New submissions for Thu, 16 Jun 22

[zhuhu00]

New submissions for Thu, 16 Jun 22

Keyword: SLAM

There is no result

Keyword: odometry

There is no result

Keyword: livox

There is no result

Keyword: loam

There is no result

Keyword: lidar

MonoGround: Detecting Monocular 3D Objects from the Ground

• Authors: Zequn Qin, Xi Li
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07372
• Pdf link: https://arxiv.org/pdf/2206.07372
• Abstract Monocular 3D object detection has attracted great attention for its advantages in simplicity and cost. Due to the ill-posed 2D to 3D mapping essence from the monocular imaging process, monocular 3D object detection suffers from inaccurate depth estimation and thus has poor 3D detection results. To alleviate this problem, we propose to introduce the ground plane as a prior in the monocular 3d object detection. The ground plane prior serves as an additional geometric condition to the ill-posed mapping and an extra source in depth estimation. In this way, we can get a more accurate depth estimation from the ground. Meanwhile, to take full advantage of the ground plane prior, we propose a depth-align training strategy and a precise two-stage depth inference method tailored for the ground plane prior. It is worth noting that the introduced ground plane prior requires no extra data sources like LiDAR, stereo images, and depth information. Extensive experiments on the KITTI benchmark show that our method could achieve state-of-the-art results compared with other methods while maintaining a very fast speed. Our code and models are available at https://github.com/cfzd/MonoGround.

PolyU-BPCoMa: A Dataset and Benchmark Towards Mobile Colorized Mapping Using a Backpack Multisensorial System

• Authors: Wenzhong Shi, Pengxin Chen, Muyang Wang, Sheng Bao, Haodong Xiang, Yue Yu, Daping Yang
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07468
• Pdf link: https://arxiv.org/pdf/2206.07468
• Abstract Constructing colorized point clouds from mobile laser scanning and images is a fundamental work in surveying and mapping. It is also an essential prerequisite for building digital twins for smart cities. However, existing public datasets are either in relatively small scales or lack accurate geometrical and color ground truth. This paper documents a multisensorial dataset named PolyU-BPCoMA which is distinctively positioned towards mobile colorized mapping. The dataset incorporates resources of 3D LiDAR, spherical imaging, GNSS and IMU on a backpack platform. Color checker boards are pasted in each surveyed area as targets and ground truth data are collected by an advanced terrestrial laser scanner (TLS). 3D geometrical and color information can be recovered in the colorized point clouds produced by the backpack system and the TLS, respectively. Accordingly, we provide an opportunity to benchmark the mapping and colorization accuracy simultaneously for a mobile multisensorial system. The dataset is approximately 800 GB in size covering both indoor and outdoor environments. The dataset and development kits are available at https://github.com/chenpengxin/PolyU-BPCoMa.git.

Real3D-Aug: Point Cloud Augmentation by Placing Real Objects with Occlusion Handling for 3D Detection and Segmentation

• Authors: Petr Šebek, Šimon Pokorný, Patrik Vacek, Tomáš Svoboda
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2206.07634
• Pdf link: https://arxiv.org/pdf/2206.07634
• Abstract Object detection and semantic segmentation with the 3D lidar point cloud data require expensive annotation. We propose a data augmentation method that takes advantage of already annotated data multiple times. We propose an augmentation framework that reuses real data, automatically finds suitable placements in the scene to be augmented, and handles occlusions explicitly. Due to the usage of the real data, the scan points of newly inserted objects in augmentation sustain the physical characteristics of the lidar, such as intensity and raydrop. The pipeline proves competitive in training top-performing models for 3D object detection and semantic segmentation. The new augmentation provides a significant performance gain in rare and essential classes, notably 6.65% average precision gain for "Hard" pedestrian class in KITTI object detection or 2.14 mean IoU gain in the SemanticKITTI segmentation challenge over the state of the art.

