalable and Explainable Models for Professional Visual Recognition

• Authors: Junde Wu, Jiayuan Zhu, Min Xu, Yueming Jin
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.05703
• Pdf link: https://arxiv.org/pdf/2403.05703
• Abstract Some visual recognition tasks are more challenging then the general ones as they require professional categories of images. The previous efforts, like fine-grained vision classification, primarily introduced models tailored to specific tasks, like identifying bird species or car brands with limited scalability and generalizability. This paper aims to design a scalable and explainable model to solve Professional Visual Recognition tasks from a generic standpoint. We introduce a biologically-inspired structure named Pro-NeXt and reveal that Pro-NeXt exhibits substantial generalizability across diverse professional fields such as fashion, medicine, and art-areas previously considered disparate. Our basic-sized Pro-NeXt-B surpasses all preceding task-specific models across 12 distinct datasets within 5 diverse domains. Furthermore, we find its good scaling property that scaling up Pro-NeXt in depth and width with increasing GFlops can consistently enhances its accuracy. Beyond scalability and adaptability, the intermediate features of Pro-NeXt achieve reliable object detection and segmentation performance without extra training, highlighting its solid explainability. We will release the code to foster further research in this area.

SAFDNet: A Simple and Effective Network for Fully Sparse 3D Object Detection

• Authors: Gang Zhang, Junnan Chen, Guohuan Gao, Jianmin Li, Si Liu, Xiaolin Hu
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.05817
• Pdf link: https://arxiv.org/pdf/2403.05817
• Abstract LiDAR-based 3D object detection plays an essential role in autonomous driving. Existing high-performing 3D object detectors usually build dense feature maps in the backbone network and prediction head. However, the computational costs introduced by the dense feature maps grow quadratically as the perception range increases, making these models hard to scale up to long-range detection. Some recent works have attempted to construct fully sparse detectors to solve this issue; nevertheless, the resulting models either rely on a complex multi-stage pipeline or exhibit inferior performance. In this work, we propose SAFDNet, a straightforward yet highly effective architecture, tailored for fully sparse 3D object detection. In SAFDNet, an adaptive feature diffusion strategy is designed to address the center feature missing problem. We conducted extensive experiments on Waymo Open, nuScenes, and Argoverse2 datasets. SAFDNet performed slightly better than the previous SOTA on the first two datasets but much better on the last dataset, which features long-range detection, verifying the efficacy of SAFDNet in scenarios where long-range detection is required. Notably, on Argoverse2, SAFDNet surpassed the previous best hybrid detector HEDNet by 2.6% mAP while being 2.1x faster, and yielded 2.1% mAP gains over the previous best sparse detector FSDv2 while being 1.3x faster. The code will be available at https://github.com/zhanggang001/HEDNet.

Frequency Attention for Knowledge Distillation

• Authors: Cuong Pham, Van-Anh Nguyen, Trung Le, Dinh Phung, Gustavo Carneiro, Thanh-Toan Do
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.05894
• Pdf link: https://arxiv.org/pdf/2403.05894
• Abstract Knowledge distillation is an attractive approach for learning compact deep neural networks, which learns a lightweight student model by distilling knowledge from a complex teacher model. Attention-based knowledge distillation is a specific form of intermediate feature-based knowledge distillation that uses attention mechanisms to encourage the student to better mimic the teacher. However, most of the previous attention-based distillation approaches perform attention in the spatial domain, which primarily affects local regions in the input image. This may not be sufficient when we need to capture the broader context or global information necessary for effective knowledge transfer. In frequency domain, since each frequency is determined from all pixels of the image in spatial domain, it can contain global information about the image. Inspired by the benefits of the frequency domain, we propose a novel module that functions as an attention mechanism in the frequency domain. The module consists of a learnable global filter that can adjust the frequencies of student's features under the guidance of the teacher's features, which encourages the student's features to have patterns similar to the teacher's features. We then propose an enhanced knowledge review-based distillation model by leveraging the proposed frequency attention module. The extensive experiments with various teacher and student architectures on image classification and object detection benchmark datasets show that the proposed approach outperforms other knowledge distillation methods.

Reframe Anything: LLM Agent for Open World Video Reframing

• Authors: Jiawang Cao, Yongliang Wu, Weiheng Chi, Wenbo Zhu, Ziyue Su, Jay Wu
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Human-Computer Interaction (cs.HC)
• Arxiv link: https://arxiv.org/abs/2403.06070
• Pdf link: https://arxiv.org/pdf/2403.06070
• Abstract The proliferation of mobile devices and social media has revolutionized content dissemination, with short-form video becoming increasingly prevalent. This shift has introduced the challenge of video reframing to fit various screen aspect ratios, a process that highlights the most compelling parts of a video. Traditionally, video reframing is a manual, time-consuming task requiring professional expertise, which incurs high production costs. A potential solution is to adopt some machine learning models, such as video salient object detection, to automate the process. However, these methods often lack generalizability due to their reliance on specific training data. The advent of powerful large language models (LLMs) open new avenues for AI capabilities. Building on this, we introduce Reframe Any Video Agent (RAVA), a LLM-based agent that leverages visual foundation models and human instructions to restructure visual content for video reframing. RAVA operates in three stages: perception, where it interprets user instructions and video content; planning, where it determines aspect ratios and reframing strategies; and execution, where it invokes the editing tools to produce the final video. Our experiments validate the effectiveness of RAVA in video salient object detection and real-world reframing tasks, demonstrating its potential as a tool for AI-powered video editing.

Enhancing 3D Object Detection with 2D Detection-Guided Query Anchors

• Authors: Haoxuanye Ji, Pengpeng Liang, Erkang Cheng
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06093
• Pdf link: https://arxiv.org/pdf/2403.06093
• Abstract Multi-camera-based 3D object detection has made notable progress in the past several years. However, we observe that there are cases (e.g. faraway regions) in which popular 2D object detectors are more reliable than state-of-the-art 3D detectors. In this paper, to improve the performance of query-based 3D object detectors, we present a novel query generating approach termed QAF2D, which infers 3D query anchors from 2D detection results. A 2D bounding box of an object in an image is lifted to a set of 3D anchors by associating each sampled point within the box with depth, yaw angle, and size candidates. Then, the validity of each 3D anchor is verified by comparing its projection in the image with its corresponding 2D box, and only valid anchors are kept and used to construct queries. The class information of the 2D bounding box associated with each query is also utilized to match the predicted boxes with ground truth for the set-based loss. The image feature extraction backbone is shared between the 3D detector and 2D detector by adding a small number of prompt parameters. We integrate QAF2D into three popular query-based 3D object detectors and carry out comprehensive evaluations on the nuScenes dataset. The largest improvement that QAF2D can bring about on the nuScenes validation subset is $2.3%$ NDS and $2.7%$ mAP. Code is available at https://github.com/nullmax-vision/QAF2D.

Cross-Cluster Shifting for Efficient and Effective 3D Object Detection in Autonomous Driving

• Authors: Zhili Chen, Kien T. Pham, Maosheng Ye, Zhiqiang Shen, Qifeng Chen
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
• Arxiv link: https://arxiv.org/abs/2403.06166
• Pdf link: https://arxiv.org/pdf/2403.06166
• Abstract We present a new 3D point-based detector model, named Shift-SSD, for precise 3D object detection in autonomous driving. Traditional point-based 3D object detectors often employ architectures that rely on a progressive downsampling of points. While this method effectively reduces computational demands and increases receptive fields, it will compromise the preservation of crucial non-local information for accurate 3D object detection, especially in the complex driving scenarios. To address this, we introduce an intriguing Cross-Cluster Shifting operation to unleash the representation capacity of the point-based detector by efficiently modeling longer-range inter-dependency while including only a negligible overhead. Concretely, the Cross-Cluster Shifting operation enhances the conventional design by shifting partial channels from neighboring clusters, which enables richer interaction with non-local regions and thus enlarges the receptive field of clusters. We conduct extensive experiments on the KITTI, Waymo, and nuScenes datasets, and the results demonstrate the state-of-the-art performance of Shift-SSD in both detection accuracy and runtime efficiency.

$V_kD:$ Improving Knowledge Distillation using Orthogonal Projections

• Authors: Roy Miles, Ismail Elezi, Jiankang Deng
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06213
• Pdf link: https://arxiv.org/pdf/2403.06213
• Abstract Knowledge distillation is an effective method for training small and efficient deep learning models. However, the efficacy of a single method can degenerate when transferring to other tasks, modalities, or even other architectures. To address this limitation, we propose a novel constrained feature distillation method. This method is derived from a small set of core principles, which results in two emerging components: an orthogonal projection and a task-specific normalisation. Equipped with both of these components, our transformer models can outperform all previous methods on ImageNet and reach up to a 4.4% relative improvement over the previous state-of-the-art methods. To further demonstrate the generality of our method, we apply it to object detection and image generation, whereby we obtain consistent and substantial performance improvements over state-of-the-art. Code and models are publicly available: https://github.com/roymiles/vkd

Poly Kernel Inception Network for Remote Sensing Detection

• Authors: Xinhao Cai, Qiuxia Lai, Yuwei Wang, Wenguan Wang, Zeren Sun, Yazhou Yao
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06258
• Pdf link: https://arxiv.org/pdf/2403.06258
• Abstract Object detection in remote sensing images (RSIs) often suffers from several increasing challenges, including the large variation in object scales and the diverse-ranging context. Prior methods tried to address these challenges by expanding the spatial receptive field of the backbone, either through large-kernel convolution or dilated convolution. However, the former typically introduces considerable background noise, while the latter risks generating overly sparse feature representations. In this paper, we introduce the Poly Kernel Inception Network (PKINet) to handle the above challenges. PKINet employs multi-scale convolution kernels without dilation to extract object features of varying scales and capture local context. In addition, a Context Anchor Attention (CAA) module is introduced in parallel to capture long-range contextual information. These two components work jointly to advance the performance of PKINet on four challenging remote sensing detection benchmarks, namely DOTA-v1.0, DOTA-v1.5, HRSC2016, and DIOR-R.