Keyword: loop detection

There is no result

Keyword: nerf

Neural Deformable Voxel Grid for Fast Optimization of Dynamic View Synthesis

• Authors: Xiang Guo, Guanying Chen, Yuchao Dai, Xiaoqing Ye, Jiadai Sun, Xiao Tan, Errui Ding
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07698
• Pdf link: https://arxiv.org/pdf/2206.07698
• Abstract Recently, Neural Radiance Fields (NeRF) is revolutionizing the task of novel view synthesis (NVS) for its superior performance. However, NeRF and its variants generally require a lengthy per-scene training procedure, where a multi-layer perceptron (MLP) is fitted to the captured images. To remedy the challenge, the voxel-grid representation has been proposed to significantly speed up the training. However, these existing methods can only deal with static scenes. How to develop an efficient and accurate dynamic view synthesis method remains an open problem. Extending the methods for static scenes to dynamic scenes is not straightforward as both the scene geometry and appearance change over time. In this paper, built on top of the recent advances in voxel-grid optimization, we propose a fast deformable radiance field method to handle dynamic scenes. Our method consists of two modules. The first module adopts a deformation grid to store 3D dynamic features, and a light-weight MLP for decoding the deformation that maps a 3D point in observation space to the canonical space using the interpolated features. The second module contains a density and a color grid to model the geometry and density of the scene. The occlusion is explicitly modeled to further improve the rendering quality. Experimental results show that our method achieves comparable performance to D-NeRF using only 20 minutes for training, which is more than 70x faster than D-NeRF, clearly demonstrating the efficiency of our proposed method.

Keyword: mapping

MonoGround: Detecting Monocular 3D Objects from the Ground

• Authors: Zequn Qin, Xi Li
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07372
• Pdf link: https://arxiv.org/pdf/2206.07372
• Abstract Monocular 3D object detection has attracted great attention for its advantages in simplicity and cost. Due to the ill-posed 2D to 3D mapping essence from the monocular imaging process, monocular 3D object detection suffers from inaccurate depth estimation and thus has poor 3D detection results. To alleviate this problem, we propose to introduce the ground plane as a prior in the monocular 3d object detection. The ground plane prior serves as an additional geometric condition to the ill-posed mapping and an extra source in depth estimation. In this way, we can get a more accurate depth estimation from the ground. Meanwhile, to take full advantage of the ground plane prior, we propose a depth-align training strategy and a precise two-stage depth inference method tailored for the ground plane prior. It is worth noting that the introduced ground plane prior requires no extra data sources like LiDAR, stereo images, and depth information. Extensive experiments on the KITTI benchmark show that our method could achieve state-of-the-art results compared with other methods while maintaining a very fast speed. Our code and models are available at https://github.com/cfzd/MonoGround.

Knowledge4COVID-19: A Semantic-based Approach for Constructing a COVID-19 related Knowledge Graph from Various Sources and Analysing Treatments' Toxicities

• Authors: Ahmad Sakor, Samaneh Jozashoori, Emetis Niazmand, Ariam Rivas, Kostantinos Bougiatiotis, Fotis Aisopos, Enrique Iglesias, Philipp D. Rohde, Trupti Padiya, Anastasia Krithara, Georgios Paliouras, Maria-Esther Vidal
• Subjects: Databases (cs.DB)
• Arxiv link: https://arxiv.org/abs/2206.07375
• Pdf link: https://arxiv.org/pdf/2206.07375
• Abstract In this paper, we present Knowledge4COVID-19, a framework that aims to showcase the power of integrating disparate sources of knowledge to discover adverse drug effects caused by drug-drug interactions among COVID-19 treatments and pre-existing condition drugs. Initially, we focus on constructing the Knowledge4COVID-19 knowledge graph (KG) from the declarative definition of mapping rules using the RDF Mapping Language. Since valuable information about drug treatments, drug-drug interactions, and side effects is present in textual descriptions in scientific databases (e.g., DrugBank) or in scientific literature (e.g., the CORD-19, the Covid-19 Open Research Dataset), the Knowledge4COVID-19 framework implements Natural Language Processing. The Knowledge4COVID-19 framework extracts relevant entities and predicates that enable the fine-grained description of COVID-19 treatments and the potential adverse events that may occur when these treatments are combined with treatments of common comorbidities, e.g., hypertension, diabetes, or asthma. Moreover, on top of the KG, several techniques for the discovery and prediction of interactions and potential adverse effects of drugs have been developed with the aim of suggesting more accurate treatments for treating the virus. We provide services to traverse the KG and visualize the effects that a group of drugs may have on a treatment outcome. Knowledge4COVID-19 was part of the Pan-European hackathon#EUvsVirus in April 2020 and is publicly available as a resource through a GitHub repository (https://github.com/SDM-TIB/Knowledge4COVID-19) and a DOI (https://zenodo.org/record/4701817#.YH336-8zbol).