Transformer based Multitask Learning for Image Captioning and Object Detection

• Authors: Debolena Basak, P.K. Srijith, Maunendra Sankar Desarkar
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2403.06292
• Pdf link: https://arxiv.org/pdf/2403.06292
• Abstract In several real-world scenarios like autonomous navigation and mobility, to obtain a better visual understanding of the surroundings, image captioning and object detection play a crucial role. This work introduces a novel multitask learning framework that combines image captioning and object detection into a joint model. We propose TICOD, Transformer-based Image Captioning and Object detection model for jointly training both tasks by combining the losses obtained from image captioning and object detection networks. By leveraging joint training, the model benefits from the complementary information shared between the two tasks, leading to improved performance for image captioning. Our approach utilizes a transformer-based architecture that enables end-to-end network integration for image captioning and object detection and performs both tasks jointly. We evaluate the effectiveness of our approach through comprehensive experiments on the MS-COCO dataset. Our model outperforms the baselines from image captioning literature by achieving a 3.65% improvement in BERTScore.

3D Semantic Segmentation-Driven Representations for 3D Object Detection

• Authors: Hayeon O, Kunsoo Huh
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06501
• Pdf link: https://arxiv.org/pdf/2403.06501
• Abstract In autonomous driving, 3D detection provides more precise information to downstream tasks, including path planning and motion estimation, compared to 2D detection. Therefore, the need for 3D detection research has emerged. However, although single and multi-view images and depth maps obtained from the camera were used, detection accuracy was relatively low compared to other modality-based detectors due to the lack of geometric information. The proposed multi-modal 3D object detection combines semantic features obtained from images and geometric features obtained from point clouds, but there are difficulties in defining unified representation to fuse data existing in different domains and synchronization between them. In this paper, we propose SeSame : point-wise semantic feature as a new presentation to ensure sufficient semantic information of the existing LiDAR-only based 3D detection. Experiments show that our approach outperforms previous state-of-the-art at different levels of difficulty in car and performance improvement on the KITTI object detection benchmark. Our code is available at https://github.com/HAMA-DL-dev/SeSame

SARDet-100K: Towards Open-Source Benchmark and ToolKit for Large-Scale SAR Object Detection

• Authors: Yuxuan Li, Xiang Li, Weijie Li, Qibin Hou, Li Liu, Ming-Ming Cheng, Jian Yang
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Computational Engineering, Finance, and Science (cs.CE); Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2403.06534
• Pdf link: https://arxiv.org/pdf/2403.06534
• Abstract Synthetic Aperture Radar (SAR) object detection has gained significant attention recently due to its irreplaceable all-weather imaging capabilities. However, this research field suffers from both limited public datasets (mostly comprising <2K images with only mono-category objects) and inaccessible source code. To tackle these challenges, we establish a new benchmark dataset and an open-source method for large-scale SAR object detection. Our dataset, SARDet-100K, is a result of intense surveying, collecting, and standardizing 10 existing SAR detection datasets, providing a large-scale and diverse dataset for research purposes. To the best of our knowledge, SARDet-100K is the first COCO-level large-scale multi-class SAR object detection dataset ever created. With this high-quality dataset, we conducted comprehensive experiments and uncovered a crucial challenge in SAR object detection: the substantial disparities between the pretraining on RGB datasets and finetuning on SAR datasets in terms of both data domain and model structure. To bridge these gaps, we propose a novel Multi-Stage with Filter Augmentation (MSFA) pretraining framework that tackles the problems from the perspective of data input, domain transition, and model migration. The proposed MSFA method significantly enhances the performance of SAR object detection models while demonstrating exceptional generalizability and flexibility across diverse models. This work aims to pave the way for further advancements in SAR object detection. The dataset and code is available at https://github.com/zcablii/SARDet_100K.

Cross-domain and Cross-dimension Learning for Image-to-Graph Transformers

• Authors: Alexander H. Berger, Laurin Lux, Suprosanna Shit, Ivan Ezhov, Georgios Kaissis, Martin J. Menten, Daniel Rueckert, Johannes C. Paetzold
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06601
• Pdf link: https://arxiv.org/pdf/2403.06601
• Abstract Direct image-to-graph transformation is a challenging task that solves object detection and relationship prediction in a single model. Due to the complexity of this task, large training datasets are rare in many domains, which makes the training of large networks challenging. This data sparsity necessitates the establishment of pre-training strategies akin to the state-of-the-art in computer vision. In this work, we introduce a set of methods enabling cross-domain and cross-dimension transfer learning for image-to-graph transformers. We propose (1) a regularized edge sampling loss for sampling the optimal number of object relationships (edges) across domains, (2) a domain adaptation framework for image-to-graph transformers that aligns features from different domains, and (3) a simple projection function that allows us to pretrain 3D transformers on 2D input data. We demonstrate our method's utility in cross-domain and cross-dimension experiments, where we pretrain our models on 2D satellite images before applying them to vastly different target domains in 2D and 3D. Our method consistently outperforms a series of baselines on challenging benchmarks, such as retinal or whole-brain vessel graph extraction.

Evaluating the Energy Efficiency of Few-Shot Learning for Object Detection in Industrial Settings

• Authors: Georgios Tsoumplekas, Vladislav Li, Ilias Siniosoglou, Vasileios Argyriou, Sotirios K. Goudos, Ioannis D. Moscholios, Panagiotis Radoglou-Grammatikis, Panagiotis Sarigiannidis
• Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06631
• Pdf link: https://arxiv.org/pdf/2403.06631
• Abstract In the ever-evolving era of Artificial Intelligence (AI), model performance has constituted a key metric driving innovation, leading to an exponential growth in model size and complexity. However, sustainability and energy efficiency have been critical requirements during deployment in contemporary industrial settings, necessitating the use of data-efficient approaches such as few-shot learning. In this paper, to alleviate the burden of lengthy model training and minimize energy consumption, a finetuning approach to adapt standard object detection models to downstream tasks is examined. Subsequently, a thorough case study and evaluation of the energy demands of the developed models, applied in object detection benchmark datasets from volatile industrial environments is presented. Specifically, different finetuning strategies as well as utilization of ancillary evaluation data during training are examined, and the trade-off between performance and efficiency is highlighted in this low-data regime. Finally, this paper introduces a novel way to quantify this trade-off through a customized Efficiency Factor metric.

Genetic Learning for Designing Sim-to-Real Data Augmentations

• Authors: Bram Vanherle, Nick Michiels, Frank Van Reeth
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06786
• Pdf link: https://arxiv.org/pdf/2403.06786
• Abstract Data augmentations are useful in closing the sim-to-real domain gap when training on synthetic data. This is because they widen the training data distribution, thus encouraging the model to generalize better to other domains. Many image augmentation techniques exist, parametrized by different settings, such as strength and probability. This leads to a large space of different possible augmentation policies. Some policies work better than others for overcoming the sim-to-real gap for specific datasets, and it is unclear why. This paper presents two different interpretable metrics that can be combined to predict how well a certain augmentation policy will work for a specific sim-to-real setting, focusing on object detection. We validate our metrics by training many models with different augmentation policies and showing a strong correlation with performance on real data. Additionally, we introduce GeneticAugment, a genetic programming method that can leverage these metrics to automatically design an augmentation policy for a specific dataset without needing to train a model.

LeOCLR: Leveraging Original Images for Contrastive Learning of Visual Representations

• Authors: Mohammad Alkhalefi, Georgios Leontidis, Mingjun Zhong
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06813
• Pdf link: https://arxiv.org/pdf/2403.06813
• Abstract Contrastive instance discrimination outperforms supervised learning in downstream tasks like image classification and object detection. However, this approach heavily relies on data augmentation during representation learning, which may result in inferior results if not properly implemented. Random cropping followed by resizing is a common form of data augmentation used in contrastive learning, but it can lead to degraded representation learning if the two random crops contain distinct semantic content. To address this issue, this paper introduces LeOCLR (Leveraging Original Images for Contrastive Learning of Visual Representations), a framework that employs a new instance discrimination approach and an adapted loss function that ensures the shared region between positive pairs is semantically correct. The experimental results show that our approach consistently improves representation learning across different datasets compared to baseline models. For example, our approach outperforms MoCo-v2 by 5.1% on ImageNet-1K in linear evaluation and several other methods on transfer learning tasks.

Real-time Transformer-based Open-Vocabulary Detection with Efficient Fusion Head

• Authors: Tiancheng Zhao, Peng Liu, Xuan He, Lu Zhang, Kyusong Lee
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2403.06892
• Pdf link: https://arxiv.org/pdf/2403.06892
• Abstract End-to-end transformer-based detectors (DETRs) have shown exceptional performance in both closed-set and open-vocabulary object detection (OVD) tasks through the integration of language modalities. However, their demanding computational requirements have hindered their practical application in real-time object detection (OD) scenarios. In this paper, we scrutinize the limitations of two leading models in the OVDEval benchmark, OmDet and Grounding-DINO, and introduce OmDet-Turbo. This novel transformer-based real-time OVD model features an innovative Efficient Fusion Head (EFH) module designed to alleviate the bottlenecks observed in OmDet and Grounding-DINO. Notably, OmDet-Turbo-Base achieves a 100.2 frames per second (FPS) with TensorRT and language cache techniques applied. Notably, in zero-shot scenarios on COCO and LVIS datasets, OmDet-Turbo achieves performance levels nearly on par with current state-of-the-art supervised models. Furthermore, it establishes new state-of-the-art benchmarks on ODinW and OVDEval, boasting an AP of 30.1 and an NMS-AP of 26.86, respectively. The practicality of OmDet-Turbo in industrial applications is underscored by its exceptional performance on benchmark datasets and superior inference speed, positioning it as a compelling choice for real-time object detection tasks. Code: \url{https://github.com/om-ai-lab/OmDet}

Keyword: transformer

Understanding the Progression of Educational Topics via Semantic Matching

• Authors: Tamador Alkhidir (1), Edmond Awad (2), Aamena Alshamsi (3) ((1) Curriculum Department, Ministry of Education, United Arab Emirates, (2) Department of Economics and Institute for Data Science and AI, University of Exeter, United Kingdom,(3) Heuristic World, Dubai, United Arab Emirates)
• Subjects: Computers and Society (cs.CY); Computation and Language (cs.CL); Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2403.05553
• Pdf link: https://arxiv.org/pdf/2403.05553
• Abstract Education systems are dynamically changing to accommodate technological advances, industrial and societal needs, and to enhance students' learning journeys. Curriculum specialists and educators constantly revise taught subjects across educational grades to identify gaps, introduce new learning topics, and enhance the learning outcomes. This process is usually done within the same subjects (e.g. math) or across related subjects (e.g. math and physics) considering the same and different educational levels, leading to massive multi-layer comparisons. Having nuanced data about subjects, topics, and learning outcomes structured within a dataset, empowers us to leverage data science to better understand the progression of various learning topics. In this paper, Bidirectional Encoder Representations from Transformers (BERT) topic modeling was used to extract topics from the curriculum, which were then used to identify relationships between subjects, track their progression, and identify conceptual gaps. We found that grouping learning outcomes by common topics helped specialists reduce redundancy and introduce new concepts in the curriculum. We built a dashboard to avail the methodology to curriculum specials. Finally, we tested the validity of the approach with subject matter experts.

Detecting a Proxy for Potential Comorbid ADHD in People Reporting Anxiety Symptoms from Social Media Data

• Authors: Claire S. Lee, Noelle Lim, Michael Guerzhoy
• Subjects: Computers and Society (cs.CY); Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2403.05561
• Pdf link: https://arxiv.org/pdf/2403.05561
• Abstract We present a novel task that can elucidate the connection between anxiety and ADHD; use Transformers to make progress toward solving a task that is not solvable by keyword-based classifiers; and discuss a method for visualization of our classifier illuminating the connection between anxiety and ADHD presentations. Up to approximately 50% of adults with ADHD may also have an anxiety disorder and approximately 30% of adults with anxiety may also have ADHD. Patients presenting with anxiety may be treated for anxiety without ADHD ever being considered, possibly affecting treatment. We show how data that bears on ADHD that is comorbid with anxiety can be obtained from social media data, and show that Transformers can be used to detect a proxy for possible comorbid ADHD in people with anxiety symptoms. We collected data from anxiety and ADHD online forums (subreddits). We identified posters who first started posting in the Anxiety subreddit and later started posting in the ADHD subreddit as well. We use this subset of the posters as a proxy for people who presented with anxiety symptoms and then became aware that they might have ADHD. We fine-tune a Transformer architecture-based classifier to classify people who started posting in the Anxiety subreddit and then started posting in the ADHD subreddit vs. people who posted in the Anxiety subreddit without later posting in the ADHD subreddit. We show that a Transformer architecture is capable of achieving reasonable results (76% correct for RoBERTa vs. under 60% correct for the best keyword-based model, both with 50% base rate).

$\mathtt{tsGT}$: Stochastic Time Series Modeling With Transformer

• Authors: Łukasz Kuciński, Witold Drzewakowski, Mateusz Olko, Piotr Kozakowski, Łukasz Maziarka, Marta Emilia Nowakowska, Łukasz Kaiser, Piotr Miłoś
• Subjects: Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2403.05713
• Pdf link: https://arxiv.org/pdf/2403.05713
• Abstract Time series methods are of fundamental importance in virtually any field of science that deals with temporally structured data. Recently, there has been a surge of deterministic transformer models with time series-specific architectural biases. In this paper, we go in a different direction by introducing $\mathtt{tsGT}$, a stochastic time series model built on a general-purpose transformer architecture. We focus on using a well-known and theoretically justified rolling window backtesting and evaluation protocol. We show that $\mathtt{tsGT}$ outperforms the state-of-the-art models on MAD and RMSE, and surpasses its stochastic peers on QL and CRPS, on four commonly used datasets. We complement these results with a detailed analysis of $\mathtt{tsGT}$'s ability to model the data distribution and predict marginal quantile values.

Algorithmic progress in language models

• Authors: Anson Ho, Tamay Besiroglu, Ege Erdil, David Owen, Robi Rahman, Zifan Carl Guo, David Atkinson, Neil Thompson, Jaime Sevilla
• Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.05812
• Pdf link: https://arxiv.org/pdf/2403.05812
• Abstract We investigate the rate at which algorithms for pre-training language models have improved since the advent of deep learning. Using a dataset of over 200 language model evaluations on Wikitext and Penn Treebank spanning 2012-2023, we find that the compute required to reach a set performance threshold has halved approximately every 8 months, with a 95% confidence interval of around 5 to 14 months, substantially faster than hardware gains per Moore's Law. We estimate augmented scaling laws, which enable us to quantify algorithmic progress and determine the relative contributions of scaling models versus innovations in training algorithms. Despite the rapid pace of algorithmic progress and the development of new architectures such as the transformer, our analysis reveals that the increase in compute made an even larger contribution to overall performance improvements over this time period. Though limited by noisy benchmark data, our analysis quantifies the rapid progress in language modeling, shedding light on the relative contributions from compute and algorithms.

Long-term Frame-Event Visual Tracking: Benchmark Dataset and Baseline

• Authors: Xiao Wang, Ju Huang, Shiao Wang, Chuanming Tang, Bo Jiang, Yonghong Tian, Jin Tang, Bin Luo
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Neural and Evolutionary Computing (cs.NE)
• Arxiv link: https://arxiv.org/abs/2403.05839
• Pdf link: https://arxiv.org/pdf/2403.05839
• Abstract Current event-/frame-event based trackers undergo evaluation on short-term tracking datasets, however, the tracking of real-world scenarios involves long-term tracking, and the performance of existing tracking algorithms in these scenarios remains unclear. In this paper, we first propose a new long-term and large-scale frame-event single object tracking dataset, termed FELT. It contains 742 videos and 1,594,474 RGB frames and event stream pairs and has become the largest frame-event tracking dataset to date. We re-train and evaluate 15 baseline trackers on our dataset for future works to compare. More importantly, we find that the RGB frames and event streams are naturally incomplete due to the influence of challenging factors and spatially sparse event flow. In response to this, we propose a novel associative memory Transformer network as a unified backbone by introducing modern Hopfield layers into multi-head self-attention blocks to fuse both RGB and event data. Extensive experiments on both FELT and RGB-T tracking dataset LasHeR fully validated the effectiveness of our model. The dataset and source code can be found at \url{https://github.com/Event-AHU/FELT_SOT_Benchmark}.

Hufu: A Modality-Agnositc Watermarking System for Pre-Trained Transformers via Permutation Equivariance

• Authors: Hengyuan Xu, Liyao Xiang, Xingjun Ma, Borui Yang, Baochun Li
• Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.05842
• Pdf link: https://arxiv.org/pdf/2403.05842
• Abstract With the blossom of deep learning models and services, it has become an imperative concern to safeguard the valuable model parameters from being stolen. Watermarking is considered an important tool for ownership verification. However, current watermarking schemes are customized for different models and tasks, hard to be integrated as an integrated intellectual protection service. We propose Hufu, a modality-agnostic watermarking system for pre-trained Transformer-based models, relying on the permutation equivariance property of Transformers. Hufu embeds watermark by fine-tuning the pre-trained model on a set of data samples specifically permuted, and the embedded model essentially contains two sets of weights -- one for normal use and the other for watermark extraction which is triggered on permuted inputs. The permutation equivariance ensures minimal interference between these two sets of model weights and thus high fidelity on downstream tasks. Since our method only depends on the model itself, it is naturally modality-agnostic, task-independent, and trigger-sample-free. Extensive experiments on the state-of-the-art vision Transformers, BERT, and GPT2 have demonstrated Hufu's superiority in meeting watermarking requirements including effectiveness, efficiency, fidelity, and robustness, showing its great potential to be deployed as a uniform ownership verification service for various Transformers.

Interest-Aware Joint Caching, Computing, and Communication Optimization for Mobile VR Delivery in MEC Networks

• Authors: Baojie Fu, Tong Tang, Dapeng Wu, Ruyan Wang
• Subjects: Multimedia (cs.MM); Emerging Technologies (cs.ET)
• Arxiv link: https://arxiv.org/abs/2403.05851
• Pdf link: https://arxiv.org/pdf/2403.05851
• Abstract In the upcoming B5G/6G era, virtual reality (VR) over wireless has become a typical application, which is an inevitable trend in the development of video. However, in immersive and interactive VR experiences, VR services typically exhibit high delay, while simultaneously posing challenges for the energy consumption of local devices. To address these issues, this paper aims to improve the performance of the VR service in the edge-terminal cooperative system. Specifically, we formulate a problem of joint caching, computing, and communication VR service policy, by optimizing the weighted sum of overall VR delivery delay and energy consumption of local devices. For the purpose of designing the optimal VR service policy, the optimization problem is decoupled into three independent subproblems to be solved separately. To enhance the caching efficiency within the network, a bidirectional encoder representations from transformers (Bert)-based user interest analysis method is first proposed to characterize the content requesting behavior accurately. On the basis of this, a service cost minimum-maximization problem is formulated with consideration of performance fairness among users. Thereafter, the joint caching and computing scheme is derived for each user with given allocation of communication resources while a bisection-based communication scheme is acquired with the given information on joint caching and computing policy. With alternative optimization, an optimal policy for joint caching, computing and communication based on user interest can be finally obtained. Simulation results are presented to demonstrate the superiority of the proposed user interest-aware caching scheme and the effective of the joint caching, computing and communication optimization policy with consideration of user fairness.

General surgery vision transformer: A video pre-trained foundation model for general surgery

• Authors: Samuel Schmidgall, Ji Woong Kim, Jeffery Jopling, Axel Krieger
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Tissues and Organs (q-bio.TO)
• Arxiv link: https://arxiv.org/abs/2403.05949
• Pdf link: https://arxiv.org/pdf/2403.05949
• Abstract The absence of openly accessible data and specialized foundation models is a major barrier for computational research in surgery. Toward this, (i) we open-source the largest dataset of general surgery videos to-date, consisting of 680 hours of surgical videos, including data from robotic and laparoscopic techniques across 28 procedures; (ii) we propose a technique for video pre-training a general surgery vision transformer (GSViT) on surgical videos based on forward video prediction that can run in real-time for surgical applications, toward which we open-source the code and weights of GSViT; (iii) we also release code and weights for procedure-specific fine-tuned versions of GSViT across 10 procedures; (iv) we demonstrate the performance of GSViT on the Cholec80 phase annotation task, displaying improved performance over state-of-the-art single frame predictors.