Blockchain-based Federated Learning for Industrial Metaverses: Incentive Scheme with Optimal AoI

• Authors: Jiawen Kang, Dongdong Ye, Jiangtian Nie, Jiang Xiao, Xianjun Deng, Siming Wang, Zehui Xiong, Rong Yu, Dusit Niyato
• Subjects: Computer Science and Game Theory (cs.GT)
• Arxiv link: https://arxiv.org/abs/2206.07384
• Pdf link: https://arxiv.org/pdf/2206.07384
• Abstract The emerging industrial metaverses realize the mapping and expanding operations of physical industry into virtual space for significantly upgrading intelligent manufacturing. The industrial metaverses obtain data from various production and operation lines by Industrial Internet of Things (IIoT), and thus conduct effective data analysis and decision-making, thereby enhancing the production efficiency of the physical space, reducing operating costs, and maximizing commercial value. However, there still exist bottlenecks when integrating metaverses into IIoT, such as the privacy leakage of sensitive data with commercial secrets, IIoT sensing data freshness, and incentives for sharing these data. In this paper, we design a user-defined privacy-preserving framework with decentralized federated learning for the industrial metaverses. To further improve privacy protection of industrial metaverse, a cross-chain empowered federated learning framework is further utilized to perform decentralized, secure, and privacy-preserving data training on both physical and virtual spaces through a hierarchical blockchain architecture with a main chain and multiple subchains. Moreover, we introduce the age of information as the data freshness metric and thus design an age-based contract model to motivate data sensing among IIoT nodes. Numerical results indicate the efficiency of the proposed framework and incentive mechanism in the industrial metaverses

PolyU-BPCoMa: A Dataset and Benchmark Towards Mobile Colorized Mapping Using a Backpack Multisensorial System

• Authors: Wenzhong Shi, Pengxin Chen, Muyang Wang, Sheng Bao, Haodong Xiang, Yue Yu, Daping Yang
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07468
• Pdf link: https://arxiv.org/pdf/2206.07468
• Abstract Constructing colorized point clouds from mobile laser scanning and images is a fundamental work in surveying and mapping. It is also an essential prerequisite for building digital twins for smart cities. However, existing public datasets are either in relatively small scales or lack accurate geometrical and color ground truth. This paper documents a multisensorial dataset named PolyU-BPCoMA which is distinctively positioned towards mobile colorized mapping. The dataset incorporates resources of 3D LiDAR, spherical imaging, GNSS and IMU on a backpack platform. Color checker boards are pasted in each surveyed area as targets and ground truth data are collected by an advanced terrestrial laser scanner (TLS). 3D geometrical and color information can be recovered in the colorized point clouds produced by the backpack system and the TLS, respectively. Accordingly, we provide an opportunity to benchmark the mapping and colorization accuracy simultaneously for a mobile multisensorial system. The dataset is approximately 800 GB in size covering both indoor and outdoor environments. The dataset and development kits are available at https://github.com/chenpengxin/PolyU-BPCoMa.git.

Keyword: localization

Structured Video Tokens @ Ego4D PNR Temporal Localization Challenge 2022

• Authors: Elad Ben-Avraham, Roei Herzig, Karttikeya Mangalam, Amir Bar, Anna Rohrbach, Leonid Karlinsky, Trevor Darrell, Amir Globerson
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07689
• Pdf link: https://arxiv.org/pdf/2206.07689
• Abstract This technical report describes the SViT approach for the Ego4D Point of No Return (PNR) Temporal Localization Challenge. We propose a learning framework StructureViT (SViT for short), which demonstrates how utilizing the structure of a small number of images only available during training can improve a video model. SViT relies on two key insights. First, as both images and videos contain structured information, we enrich a transformer model with a set of \emph{object tokens} that can be used across images and videos. Second, the scene representations of individual frames in video should "align" with those of still images. This is achieved via a "Frame-Clip Consistency" loss, which ensures the flow of structured information between images and videos. SViT obtains strong performance on the challenge test set with 0.656 absolute temporal localization error.

LET-3D-AP: Longitudinal Error Tolerant 3D Average Precision for Camera-Only 3D Detection

• Authors: Wei-Chih Hung, Henrik Kretzschmar, Vincent Casser, Jyh-Jing Hwang, Dragomir Anguelov
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07705
• Pdf link: https://arxiv.org/pdf/2206.07705
• Abstract The popular object detection metric 3D Average Precision (3D AP) relies on the intersection over union between predicted bounding boxes and ground truth bounding boxes. However, depth estimation based on cameras has limited accuracy, which may cause otherwise reasonable predictions that suffer from such longitudinal localization errors to be treated as false positives and false negatives. We therefore propose variants of the popular 3D AP metric that are designed to be more permissive with respect to depth estimation errors. Specifically, our novel longitudinal error tolerant metrics, LET-3D-AP and LET-3D-APL, allow longitudinal localization errors of the predicted bounding boxes up to a given tolerance. The proposed metrics have been used in the Waymo Open Dataset 3D Camera-Only Detection Challenge. We believe that they will facilitate advances in the field of camera-only 3D detection by providing more informative performance signals.

Keyword: transformer

It's Time for Artistic Correspondence in Music and Video

• Authors: Didac Suris, Carl Vondrick, Bryan Russell, Justin Salamon
• Subjects: Multimedia (cs.MM); Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07148
• Pdf link: https://arxiv.org/pdf/2206.07148
• Abstract We present an approach for recommending a music track for a given video, and vice versa, based on both their temporal alignment and their correspondence at an artistic level. We propose a self-supervised approach that learns this correspondence directly from data, without any need of human annotations. In order to capture the high-level concepts that are required to solve the task, we propose modeling the long-term temporal context of both the video and the music signals, using Transformer networks for each modality. Experiments show that this approach strongly outperforms alternatives that do not exploit the temporal context. The combination of our contributions improve retrieval accuracy up to 10x over prior state of the art. This strong improvement allows us to introduce a wide range of analyses and applications. For instance, we can condition music retrieval based on visually defined attributes.

Surgical Phase Recognition in Laparoscopic Cholecystectomy

• Authors: Yunfan Li, Vinayak Shenoy, Prateek Prasanna, I.V. Ramakrishnan, Haibin Ling, Himanshu Gupta
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07198
• Pdf link: https://arxiv.org/pdf/2206.07198
• Abstract Automatic recognition of surgical phases in surgical videos is a fundamental task in surgical workflow analysis. In this report, we propose a Transformer-based method that utilizes calibrated confidence scores for a 2-stage inference pipeline, which dynamically switches between a baseline model and a separately trained transition model depending on the calibrated confidence level. Our method outperforms the baseline model on the Cholec80 dataset, and can be applied to a variety of action segmentation methods.

Born for Auto-Tagging: Faster and better with new objective functions

• Authors: Chiung-ju Liu, Huang-Ting Shieh
• Subjects: Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2206.07264
• Pdf link: https://arxiv.org/pdf/2206.07264
• Abstract Keyword extraction is a task of text mining. It is applied to increase search volume in SEO and ads. Implemented in auto-tagging, it makes tagging on a mass scale of online articles and photos efficiently and accurately. BAT is invented for auto-tagging which served as awoo's AI marketing platform (AMP). awoo AMP not only provides service as a customized recommender system but also increases the converting rate in E-commerce. The strength of BAT converges faster and better than other SOTA models, as its 4-layer structure achieves the best F scores at 50 epochs. In other words, it performs better than other models which require deeper layers at 100 epochs. To generate rich and clean tags, awoo creates new objective functions to maintain similar ${\rm F_1}$ scores with cross-entropy while enhancing ${\rm F_2}$ scores simultaneously. To assure the even better performance of F scores awoo revamps the learning rate strategy proposed by Transformer \cite{Transformer} to increase ${\rm F_1}$ and ${\rm F_2}$ scores at the same time.

Rethinking Generalization in Few-Shot Classification

• Authors: Markus Hiller, Rongkai Ma, Mehrtash Harandi, Tom Drummond
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07267
• Pdf link: https://arxiv.org/pdf/2206.07267
• Abstract Single image-level annotations only correctly describe an often small subset of an image's content, particularly when complex real-world scenes are depicted. While this might be acceptable in many classification scenarios, it poses a significant challenge for applications where the set of classes differs significantly between training and test time. In this paper, we take a closer look at the implications in the context of $\textit{few-shot learning}$. Splitting the input samples into patches and encoding these via the help of Vision Transformers allows us to establish semantic correspondences between local regions across images and independent of their respective class. The most informative patch embeddings for the task at hand are then determined as a function of the support set via online optimization at inference time, additionally providing visual interpretability of `$\textit{what matters most}$' in the image. We build on recent advances in unsupervised training of networks via masked image modelling to overcome the lack of fine-grained labels and learn the more general statistical structure of the data while avoiding negative image-level annotation influence, $\textit{aka}$ supervision collapse. Experimental results show the competitiveness of our approach, achieving new state-of-the-art results on four popular few-shot classification benchmarks for $5$-shot and $1$-shot scenarios.