CarbonNet: How Computer Vision Plays a Role in Climate Change? Application: Learning Geomechanics from Subsurface Geometry of CCS to Mitigate Global Warming

• Authors: Wei Chen, Yunan Li, Yuan Tian
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06025
• Pdf link: https://arxiv.org/pdf/2403.06025
• Abstract We introduce a new approach using computer vision to predict the land surface displacement from subsurface geometry images for Carbon Capture and Sequestration (CCS). CCS has been proved to be a key component for a carbon neutral society. However, scientists see there are challenges along the way including the high computational cost due to the large model scale and limitations to generalize a pre-trained model with complex physics. We tackle those challenges by training models directly from the subsurface geometry images. The goal is to understand the respons of land surface displacement due to carbon injection and utilize our trained models to inform decision making in CCS projects. We implement multiple models (CNN, ResNet, and ResNetUNet) for static mechanics problem, which is a image prediction problem. Next, we use the LSTM and transformer for transient mechanics scenario, which is a video prediction problem. It shows ResNetUNet outperforms the others thanks to its architecture in static mechanics problem, and LSTM shows comparable performance to transformer in transient problem. This report proceeds by outlining our dataset in detail followed by model descriptions in method section. Result and discussion state the key learning, observations, and conclusion with future work rounds out the paper.

Target-constrained Bidirectional Planning for Generation of Target-oriented Proactive Dialogue

• Authors: Jian Wang, Dongding Lin, Wenjie Li
• Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06063
• Pdf link: https://arxiv.org/pdf/2403.06063
• Abstract Target-oriented proactive dialogue systems aim to lead conversations from a dialogue context toward a pre-determined target, such as making recommendations on designated items or introducing new specific topics. To this end, it is critical for such dialogue systems to plan reasonable actions to drive the conversation proactively, and meanwhile, to plan appropriate topics to move the conversation forward to the target topic smoothly. In this work, we mainly focus on effective dialogue planning for target-oriented dialogue generation. Inspired by decision-making theories in cognitive science, we propose a novel target-constrained bidirectional planning (TRIP) approach, which plans an appropriate dialogue path by looking ahead and looking back. By formulating the planning as a generation task, our TRIP bidirectionally generates a dialogue path consisting of a sequence of <action, topic> pairs using two Transformer decoders. They are expected to supervise each other and converge on consistent actions and topics by minimizing the decision gap and contrastive generation of targets. Moreover, we propose a target-constrained decoding algorithm with a bidirectional agreement to better control the planning process. Subsequently, we adopt the planned dialogue paths to guide dialogue generation in a pipeline manner, where we explore two variants: prompt-based generation and plan-controlled generation. Extensive experiments are conducted on two challenging dialogue datasets, which are re-purposed for exploring target-oriented dialogue. Our automatic and human evaluations demonstrate that the proposed methods significantly outperform various baseline models.

FrameQuant: Flexible Low-Bit Quantization for Transformers

• Authors: Harshavardhan Adepu, Zhanpeng Zeng, Li Zhang, Vikas Singh
• Subjects: Machine Learning (cs.LG); Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2403.06082
• Pdf link: https://arxiv.org/pdf/2403.06082
• Abstract Transformers are the backbone of powerful foundation models for many Vision and Natural Language Processing tasks. But their compute and memory/storage footprint is large, and so, serving such models is expensive often requiring high-end hardware. To mitigate this difficulty, Post-Training Quantization seeks to modify a pre-trained model and quantize it to eight bits or lower, significantly boosting compute/memory/latency efficiency. Such models have been successfully quantized to four bits with some performance loss. In this work, we outline a simple scheme to quantize Transformer-based models to just two bits (plus some overhead) with only a small drop in accuracy. Key to our formulation is a concept borrowed from Harmonic analysis called Fusion Frames. Our main finding is that the quantization must take place not in the original weight space, but instead in the Fusion Frame representations. If quantization is interpreted as the addition of noise, our casting of the problem allows invoking an extensive body of known consistent recovery and noise robustness guarantees. Further, if desired, de-noising filters are known in closed form. We show empirically, via a variety of experiments, that (almost) two-bit quantization for Transformer models promises sizable efficiency gains.

Towards In-Vehicle Multi-Task Facial Attribute Recognition: Investigating Synthetic Data and Vision Foundation Models

• Authors: Esmaeil Seraj, Walter Talamonti
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
• Arxiv link: https://arxiv.org/abs/2403.06088
• Pdf link: https://arxiv.org/pdf/2403.06088
• Abstract In the burgeoning field of intelligent transportation systems, enhancing vehicle-driver interaction through facial attribute recognition, such as facial expression, eye gaze, age, etc., is of paramount importance for safety, personalization, and overall user experience. However, the scarcity of comprehensive large-scale, real-world datasets poses a significant challenge for training robust multi-task models. Existing literature often overlooks the potential of synthetic datasets and the comparative efficacy of state-of-the-art vision foundation models in such constrained settings. This paper addresses these gaps by investigating the utility of synthetic datasets for training complex multi-task models that recognize facial attributes of passengers of a vehicle, such as gaze plane, age, and facial expression. Utilizing transfer learning techniques with both pre-trained Vision Transformer (ViT) and Residual Network (ResNet) models, we explore various training and adaptation methods to optimize performance, particularly when data availability is limited. We provide extensive post-evaluation analysis, investigating the effects of synthetic data distributions on model performance in in-distribution data and out-of-distribution inference. Our study unveils counter-intuitive findings, notably the superior performance of ResNet over ViTs in our specific multi-task context, which is attributed to the mismatch in model complexity relative to task complexity. Our results highlight the challenges and opportunities for enhancing the use of synthetic data and vision foundation models in practical applications.

Diffusion Models Trained with Large Data Are Transferable Visual Models

• Authors: Guangkai Xu, Yongtao Ge, Mingyu Liu, Chengxiang Fan, Kangyang Xie, Zhiyue Zhao, Hao Chen, Chunhua Shen
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06090
• Pdf link: https://arxiv.org/pdf/2403.06090
• Abstract We show that, simply initializing image understanding models using a pre-trained UNet (or transformer) of diffusion models, it is possible to achieve remarkable transferable performance on fundamental vision perception tasks using a moderate amount of target data (even synthetic data only), including monocular depth, surface normal, image segmentation, matting, human pose estimation, among virtually many others. Previous works have adapted diffusion models for various perception tasks, often reformulating these tasks as generation processes to align with the diffusion process. In sharp contrast, we demonstrate that fine-tuning these models with minimal adjustments can be a more effective alternative, offering the advantages of being embarrassingly simple and significantly faster. As the backbone network of Stable Diffusion models is trained on giant datasets comprising billions of images, we observe very robust generalization capabilities of the diffusion backbone. Experimental results showcase the remarkable transferability of the backbone of diffusion models across diverse tasks and real-world datasets.

CLEAR: Cross-Transformers with Pre-trained Language Model is All you need for Person Attribute Recognition and Retrieval

• Authors: Doanh C. Bui, Thinh V. Le, Hung Ba Ngo, Tae Jong Choi
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06119
• Pdf link: https://arxiv.org/pdf/2403.06119
• Abstract Person attribute recognition and attribute-based retrieval are two core human-centric tasks. In the recognition task, the challenge is specifying attributes depending on a person's appearance, while the retrieval task involves searching for matching persons based on attribute queries. There is a significant relationship between recognition and retrieval tasks. In this study, we demonstrate that if there is a sufficiently robust network to solve person attribute recognition, it can be adapted to facilitate better performance for the retrieval task. Another issue that needs addressing in the retrieval task is the modality gap between attribute queries and persons' images. Therefore, in this paper, we present CLEAR, a unified network designed to address both tasks. We introduce a robust cross-transformers network to handle person attribute recognition. Additionally, leveraging a pre-trained language model, we construct pseudo-descriptions for attribute queries and introduce an effective training strategy to train only a few additional parameters for adapters, facilitating the handling of the retrieval task. Finally, the unified CLEAR model is evaluated on five benchmarks: PETA, PA100K, Market-1501, RAPv2, and UPAR-2024. Without bells and whistles, CLEAR achieves state-of-the-art performance or competitive results for both tasks, significantly outperforming other competitors in terms of person retrieval performance on the widely-used Market-1501 dataset.

Can Large Language Models Automatically Score Proficiency of Written Essays?

• Authors: Watheq Mansour, Salam Albatarni, Sohaila Eltanbouly, Tamer Elsayed
• Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06149
• Pdf link: https://arxiv.org/pdf/2403.06149
• Abstract Although several methods were proposed to address the problem of automated essay scoring (AES) in the last 50 years, there is still much to desire in terms of effectiveness. Large Language Models (LLMs) are transformer-based models that demonstrate extraordinary capabilities on various tasks. In this paper, we test the ability of LLMs, given their powerful linguistic knowledge, to analyze and effectively score written essays. We experimented with two popular LLMs, namely ChatGPT and Llama. We aim to check if these models can do this task and, if so, how their performance is positioned among the state-of-the-art (SOTA) models across two levels, holistically and per individual writing trait. We utilized prompt-engineering tactics in designing four different prompts to bring their maximum potential to this task. Our experiments conducted on the ASAP dataset revealed several interesting observations. First, choosing the right prompt depends highly on the model and nature of the task. Second, the two LLMs exhibited comparable average performance in AES, with a slight advantage for ChatGPT. Finally, despite the performance gap between the two LLMs and SOTA models in terms of predictions, they provide feedback to enhance the quality of the essays, which can potentially help both teachers and students.

On depth prediction for autonomous driving using self-supervised learning

• Authors: Houssem Boulahbal
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06194
• Pdf link: https://arxiv.org/pdf/2403.06194
• Abstract Perception of the environment is a critical component for enabling autonomous driving. It provides the vehicle with the ability to comprehend its surroundings and make informed decisions. Depth prediction plays a pivotal role in this process, as it helps the understanding of the geometry and motion of the environment. This thesis focuses on the challenge of depth prediction using monocular self-supervised learning techniques. The problem is approached from a broader perspective first, exploring conditional generative adversarial networks (cGANs) as a potential technique to achieve better generalization was performed. In doing so, a fundamental contribution to the conditional GANs, the acontrario cGAN was proposed. The second contribution entails a single image-to-depth self-supervised method, proposing a solution for the rigid-scene assumption using a novel transformer-based method that outputs a pose for each dynamic object. The third significant aspect involves the introduction of a video-to-depth map forecasting approach. This method serves as an extension of self-supervised techniques to predict future depths. This involves the creation of a novel transformer model capable of predicting the future depth of a given scene. Moreover, the various limitations of the aforementioned methods were addressed and a video-to-video depth maps model was proposed. This model leverages the spatio-temporal consistency of the input and output sequence to predict a more accurate depth sequence output. These methods have significant applications in autonomous driving (AD) and advanced driver assistance systems (ADAS).