Streaming non-autoregressive model for any-to-many voice conversion

• Authors: Ziyi Chen, Haoran Miao, Pengyuan Zhang
• Subjects: Sound (cs.SD); Audio and Speech Processing (eess.AS)
• Arxiv link: https://arxiv.org/abs/2206.07288
• Pdf link: https://arxiv.org/pdf/2206.07288
• Abstract Voice conversion models have developed for decades, and current mainstream research focuses on non-streaming voice conversion. However, streaming voice conversion is more suitable for practical application scenarios than non-streaming voice conversion. In this paper, we propose a streaming any-to-many voice conversion based on fully non-autoregressive model, which includes a streaming transformer based acoustic model and a streaming vocoder. Streaming transformer based acoustic model is composed of a pre-trained encoder from streaming end-to-end based automatic speech recognition model and a decoder modified on FastSpeech blocks. Streaming vocoder is designed for streaming task with pseudo quadrature mirror filter bank and causal convolution. Experimental results show that the proposed method achieves significant performance both in latency and conversion quality and can be real-time on CPU and GPU.

VCT: A Video Compression Transformer

• Authors: Fabian Mentzer, George Toderici, David Minnen, Sung-Jin Hwang, Sergi Caelles, Mario Lucic, Eirikur Agustsson
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Image and Video Processing (eess.IV)
• Arxiv link: https://arxiv.org/abs/2206.07307
• Pdf link: https://arxiv.org/pdf/2206.07307
• Abstract We show how transformers can be used to vastly simplify neural video compression. Previous methods have been relying on an increasing number of architectural biases and priors, including motion prediction and warping operations, resulting in complex models. Instead, we independently map input frames to representations and use a transformer to model their dependencies, letting it predict the distribution of future representations given the past. The resulting video compression transformer outperforms previous methods on standard video compression data sets. Experiments on synthetic data show that our model learns to handle complex motion patterns such as panning, blurring and fading purely from data. Our approach is easy to implement, and we release code to facilitate future research.

A Survey : Neural Networks for AMR-to-Text

• Authors: Hongyu Hao, Guangtong Li, Zhiming Hu, Huafeng Wang
• Subjects: Computation and Language (cs.CL); Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2206.07328
• Pdf link: https://arxiv.org/pdf/2206.07328
• Abstract AMR-to-text is one of the key techniques in the NLP community that aims at generating sentences from the Abstract Meaning Representation (AMR) graphs. Since AMR was proposed in 2013, the study on AMR-to-Text has become increasingly prevalent as an essential branch of structured data to text because of the unique advantages of AMR as a high-level semantic description of natural language. In this paper, we provide a brief survey of AMR-to-Text. Firstly, we introduce the current scenario of this technique and point out its difficulties. Secondly, based on the methods used in previous studies, we roughly divided them into five categories according to their respective mechanisms, i.e., Rules-based, Seq-to-Seq-based, Graph-to-Seq-based, Transformer-based, and Pre-trained Language Model (PLM)-based. In particular, we detail the neural network-based method and present the latest progress of AMR-to-Text, which refers to AMR reconstruction, Decoder optimization, etc. Furthermore, we present the benchmarks and evaluation methods of AMR-to-Text. Eventually, we provide a summary of current techniques and the outlook for future research.

ETMA: Efficient Transformer Based Multilevel Attention framework for Multimodal Fake News Detection

• Authors: Ashima Yadav, Shivani Gaba, Ishan Budhiraja, Neeraj Kumar
• Subjects: Multimedia (cs.MM)
• Arxiv link: https://arxiv.org/abs/2206.07331
• Pdf link: https://arxiv.org/pdf/2206.07331
• Abstract In this new digital era, social media has created a severe impact on the lives of people. In recent times, fake news content on social media has become one of the major challenging problems for society. The dissemination of fabricated and false news articles includes multimodal data in the form of text and images. The previous methods have mainly focused on unimodal analysis. Moreover, for multimodal analysis, researchers fail to keep the unique characteristics corresponding to each modality. This paper aims to overcome these limitations by proposing an Efficient Transformer based Multilevel Attention (ETMA) framework for multimodal fake news detection, which comprises the following components: visual attention-based encoder, textual attention-based encoder, and joint attention-based learning. Each component utilizes the different forms of attention mechanism and uniquely deals with multimodal data to detect fraudulent content. The efficacy of the proposed network is validated by conducting several experiments on four real-world fake news datasets: Twitter, Jruvika Fake News Dataset, Pontes Fake News Dataset, and Risdal Fake News Dataset using multiple evaluation metrics. The results show that the proposed method outperforms the baseline methods on all four datasets. Further, the computation time of the model is also lower than the state-of-the-art methods.

XMorpher: Full Transformer for Deformable Medical Image Registration via Cross Attention

• Authors: Jiacheng Shi, Yuting He, Youyong Kong, Jean-Louis Coatrieux, Huazhong Shu, Guanyu Yang, Shuo Li
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07349
• Pdf link: https://arxiv.org/pdf/2206.07349
• Abstract An effective backbone network is important to deep learning-based Deformable Medical Image Registration (DMIR), because it extracts and matches the features between two images to discover the mutual correspondence for fine registration. However, the existing deep networks focus on single image situation and are limited in registration task which is performed on paired images. Therefore, we advance a novel backbone network, XMorpher, for the effective corresponding feature representation in DMIR. 1) It proposes a novel full transformer architecture including dual parallel feature extraction networks which exchange information through cross attention, thus discovering multi-level semantic correspondence while extracting respective features gradually for final effective registration. 2) It advances the Cross Attention Transformer (CAT) blocks to establish the attention mechanism between images which is able to find the correspondence automatically and prompts the features to fuse efficiently in the network. 3) It constrains the attention computation between base windows and searching windows with different sizes, and thus focuses on the local transformation of deformable registration and enhances the computing efficiency at the same time. Without any bells and whistles, our XMorpher gives Voxelmorph 2.8% improvement on DSC , demonstrating its effective representation of the features from the paired images in DMIR. We believe that our XMorpher has great application potential in more paired medical images. Our XMorpher is open on https://github.com/Solemoon/XMorpher

RecBole 2.0: Towards a More Up-to-Date Recommendation Library

• Authors: Wayne Xin Zhao, Yupeng Hou, Xingyu Pan, Chen Yang, Zeyu Zhang, Zihan Lin, Jingsen Zhang, Shuqing Bian, Jiakai Tang, Wenqi Sun, Yushuo Chen, Lanling Xu, Gaowei Zhang, Zhen Tian, Changxin Tian, Shanlei Mu, Xinyan Fan, Xu Chen, Ji-Rong Wen
• Subjects: Information Retrieval (cs.IR)
• Arxiv link: https://arxiv.org/abs/2206.07351
• Pdf link: https://arxiv.org/pdf/2206.07351
• Abstract In order to support the study of recent advances in recommender systems, this paper presents an extended recommendation library consisting of eight packages for up-to-date topics and architectures. First of all, from a data perspective, we consider three important topics related to data issues (i.e., sparsity, bias and distribution shift), and develop five packages accordingly: meta-learning, data augmentation, debiasing, fairness and cross-domain recommendation. Furthermore, from a model perspective, we develop two benchmarking packages for Transformer-based and graph neural network (GNN)-based models, respectively. All the packages (consisting of 65 new models) are developed based on a popular recommendation framework RecBole, ensuring that both the implementation and interface are unified. For each package, we provide complete implementations from data loading, experimental setup, evaluation and algorithm implementation. This library provides a valuable resource to facilitate the up-to-date research in recommender systems. The project is released at the link: https://github.com/RUCAIBox/RecBole2.0.

NatiQ: An End-to-end Text-to-Speech System for Arabic

• Authors: Ahmed Abdelali, Nadir Durrani, Cenk Demiroglu, Fahim Dalvi, Hamdy Mubarak, Kareem Darwish
• Subjects: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
• Arxiv link: https://arxiv.org/abs/2206.07373
• Pdf link: https://arxiv.org/pdf/2206.07373
• Abstract NatiQ is end-to-end text-to-speech system for Arabic. Our speech synthesizer uses an encoder-decoder architecture with attention. We used both tacotron-based models (tacotron-1 and tacotron-2) and the faster transformer model for generating mel-spectrograms from characters. We concatenated Tacotron1 with the WaveRNN vocoder, Tacotron2 with the WaveGlow vocoder and ESPnet transformer with the parallel wavegan vocoder to synthesize waveforms from the spectrograms. We used in-house speech data for two voices: 1) neutral male "Hamza"- narrating general content and news, and 2) expressive female "Amina"- narrating children story books to train our models. Our best systems achieve an average Mean Opinion Score (MOS) of 4.21 and 4.40 for Amina and Hamza respectively. The objective evaluation of the systems using word and character error rate (WER and CER) as well as the response time measured by real-time factor favored the end-to-end architecture ESPnet. NatiQ demo is available on-line at https://tts.qcri.org

Estimating Confidence of Predictions of Individual Classifiers and Their Ensembles for the Genre Classification Task

• Authors: Mikhail Lepekhin, Serge Sharoff
• Subjects: Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2206.07427
• Pdf link: https://arxiv.org/pdf/2206.07427
• Abstract Genre identification is a subclass of non-topical text classification. The main difference between this task and topical classification is that genres, unlike topics, usually do not correspond to simple keywords, and thus they need to be defined in terms of their functions in communication. Neural models based on pre-trained transformers, such as BERT or XLM-RoBERTa, demonstrate SOTA results in many NLP tasks, including non-topical classification. However, in many cases, their downstream application to very large corpora, such as those extracted from social media, can lead to unreliable results because of dataset shifts, when some raw texts do not match the profile of the training set. To mitigate this problem, we experiment with individual models as well as with their ensembles. To evaluate the robustness of all models we use a prediction confidence metric, which estimates the reliability of a prediction in the absence of a gold standard label. We can evaluate robustness via the confidence gap between the correctly classified texts and the misclassified ones on a labeled test corpus, higher gaps make it easier to improve our confidence that our classifier made the right decision. Our results show that for all of the classifiers tested in this study, there is a confidence gap, but for the ensembles, the gap is bigger, meaning that ensembles are more robust than their individual models.