$V_kD:$ Improving Knowledge Distillation using Orthogonal Projections

• Authors: Roy Miles, Ismail Elezi, Jiankang Deng
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06213
• Pdf link: https://arxiv.org/pdf/2403.06213
• Abstract Knowledge distillation is an effective method for training small and efficient deep learning models. However, the efficacy of a single method can degenerate when transferring to other tasks, modalities, or even other architectures. To address this limitation, we propose a novel constrained feature distillation method. This method is derived from a small set of core principles, which results in two emerging components: an orthogonal projection and a task-specific normalisation. Equipped with both of these components, our transformer models can outperform all previous methods on ImageNet and reach up to a 4.4% relative improvement over the previous state-of-the-art methods. To further demonstrate the generality of our method, we apply it to object detection and image generation, whereby we obtain consistent and substantial performance improvements over state-of-the-art. Code and models are publicly available: https://github.com/roymiles/vkd

MoST: Motion Style Transformer between Diverse Action Contents

• Authors: Boeun Kim, Jungho Kim, Hyung Jin Chang, Jin Young Choi
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06225
• Pdf link: https://arxiv.org/pdf/2403.06225
• Abstract While existing motion style transfer methods are effective between two motions with identical content, their performance significantly diminishes when transferring style between motions with different contents. This challenge lies in the lack of clear separation between content and style of a motion. To tackle this challenge, we propose a novel motion style transformer that effectively disentangles style from content and generates a plausible motion with transferred style from a source motion. Our distinctive approach to achieving the goal of disentanglement is twofold: (1) a new architecture for motion style transformer with 'part-attentive style modulator across body parts' and 'Siamese encoders that encode style and content features separately'; (2) style disentanglement loss. Our method outperforms existing methods and demonstrates exceptionally high quality, particularly in motion pairs with different contents, without the need for heuristic post-processing. Codes are available at https://github.com/Boeun-Kim/MoST.

Finding Visual Saliency in Continuous Spike Stream

• Authors: Lin Zhu, Xianzhang Chen, Xiao Wang, Hua Huang
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06233
• Pdf link: https://arxiv.org/pdf/2403.06233
• Abstract As a bio-inspired vision sensor, the spike camera emulates the operational principles of the fovea, a compact retinal region, by employing spike discharges to encode the accumulation of per-pixel luminance intensity. Leveraging its high temporal resolution and bio-inspired neuromorphic design, the spike camera holds significant promise for advancing computer vision applications. Saliency detection mimics the behavior of human beings and captures the most salient region from the scenes. In this paper, we investigate the visual saliency in the continuous spike stream for the first time. To effectively process the binary spike stream, we propose a Recurrent Spiking Transformer (RST) framework, which is based on a full spiking neural network. Our framework enables the extraction of spatio-temporal features from the continuous spatio-temporal spike stream while maintaining low power consumption. To facilitate the training and validation of our proposed model, we build a comprehensive real-world spike-based visual saliency dataset, enriched with numerous light conditions. Extensive experiments demonstrate the superior performance of our Recurrent Spiking Transformer framework in comparison to other spike neural network-based methods. Our framework exhibits a substantial margin of improvement in capturing and highlighting visual saliency in the spike stream, which not only provides a new perspective for spike-based saliency segmentation but also shows a new paradigm for full SNN-based transformer models. The code and dataset are available at \url{https://github.com/BIT-Vision/SVS}.

Transformer based Multitask Learning for Image Captioning and Object Detection

• Authors: Debolena Basak, P.K. Srijith, Maunendra Sankar Desarkar
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2403.06292
• Pdf link: https://arxiv.org/pdf/2403.06292
• Abstract In several real-world scenarios like autonomous navigation and mobility, to obtain a better visual understanding of the surroundings, image captioning and object detection play a crucial role. This work introduces a novel multitask learning framework that combines image captioning and object detection into a joint model. We propose TICOD, Transformer-based Image Captioning and Object detection model for jointly training both tasks by combining the losses obtained from image captioning and object detection networks. By leveraging joint training, the model benefits from the complementary information shared between the two tasks, leading to improved performance for image captioning. Our approach utilizes a transformer-based architecture that enables end-to-end network integration for image captioning and object detection and performs both tasks jointly. We evaluate the effectiveness of our approach through comprehensive experiments on the MS-COCO dataset. Our model outperforms the baselines from image captioning literature by achieving a 3.65% improvement in BERTScore.

Point Mamba: A Novel Point Cloud Backbone Based on State Space Model with Octree-Based Ordering Strategy

• Authors: Jiuming Liu, Ruiji Yu, Yian Wang, Yu Zheng, Tianchen Deng, Weicai Ye, Hesheng Wang
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06467
• Pdf link: https://arxiv.org/pdf/2403.06467
• Abstract Recently, state space model (SSM) has gained great attention due to its promising performance, linear complexity, and long sequence modeling ability in both language and image domains. However, it is non-trivial to extend SSM to the point cloud field, because of the causality requirement of SSM and the disorder and irregularity nature of point clouds. In this paper, we propose a novel SSM-based point cloud processing backbone, named Point Mamba, with a causality-aware ordering mechanism. To construct the causal dependency relationship, we design an octree-based ordering strategy on raw irregular points, globally sorting points in a z-order sequence and also retaining their spatial proximity. Our method achieves state-of-the-art performance compared with transformer-based counterparts, with 93.4% accuracy and 75.7 mIOU respectively on the ModelNet40 classification dataset and ScanNet semantic segmentation dataset. Furthermore, our Point Mamba has linear complexity, which is more efficient than transformer-based methods. Our method demonstrates the great potential that SSM can serve as a generic backbone in point cloud understanding. Codes are released at https://github.com/IRMVLab/Point-Mamba.

Multilingual Turn-taking Prediction Using Voice Activity Projection

• Authors: Koji Inoue, Bing'er Jiang, Erik Ekstedt, Tatsuya Kawahara, Gabriel Skantze
• Subjects: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
• Arxiv link: https://arxiv.org/abs/2403.06487
• Pdf link: https://arxiv.org/pdf/2403.06487
• Abstract This paper investigates the application of voice activity projection (VAP), a predictive turn-taking model for spoken dialogue, on multilingual data, encompassing English, Mandarin, and Japanese. The VAP model continuously predicts the upcoming voice activities of participants in dyadic dialogue, leveraging a cross-attention Transformer to capture the dynamic interplay between participants. The results show that a monolingual VAP model trained on one language does not make good predictions when applied to other languages. However, a multilingual model, trained on all three languages, demonstrates predictive performance on par with monolingual models across all languages. Further analyses show that the multilingual model has learned to discern the language of the input signal. We also analyze the sensitivity to pitch, a prosodic cue that is thought to be important for turn-taking. Finally, we compare two different audio encoders, contrastive predictive coding (CPC) pre-trained on English, with a recent model based on multilingual wav2vec 2.0 (MMS).

QuantTune: Optimizing Model Quantization with Adaptive Outlier-Driven Fine Tuning

• Authors: Jiun-Man Chen, Yu-Hsuan Chao, Yu-Jie Wang, Ming-Der Shieh, Chih-Chung Hsu, Wei-Fen Lin
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
• Arxiv link: https://arxiv.org/abs/2403.06497
• Pdf link: https://arxiv.org/pdf/2403.06497
• Abstract Transformer-based models have gained widespread popularity in both the computer vision (CV) and natural language processing (NLP) fields. However, significant challenges arise during post-training linear quantization, leading to noticeable reductions in inference accuracy. Our study focuses on uncovering the underlying causes of these accuracy drops and proposing a quantization-friendly fine-tuning method, \textbf{QuantTune}. Firstly, our analysis revealed that, on average, 65% of quantization errors result from the precision loss incurred by the dynamic range amplification effect of outliers across the target Transformer-based models. Secondly, \textbf{QuantTune} adjusts weights based on the deviation of outlier activations and effectively constrains the dynamic ranges of the problematic activations. As a result, it successfully mitigates the negative impact of outliers on the inference accuracy of quantized models. Lastly, \textbf{QuantTune} can be seamlessly integrated into the back-propagation pass in the fine-tuning process without requiring extra complexity in inference software and hardware design. Our approach showcases significant improvements in post-training quantization across a range of Transformer-based models, including ViT, Bert-base, and OPT. QuantTune reduces accuracy drops by 12.09% at 8-bit quantization and 33.8% at 7-bit compared to top calibration methods, outperforming state-of-the-art solutions by over 18.84% across ViT models.

Autonomous Overhead Powerline Recharging for Uninterrupted Drone Operations

• Authors: Viet Duong Hoang, Frederik Falk Nyboe, Nicolaj Haarhøj Malle, Emad Ebeid
• Subjects: Robotics (cs.RO)
• Arxiv link: https://arxiv.org/abs/2403.06533
• Pdf link: https://arxiv.org/pdf/2403.06533
• Abstract We present a fully autonomous self-recharging drone system capable of long-duration sustained operations near powerlines. The drone is equipped with a robust onboard perception and navigation system that enables it to locate powerlines and approach them for landing. A passively actuated gripping mechanism grasps the powerline cable during landing after which a control circuit regulates the magnetic field inside a split-core current transformer to provide sufficient holding force as well as battery recharging. The system is evaluated in an active outdoor three-phase powerline environment. We demonstrate multiple contiguous hours of fully autonomous uninterrupted drone operations composed of several cycles of flying, landing, recharging, and takeoff, validating the capability of extended, essentially unlimited, operational endurance.