Forecasting of depth and ego-motion with transformers and self-supervision

• Authors: Houssem Boulahbal, Adrian Voicila, Andrew Comport
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07435
• Pdf link: https://arxiv.org/pdf/2206.07435
• Abstract This paper addresses the problem of end-to-end self-supervised forecasting of depth and ego motion. Given a sequence of raw images, the aim is to forecast both the geometry and ego-motion using a self supervised photometric loss. The architecture is designed using both convolution and transformer modules. This leverages the benefits of both modules: Inductive bias of CNN, and the multi-head attention of transformers, thus enabling a rich spatio-temporal representation that enables accurate depth forecasting. Prior work attempts to solve this problem using multi-modal input/output with supervised ground-truth data which is not practical since a large annotated dataset is required. Alternatively to prior methods, this paper forecasts depth and ego motion using only self-supervised raw images as input. The approach performs significantly well on the KITTI dataset benchmark with several performance criteria being even comparable to prior non-forecasting self-supervised monocular depth inference methods.

AMR Alignment: Paying Attention to Cross-Attention

• Authors: Pere-Lluís Huguet Cabot, Abelardo Carlos Martínez Lorenzo, Roberto Navigli
• Subjects: Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2206.07587
• Pdf link: https://arxiv.org/pdf/2206.07587
• Abstract With the surge of Transformer models, many have investigated how attention acts on the learned representations. However, attention is still overlooked for specific tasks, such as Semantic Parsing. A popular approach to the formal representation of a sentence's meaning is Abstract Meaning Representation (AMR). Until now, the alignment between a sentence and its AMR representation has been explored in different ways, such as through rules or via the Expectation Maximization (EM) algorithm. In this paper, we investigate the ability of Transformer-based parsing models to yield effective alignments without ad-hoc strategies. We present the first in-depth exploration of cross-attention for AMR by proxy of alignment between the sentence spans and the semantic units in the graph. We show how current Transformer-based parsers implicitly encode the alignment information in the cross-attention weights and how to leverage it to extract such alignment. Furthermore, we supervise and guide cross-attention using alignment, dropping the need for English- and AMR-specific rules.

Exploring Capabilities of Monolingual Audio Transformers using Large Datasets in Automatic Speech Recognition of Czech

• Authors: Jan Lehečka, Jan Švec, Aleš Pražák, Josef V. Psutka
• Subjects: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
• Arxiv link: https://arxiv.org/abs/2206.07627
• Pdf link: https://arxiv.org/pdf/2206.07627
• Abstract In this paper, we present our progress in pretraining Czech monolingual audio transformers from a large dataset containing more than 80 thousand hours of unlabeled speech, and subsequently fine-tuning the model on automatic speech recognition tasks using a combination of in-domain data and almost 6 thousand hours of out-of-domain transcribed speech. We are presenting a large palette of experiments with various fine-tuning setups evaluated on two public datasets (CommonVoice and VoxPopuli) and one extremely challenging dataset from the MALACH project. Our results show that monolingual Wav2Vec 2.0 models are robust ASR systems, which can take advantage of large labeled and unlabeled datasets and successfully compete with state-of-the-art LVCSR systems. Moreover, Wav2Vec models proved to be good zero-shot learners when no training data are available for the target ASR task.