Multi-Scale Implicit Transformer with Re-parameterize for Arbitrary-Scale Super-Resolution

• Authors: Jinchen Zhu, Mingjian Zhang, Ling Zheng, Shizhuang Weng
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06536
• Pdf link: https://arxiv.org/pdf/2403.06536
• Abstract Recently, the methods based on implicit neural representations have shown excellent capabilities for arbitrary-scale super-resolution (ASSR). Although these methods represent the features of an image by generating latent codes, these latent codes are difficult to adapt for different magnification factors of super-resolution, which seriously affects their performance. Addressing this, we design Multi-Scale Implicit Transformer (MSIT), consisting of an Multi-scale Neural Operator (MSNO) and Multi-Scale Self-Attention (MSSA). Among them, MSNO obtains multi-scale latent codes through feature enhancement, multi-scale characteristics extraction, and multi-scale characteristics merging. MSSA further enhances the multi-scale characteristics of latent codes, resulting in better performance. Furthermore, to improve the performance of network, we propose the Re-Interaction Module (RIM) combined with the cumulative training strategy to improve the diversity of learned information for the network. We have systematically introduced multi-scale characteristics for the first time in ASSR, extensive experiments are performed to validate the effectiveness of MSIT, and our method achieves state-of-the-art performance in arbitrary super-resolution tasks.

Transformer-based Fusion of 2D-pose and Spatio-temporal Embeddings for Distracted Driver Action Recognition

• Authors: Erkut Akdag, Zeqi Zhu, Egor Bondarev, Peter H. N. De With
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06577
• Pdf link: https://arxiv.org/pdf/2403.06577
• Abstract Classification and localization of driving actions over time is important for advanced driver-assistance systems and naturalistic driving studies. Temporal localization is challenging because it requires robustness, reliability, and accuracy. In this study, we aim to improve the temporal localization and classification accuracy performance by adapting video action recognition and 2D human-pose estimation networks to one model. Therefore, we design a transformer-based fusion architecture to effectively combine 2D-pose features and spatio-temporal features. The model uses 2D-pose features as the positional embedding of the transformer architecture and spatio-temporal features as the main input to the encoder of the transformer. The proposed solution is generic and independent of the camera numbers and positions, giving frame-based class probabilities as output. Finally, the post-processing step combines information from different camera views to obtain final predictions and eliminate false positives. The model performs well on the A2 test set of the 2023 NVIDIA AI City Challenge for naturalistic driving action recognition, achieving the overlap score of the organizer-defined distracted driver behaviour metric of 0.5079.

Cross-domain and Cross-dimension Learning for Image-to-Graph Transformers

• Authors: Alexander H. Berger, Laurin Lux, Suprosanna Shit, Ivan Ezhov, Georgios Kaissis, Martin J. Menten, Daniel Rueckert, Johannes C. Paetzold
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06601
• Pdf link: https://arxiv.org/pdf/2403.06601
• Abstract Direct image-to-graph transformation is a challenging task that solves object detection and relationship prediction in a single model. Due to the complexity of this task, large training datasets are rare in many domains, which makes the training of large networks challenging. This data sparsity necessitates the establishment of pre-training strategies akin to the state-of-the-art in computer vision. In this work, we introduce a set of methods enabling cross-domain and cross-dimension transfer learning for image-to-graph transformers. We propose (1) a regularized edge sampling loss for sampling the optimal number of object relationships (edges) across domains, (2) a domain adaptation framework for image-to-graph transformers that aligns features from different domains, and (3) a simple projection function that allows us to pretrain 3D transformers on 2D input data. We demonstrate our method's utility in cross-domain and cross-dimension experiments, where we pretrain our models on 2D satellite images before applying them to vastly different target domains in 2D and 3D. Our method consistently outperforms a series of baselines on challenging benchmarks, such as retinal or whole-brain vessel graph extraction.

Stealing Part of a Production Language Model

• Authors: Nicholas Carlini, Daniel Paleka, Krishnamurthy Dj Dvijotham, Thomas Steinke, Jonathan Hayase, A. Feder Cooper, Katherine Lee, Matthew Jagielski, Milad Nasr, Arthur Conmy, Eric Wallace, David Rolnick, Florian Tramèr
• Subjects: Cryptography and Security (cs.CR)
• Arxiv link: https://arxiv.org/abs/2403.06634
• Pdf link: https://arxiv.org/pdf/2403.06634
• Abstract We introduce the first model-stealing attack that extracts precise, nontrivial information from black-box production language models like OpenAI's ChatGPT or Google's PaLM-2. Specifically, our attack recovers the embedding projection layer (up to symmetries) of a transformer model, given typical API access. For under $20 USD, our attack extracts the entire projection matrix of OpenAI's Ada and Babbage language models. We thereby confirm, for the first time, that these black-box models have a hidden dimension of 1024 and 2048, respectively. We also recover the exact hidden dimension size of the gpt-3.5-turbo model, and estimate it would cost under $2,000 in queries to recover the entire projection matrix. We conclude with potential defenses and mitigations, and discuss the implications of possible future work that could extend our attack.

CEAT: Continual Expansion and Absorption Transformer for Non-Exemplar Class-Incremental Learnin

• Authors: Xinyuan Gao, Songlin Dong, Yuhang He, Xing Wei, Yihong Gong
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06670
• Pdf link: https://arxiv.org/pdf/2403.06670
• Abstract In real-world applications, dynamic scenarios require the models to possess the capability to learn new tasks continuously without forgetting the old knowledge. Experience-Replay methods store a subset of the old images for joint training. In the scenario of more strict privacy protection, storing the old images becomes infeasible, which leads to a more severe plasticity-stability dilemma and classifier bias. To meet the above challenges, we propose a new architecture, named continual expansion and absorption transformer~(CEAT). The model can learn the novel knowledge by extending the expanded-fusion layers in parallel with the frozen previous parameters. After the task ends, we losslessly absorb the extended parameters into the backbone to ensure that the number of parameters remains constant. To improve the learning ability of the model, we designed a novel prototype contrastive loss to reduce the overlap between old and new classes in the feature space. Besides, to address the classifier bias towards the new classes, we propose a novel approach to generate the pseudo-features to correct the classifier. We experiment with our methods on three standard Non-Exemplar Class-Incremental Learning~(NECIL) benchmarks. Extensive experiments demonstrate that our model gets a significant improvement compared with the previous works and achieves 5.38%, 5.20%, and 4.92% improvement on CIFAR-100, TinyImageNet, and ImageNet-Subset.

CAM Back Again: Large Kernel CNNs from a Weakly Supervised Object Localization Perspective

• Authors: Shunsuke Yasuki, Masato Taki
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06676
• Pdf link: https://arxiv.org/pdf/2403.06676
• Abstract Recently, convolutional neural networks (CNNs) with large size kernels have attracted much attention in the computer vision field, following the success of the Vision Transformers. Large kernel CNNs have been reported to perform well in downstream vision tasks as well as in classification performance. The reason for the high-performance of large kernel CNNs in downstream tasks has been attributed to the large effective receptive field (ERF) produced by large size kernels, but this view has not been fully tested. We therefore revisit the performance of large kernel CNNs in downstream task, focusing on the weakly supervised object localization (WSOL) task. WSOL, a difficult downstream task that is not fully supervised, provides a new angle to explore the capabilities of the large kernel CNNs. Our study compares the modern large kernel CNNs ConvNeXt, RepLKNet, and SLaK to test the validity of the naive expectation that ERF size is important for improving downstream task performance. Our analysis of the factors contributing to high performance provides a different perspective, in which the main factor is feature map improvement. Furthermore, we find that modern CNNs are robust to the CAM problems of local regions of objects being activated, which has long been discussed in WSOL. CAM is the most classic WSOL method, but because of the above-mentioned problems, it is often used as a baseline method for comparison. However, experiments on the CUB-200-2011 dataset show that simply combining a large kernel CNN, CAM, and simple data augmentation methods can achieve performance (90.99% MaxBoxAcc) comparable to the latest WSOL method, which is CNN-based and requires special training or complex post-processing. The code is available at https://github.com/snskysk/CAM-Back-Again.

Advancing Graph Neural Networks with HL-HGAT: A Hodge-Laplacian and Attention Mechanism Approach for Heterogeneous Graph-Structured Data

• Authors: Jinghan Huang, Qiufeng Chen, Yijun Bian, Pengli Zhu, Nanguang Chen, Moo K. Chung, Anqi Qiu
• Subjects: Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06687
• Pdf link: https://arxiv.org/pdf/2403.06687
• Abstract Graph neural networks (GNNs) have proven effective in capturing relationships among nodes in a graph. This study introduces a novel perspective by considering a graph as a simplicial complex, encompassing nodes, edges, triangles, and $k$-simplices, enabling the definition of graph-structured data on any $k$-simplices. Our contribution is the Hodge-Laplacian heterogeneous graph attention network (HL-HGAT), designed to learn heterogeneous signal representations across $k$-simplices. The HL-HGAT incorporates three key components: HL convolutional filters (HL-filters), simplicial projection (SP), and simplicial attention pooling (SAP) operators, applied to $k$-simplices. HL-filters leverage the unique topology of $k$-simplices encoded by the Hodge-Laplacian (HL) operator, operating within the spectral domain of the $k$-th HL operator. To address computation challenges, we introduce a polynomial approximation for HL-filters, exhibiting spatial localization properties. Additionally, we propose a pooling operator to coarsen $k$-simplices, combining features through simplicial attention mechanisms of self-attention and cross-attention via transformers and SP operators, capturing topological interconnections across multiple dimensions of simplices. The HL-HGAT is comprehensively evaluated across diverse graph applications, including NP-hard problems, graph multi-label and classification challenges, and graph regression tasks in logistics, computer vision, biology, chemistry, and neuroscience. The results demonstrate the model's efficacy and versatility in handling a wide range of graph-based scenarios.

Multimodal Transformers for Real-Time Surgical Activity Prediction

• Authors: Keshara Weerasinghe, Seyed Hamid Reza Roodabeh, Kay Hutchinson, Homa Alemzadeh
• Subjects: Robotics (cs.RO)
• Arxiv link: https://arxiv.org/abs/2403.06705
• Pdf link: https://arxiv.org/pdf/2403.06705
• Abstract Real-time recognition and prediction of surgical activities are fundamental to advancing safety and autonomy in robot-assisted surgery. This paper presents a multimodal transformer architecture for real-time recognition and prediction of surgical gestures and trajectories based on short segments of kinematic and video data. We conduct an ablation study to evaluate the impact of fusing different input modalities and their representations on gesture recognition and prediction performance. We perform an end-to-end assessment of the proposed architecture using the JHU-ISI Gesture and Skill Assessment Working Set (JIGSAWS) dataset. Our model outperforms the state-of-the-art (SOTA) with 89.5% accuracy for gesture prediction through effective fusion of kinematic features with spatial and contextual video features. It achieves the real-time performance of 1.1-1.3ms for processing a 1-second input window by relying on a computationally efficient model.

Improving Low-Resource Knowledge Tracing Tasks by Supervised Pre-training and Importance Mechanism Fine-tuning

• Authors: Hengyuan Zhang, Zitao Liu, Shuyan Huang, Chenming Shang, Bojun Zhan, Yong Jiang
• Subjects: Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2403.06725
• Pdf link: https://arxiv.org/pdf/2403.06725
• Abstract Knowledge tracing (KT) aims to estimate student's knowledge mastery based on their historical interactions. Recently, the deep learning based KT (DLKT) approaches have achieved impressive performance in the KT task. These DLKT models heavily rely on the large number of available student interactions. However, due to various reasons such as budget constraints and privacy concerns, observed interactions are very limited in many real-world scenarios, a.k.a, low-resource KT datasets. Directly training a DLKT model on a low-resource KT dataset may lead to overfitting and it is difficult to choose the appropriate deep neural architecture. Therefore, in this paper, we propose a low-resource KT framework called LoReKT to address above challenges. Inspired by the prevalent "pre-training and fine-tuning" paradigm, we aim to learn transferable parameters and representations from rich-resource KT datasets during the pre-training stage and subsequently facilitate effective adaptation to low-resource KT datasets. Specifically, we simplify existing sophisticated DLKT model architectures with purely a stack of transformer decoders. We design an encoding mechanism to incorporate student interactions from multiple KT data sources and develop an importance mechanism to prioritize updating parameters with high importance while constraining less important ones during the fine-tuning stage. We evaluate LoReKT on six public KT datasets and experimental results demonstrate the superiority of our approach in terms of AUC and Accuracy. To encourage reproducible research, we make our data and code publicly available at https://anonymous.4open.science/r/LoReKT-C619.

In-context Exploration-Exploitation for Reinforcement Learning

• Authors: Zhenwen Dai, Federico Tomasi, Sina Ghiassian
• Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
• Arxiv link: https://arxiv.org/abs/2403.06826
• Pdf link: https://arxiv.org/pdf/2403.06826
• Abstract In-context learning is a promising approach for online policy learning of offline reinforcement learning (RL) methods, which can be achieved at inference time without gradient optimization. However, this method is hindered by significant computational costs resulting from the gathering of large training trajectory sets and the need to train large Transformer models. We address this challenge by introducing an In-context Exploration-Exploitation (ICEE) algorithm, designed to optimize the efficiency of in-context policy learning. Unlike existing models, ICEE performs an exploration-exploitation trade-off at inference time within a Transformer model, without the need for explicit Bayesian inference. Consequently, ICEE can solve Bayesian optimization problems as efficiently as Gaussian process biased methods do, but in significantly less time. Through experiments in grid world environments, we demonstrate that ICEE can learn to solve new RL tasks using only tens of episodes, marking a substantial improvement over the hundreds of episodes needed by the previous in-context learning method.

HDRTransDC: High Dynamic Range Image Reconstruction with Transformer Deformation Convolution

• Authors: Shuaikang Shang, Xuejing Kang, Anlong Ming
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06831
• Pdf link: https://arxiv.org/pdf/2403.06831
• Abstract High Dynamic Range (HDR) imaging aims to generate an artifact-free HDR image with realistic details by fusing multi-exposure Low Dynamic Range (LDR) images. Caused by large motion and severe under-/over-exposure among input LDR images, HDR imaging suffers from ghosting artifacts and fusion distortions. To address these critical issues, we propose an HDR Transformer Deformation Convolution (HDRTransDC) network to generate high-quality HDR images, which consists of the Transformer Deformable Convolution Alignment Module (TDCAM) and the Dynamic Weight Fusion Block (DWFB). To solve the ghosting artifacts, the proposed TDCAM extracts long-distance content similar to the reference feature in the entire non-reference features, which can accurately remove misalignment and fill the content occluded by moving objects. For the purpose of eliminating fusion distortions, we propose DWFB to spatially adaptively select useful information across frames to effectively fuse multi-exposed features. Extensive experiments show that our method quantitatively and qualitatively achieves state-of-the-art performance.

The Power of Noise: Toward a Unified Multi-modal Knowledge Graph Representation Framework

• Authors: Zhuo Chen, Yin Fang, Yichi Zhang, Lingbing Guo, Jiaoyan Chen, Huajun Chen, Wen Zhang
• Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06832
• Pdf link: https://arxiv.org/pdf/2403.06832
• Abstract The advancement of Multi-modal Pre-training highlights the necessity for a robust Multi-Modal Knowledge Graph (MMKG) representation learning framework. This framework is crucial for integrating structured knowledge into multi-modal Large Language Models (LLMs) at scale, aiming to alleviate issues like knowledge misconceptions and multi-modal hallucinations. In this work, to evaluate models' ability to accurately embed entities within MMKGs, we focus on two widely researched tasks: Multi-modal Knowledge Graph Completion (MKGC) and Multi-modal Entity Alignment (MMEA). Building on this foundation, we propose a novel SNAG method that utilizes a Transformer-based architecture equipped with modality-level noise masking for the robust integration of multi-modal entity features in KGs. By incorporating specific training objectives for both MKGC and MMEA, our approach achieves SOTA performance across a total of ten datasets (three for MKGC and seven for MEMA), demonstrating its robustness and versatility. Besides, SNAG can not only function as a standalone model but also enhance other existing methods, providing stable performance improvements. Our code and data are available at: https://github.com/zjukg/SNAG.

On the Generalization Ability of Unsupervised Pretraining

• Authors: Yuyang Deng, Junyuan Hong, Jiayu Zhou, Mehrdad Mahdavi
• Subjects: Machine Learning (cs.LG); Machine Learning (stat.ML)
• Arxiv link: https://arxiv.org/abs/2403.06871
• Pdf link: https://arxiv.org/pdf/2403.06871
• Abstract Recent advances in unsupervised learning have shown that unsupervised pre-training, followed by fine-tuning, can improve model generalization. However, a rigorous understanding of how the representation function learned on an unlabeled dataset affects the generalization of the fine-tuned model is lacking. Existing theoretical research does not adequately account for the heterogeneity of the distribution and tasks in pre-training and fine-tuning stage. To bridge this gap, this paper introduces a novel theoretical framework that illuminates the critical factor influencing the transferability of knowledge acquired during unsupervised pre-training to the subsequent fine-tuning phase, ultimately affecting the generalization capabilities of the fine-tuned model on downstream tasks. We apply our theoretical framework to analyze generalization bound of two distinct scenarios: Context Encoder pre-training with deep neural networks and Masked Autoencoder pre-training with deep transformers, followed by fine-tuning on a binary classification task. Finally, inspired by our findings, we propose a novel regularization method during pre-training to further enhances the generalization of fine-tuned model. Overall, our results contribute to a better understanding of unsupervised pre-training and fine-tuning paradigm, and can shed light on the design of more effective pre-training algorithms.

Exploring Large Language Models and Hierarchical Frameworks for Classification of Large Unstructured Legal Documents

• Authors: Nishchal Prasad, Mohand Boughanem, Taoufiq Dkaki
• Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2403.06872
• Pdf link: https://arxiv.org/pdf/2403.06872
• Abstract Legal judgment prediction suffers from the problem of long case documents exceeding tens of thousands of words, in general, and having a non-uniform structure. Predicting judgments from such documents becomes a challenging task, more so on documents with no structural annotation. We explore the classification of these large legal documents and their lack of structural information with a deep-learning-based hierarchical framework which we call MESc; "Multi-stage Encoder-based Supervised with-clustering"; for judgment prediction. Specifically, we divide a document into parts to extract their embeddings from the last four layers of a custom fine-tuned Large Language Model, and try to approximate their structure through unsupervised clustering. Which we use in another set of transformer encoder layers to learn the inter-chunk representations. We analyze the adaptability of Large Language Models (LLMs) with multi-billion parameters (GPT-Neo, and GPT-J) with the hierarchical framework of MESc and compare them with their standalone performance on legal texts. We also study their intra-domain(legal) transfer learning capability and the impact of combining embeddings from their last layers in MESc. We test these methods and their effectiveness with extensive experiments and ablation studies on legal documents from India, the European Union, and the United States with the ILDC dataset and a subset of the LexGLUE dataset. Our approach achieves a minimum total performance gain of approximately 2 points over previous state-of-the-art methods.

Real-time Transformer-based Open-Vocabulary Detection with Efficient Fusion Head

• Authors: Tiancheng Zhao, Peng Liu, Xuan He, Lu Zhang, Kyusong Lee
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2403.06892
• Pdf link: https://arxiv.org/pdf/2403.06892
• Abstract End-to-end transformer-based detectors (DETRs) have shown exceptional performance in both closed-set and open-vocabulary object detection (OVD) tasks through the integration of language modalities. However, their demanding computational requirements have hindered their practical application in real-time object detection (OD) scenarios. In this paper, we scrutinize the limitations of two leading models in the OVDEval benchmark, OmDet and Grounding-DINO, and introduce OmDet-Turbo. This novel transformer-based real-time OVD model features an innovative Efficient Fusion Head (EFH) module designed to alleviate the bottlenecks observed in OmDet and Grounding-DINO. Notably, OmDet-Turbo-Base achieves a 100.2 frames per second (FPS) with TensorRT and language cache techniques applied. Notably, in zero-shot scenarios on COCO and LVIS datasets, OmDet-Turbo achieves performance levels nearly on par with current state-of-the-art supervised models. Furthermore, it establishes new state-of-the-art benchmarks on ODinW and OVDEval, boasting an AP of 30.1 and an NMS-AP of 26.86, respectively. The practicality of OmDet-Turbo in industrial applications is underscored by its exceptional performance on benchmark datasets and superior inference speed, positioning it as a compelling choice for real-time object detection tasks. Code: \url{https://github.com/om-ai-lab/OmDet}

GRITv2: Efficient and Light-weight Social Relation Recognition

• Authors: N K Sagar Reddy, Neeraj Kasera, Avinash Thakur
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06895
• Pdf link: https://arxiv.org/pdf/2403.06895
• Abstract Our research focuses on the analysis and improvement of the Graph-based Relation Inference Transformer (GRIT), which serves as an important benchmark in the field. We conduct a comprehensive ablation study using the PISC-fine dataset, to find and explore improvement in efficiency and performance of GRITv2. Our research has provided a new state-of-the-art relation recognition model on the PISC relation dataset. We introduce several features in the GRIT model and analyse our new benchmarks in two versions: GRITv2-L (large) and GRITv2-S (small). Our proposed GRITv2-L surpasses existing methods on relation recognition and the GRITv2-S is within 2% performance gap of GRITv2-L, which has only 0.0625x the model size and parameters of GRITv2-L. Furthermore, we also address the need for model compression, an area crucial for deploying efficient models on resource-constrained platforms. By applying quantization techniques, we efficiently reduced the GRITv2-S size to 22MB and deployed it on the flagship OnePlus 12 mobile which still surpasses the PISC-fine benchmarks in performance, highlighting the practical viability and improved efficiency of our model on mobile devices.

Simplicity Bias of Transformers to Learn Low Sensitivity Functions

• Authors: Bhavya Vasudeva, Deqing Fu, Tianyi Zhou, Elliott Kau, Youqi Huang, Vatsal Sharan
• Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (stat.ML)
• Arxiv link: https://arxiv.org/abs/2403.06925
• Pdf link: https://arxiv.org/pdf/2403.06925
• Abstract Transformers achieve state-of-the-art accuracy and robustness across many tasks, but an understanding of the inductive biases that they have and how those biases are different from other neural network architectures remains elusive. Various neural network architectures such as fully connected networks have been found to have a simplicity bias towards simple functions of the data; one version of this simplicity bias is a spectral bias to learn simple functions in the Fourier space. In this work, we identify the notion of sensitivity of the model to random changes in the input as a notion of simplicity bias which provides a unified metric to explain the simplicity and spectral bias of transformers across different data modalities. We show that transformers have lower sensitivity than alternative architectures, such as LSTMs, MLPs and CNNs, across both vision and language tasks. We also show that low-sensitivity bias correlates with improved robustness; furthermore, it can also be used as an efficient intervention to further improve the robustness of transformers.

Explainable Transformer Prototypes for Medical Diagnoses

• Authors: Ugur Demir, Debesh Jha, Zheyuan Zhang, Elif Keles, Bradley Allen, Aggelos K. Katsaggelos, Ulas Bagci
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06961
• Pdf link: https://arxiv.org/pdf/2403.06961
• Abstract Deployments of artificial intelligence in medical diagnostics mandate not just accuracy and efficacy but also trust, emphasizing the need for explainability in machine decisions. The recent trend in automated medical image diagnostics leans towards the deployment of Transformer-based architectures, credited to their impressive capabilities. Since the self-attention feature of transformers contributes towards identifying crucial regions during the classification process, they enhance the trustability of the methods. However, the complex intricacies of these attention mechanisms may fall short of effectively pinpointing the regions of interest directly influencing AI decisions. Our research endeavors to innovate a unique attention block that underscores the correlation between 'regions' rather than 'pixels'. To address this challenge, we introduce an innovative system grounded in prototype learning, featuring an advanced self-attention mechanism that goes beyond conventional ad-hoc visual explanation techniques by offering comprehensible visual insights. A combined quantitative and qualitative methodological approach was used to demonstrate the effectiveness of the proposed method on the large-scale NIH chest X-ray dataset. Experimental results showed that our proposed method offers a promising direction for explainability, which can lead to the development of more trustable systems, which can facilitate easier and rapid adoption of such technology into routine clinics. The code is available at www.github.com/NUBagcilab/r2r_proto.

The pitfalls of next-token prediction

• Authors: Gregor Bachmann, Vaishnavh Nagarajan
• Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2403.06963
• Pdf link: https://arxiv.org/pdf/2403.06963
• Abstract Can a mere next-token predictor faithfully model human intelligence? We crystallize this intuitive concern, which is fragmented in the literature. As a starting point, we argue that the two often-conflated phases of next-token prediction -- autoregressive inference and teacher-forced training -- must be treated distinctly. The popular criticism that errors can compound during autoregressive inference, crucially assumes that teacher-forcing has learned an accurate next-token predictor. This assumption sidesteps a more deep-rooted problem we expose: in certain classes of tasks, teacher-forcing can simply fail to learn an accurate next-token predictor in the first place. We describe a general mechanism of how teacher-forcing can fail, and design a minimal planning task where both the Transformer and the Mamba architecture empirically fail in that manner -- remarkably, despite the task being straightforward to learn. We provide preliminary evidence that this failure can be resolved when training to predict multiple tokens in advance. We hope this finding can ground future debates and inspire explorations beyond the next-token prediction paradigm. We make our code available under https://github.com/gregorbachmann/Next-Token-Failures

VideoMamba: State Space Model for Efficient Video Understanding

• Authors: Kunchang Li, Xinhao Li, Yi Wang, Yinan He, Yali Wang, Limin Wang, Yu Qiao
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06977
• Pdf link: https://arxiv.org/pdf/2403.06977
• Abstract Addressing the dual challenges of local redundancy and global dependencies in video understanding, this work innovatively adapts the Mamba to the video domain. The proposed VideoMamba overcomes the limitations of existing 3D convolution neural networks and video transformers. Its linear-complexity operator enables efficient long-term modeling, which is crucial for high-resolution long video understanding. Extensive evaluations reveal VideoMamba's four core abilities: (1) Scalability in the visual domain without extensive dataset pretraining, thanks to a novel self-distillation technique; (2) Sensitivity for recognizing short-term actions even with fine-grained motion differences; (3) Superiority in long-term video understanding, showcasing significant advancements over traditional feature-based models; and (4) Compatibility with other modalities, demonstrating robustness in multi-modal contexts. Through these distinct advantages, VideoMamba sets a new benchmark for video understanding, offering a scalable and efficient solution for comprehensive video understanding. All the code and models are available at https://github.com/OpenGVLab/VideoMamba.

Attention Prompt Tuning: Parameter-efficient Adaptation of Pre-trained Models for Spatiotemporal Modeling

• Authors: Wele Gedara Chaminda Bandara, Vishal M. Patel
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06978
• Pdf link: https://arxiv.org/pdf/2403.06978
• Abstract In this paper, we introduce Attention Prompt Tuning (APT) - a computationally efficient variant of prompt tuning for video-based applications such as action recognition. Prompt tuning approaches involve injecting a set of learnable prompts along with data tokens during fine-tuning while keeping the backbone frozen. This approach greatly reduces the number of learnable parameters compared to full tuning. For image-based downstream tasks, normally a couple of learnable prompts achieve results close to those of full tuning. However, videos, which contain more complex spatiotemporal information, require hundreds of tunable prompts to achieve reasonably good results. This reduces the parameter efficiency observed in images and significantly increases latency and the number of floating-point operations (FLOPs) during inference. To tackle these issues, we directly inject the prompts into the keys and values of the non-local attention mechanism within the transformer block. Additionally, we introduce a novel prompt reparameterization technique to make APT more robust against hyperparameter selection. The proposed APT approach greatly reduces the number of FLOPs and latency while achieving a significant performance boost over the existing parameter-efficient tuning methods on UCF101, HMDB51, and SSv2 datasets for action recognition. The code and pre-trained models are available at https://github.com/wgcban/apt

Keyword: scene understanding

Optimizing Latent Graph Representations of Surgical Scenes for Zero-Shot Domain Transfer

• Authors: Siddhant Satyanaik, Aditya Murali, Deepak Alapatt, Xin Wang, Pietro Mascagni, Nicolas Padoy
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06953
• Pdf link: https://arxiv.org/pdf/2403.06953
• Abstract Purpose: Advances in deep learning have resulted in effective models for surgical video analysis; however, these models often fail to generalize across medical centers due to domain shift caused by variations in surgical workflow, camera setups, and patient demographics. Recently, object-centric learning has emerged as a promising approach for improved surgical scene understanding, capturing and disentangling visual and semantic properties of surgical tools and anatomy to improve downstream task performance. In this work, we conduct a multi-centric performance benchmark of object-centric approaches, focusing on Critical View of Safety assessment in laparoscopic cholecystectomy, then propose an improved approach for unseen domain generalization. Methods: We evaluate four object-centric approaches for domain generalization, establishing baseline performance. Next, leveraging the disentangled nature of object-centric representations, we dissect one of these methods through a series of ablations (e.g. ignoring either visual or semantic features for downstream classification). Finally, based on the results of these ablations, we develop an optimized method specifically tailored for domain generalization, LG-DG, that includes a novel disentanglement loss function. Results: Our optimized approach, LG-DG, achieves an improvement of 9.28% over the best baseline approach. More broadly, we show that object-centric approaches are highly effective for domain generalization thanks to their modular approach to representation learning. Conclusion: We investigate the use of object-centric methods for unseen domain generalization, identify method-agnostic factors critical for performance, and present an optimized approach that substantially outperforms existing methods.

Keyword: visual reasoning

Test-time Distribution Learning Adapter for Cross-modal Visual Reasoning

• Authors: Yi Zhang, Ce Zhang
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2403.06059
• Pdf link: https://arxiv.org/pdf/2403.06059
• Abstract Vision-Language Pre-Trained (VLP) models, such as CLIP, have demonstrated remarkable effectiveness in learning generic visual representations. Several approaches aim to efficiently adapt VLP models to downstream tasks with limited supervision, aiming to leverage the acquired knowledge from VLP models. However, these methods suffer from either introducing biased representations or requiring high computational complexity, which hinders their effectiveness in fine-tuning the CLIP model. Moreover, when a model is trained on data specific to a particular domain, its ability to generalize to uncharted domains diminishes. In this work, we propose Test-Time Distribution LearNing Adapter (TT-DNA) which directly works during the testing period. Specifically, we estimate Gaussian distributions to model visual features of the few-shot support images to capture the knowledge from the support set. The cosine similarity between query image and the feature distribution of support images is used as the prediction of visual adapter. Subsequently, the visual adapter's prediction merges with the original CLIP prediction via a residual connection, resulting in the final prediction. Our extensive experimental results on visual reasoning for human object interaction demonstrate that our proposed TT-DNA outperforms existing state-of-the-art methods by large margins.