Coarse-to-Fine Vision-Language Pre-training with Fusion in the Backbone

• Authors: Zi-Yi Dou, Aishwarya Kamath, Zhe Gan, Pengchuan Zhang, Jianfeng Wang, Linjie Li, Zicheng Liu, Ce Liu, Yann LeCun, Nanyun Peng, Jianfeng Gao, Lijuan Wang
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2206.07643
• Pdf link: https://arxiv.org/pdf/2206.07643
• Abstract Vision-language (VL) pre-training has recently received considerable attention. However, most existing end-to-end pre-training approaches either only aim to tackle VL tasks such as image-text retrieval, visual question answering (VQA) and image captioning that test high-level understanding of images, or only target region-level understanding for tasks such as phrase grounding and object detection. We present FIBER (Fusion-In-the-Backbone-based transformER), a new VL model architecture that can seamlessly handle both these types of tasks. Instead of having dedicated transformer layers for fusion after the uni-modal backbones, FIBER pushes multimodal fusion deep into the model by inserting cross-attention into the image and text backbones, bringing gains in terms of memory and performance. In addition, unlike previous work that is either only pre-trained on image-text data or on fine-grained data with box-level annotations, we present a two-stage pre-training strategy that uses both these kinds of data efficiently: (i) coarse-grained pre-training based on image-text data; followed by (ii) fine-grained pre-training based on image-text-box data. We conduct comprehensive experiments on a wide range of VL tasks, ranging from VQA, image captioning, and retrieval, to phrase grounding, referring expression comprehension, and object detection. Using deep multimodal fusion coupled with the two-stage pre-training, FIBER provides consistent performance improvements over strong baselines across all tasks, often outperforming methods using magnitudes more data. Code is available at https://github.com/microsoft/FIBER.

SP-ViT: Learning 2D Spatial Priors for Vision Transformers

• Authors: Yuxuan Zhou, Wangmeng Xiang, Chao Li, Biao Wang, Xihan Wei, Lei Zhang, Margret Keuper, Xiansheng Hua
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07662
• Pdf link: https://arxiv.org/pdf/2206.07662
• Abstract Recently, transformers have shown great potential in image classification and established state-of-the-art results on the ImageNet benchmark. However, compared to CNNs, transformers converge slowly and are prone to overfitting in low-data regimes due to the lack of spatial inductive biases. Such spatial inductive biases can be especially beneficial since the 2D structure of an input image is not well preserved in transformers. In this work, we present Spatial Prior-enhanced Self-Attention (SP-SA), a novel variant of vanilla Self-Attention (SA) tailored for vision transformers. Spatial Priors (SPs) are our proposed family of inductive biases that highlight certain groups of spatial relations. Unlike convolutional inductive biases, which are forced to focus exclusively on hard-coded local regions, our proposed SPs are learned by the model itself and take a variety of spatial relations into account. Specifically, the attention score is calculated with emphasis on certain kinds of spatial relations at each head, and such learned spatial foci can be complementary to each other. Based on SP-SA we propose the SP-ViT family, which consistently outperforms other ViT models with similar GFlops or parameters. Our largest model SP-ViT-L achieves a record-breaking 86.3% Top-1 accuracy with a reduction in the number of parameters by almost 50% compared to previous state-of-the-art model (150M for SP-ViT-L vs 271M for CaiT-M-36) among all ImageNet-1K models trained on 224x224 and fine-tuned on 384x384 resolution w/o extra data.

Transformer-based Automatic Speech Recognition of Formal and Colloquial Czech in MALACH Project

• Authors: Jan Lehečka, Josef V. Psutka, Josef Psutka
• Subjects: Computation and Language (cs.CL)
• Arxiv link: https://arxiv.org/abs/2206.07666
• Pdf link: https://arxiv.org/pdf/2206.07666
• Abstract Czech is a very specific language due to its large differences between the formal and the colloquial form of speech. While the formal (written) form is used mainly in official documents, literature, and public speeches, the colloquial (spoken) form is used widely among people in casual speeches. This gap introduces serious problems for ASR systems, especially when training or evaluating ASR models on datasets containing a lot of colloquial speech, such as the MALACH project. In this paper, we are addressing this problem in the light of a new paradigm in end-to-end ASR systems -- recently introduced self-supervised audio Transformers. Specifically, we are investigating the influence of colloquial speech on the performance of Wav2Vec 2.0 models and their ability to transcribe colloquial speech directly into formal transcripts. We are presenting results with both formal and colloquial forms in the training transcripts, language models, and evaluation transcripts.

AVATAR: Unconstrained Audiovisual Speech Recognition

• Authors: Valentin Gabeur, Paul Hongsuck Seo, Arsha Nagrani, Chen Sun, Karteek Alahari, Cordelia Schmid
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM); Sound (cs.SD); Audio and Speech Processing (eess.AS)
• Arxiv link: https://arxiv.org/abs/2206.07684
• Pdf link: https://arxiv.org/pdf/2206.07684
• Abstract Audio-visual automatic speech recognition (AV-ASR) is an extension of ASR that incorporates visual cues, often from the movements of a speaker's mouth. Unlike works that simply focus on the lip motion, we investigate the contribution of entire visual frames (visual actions, objects, background etc.). This is particularly useful for unconstrained videos, where the speaker is not necessarily visible. To solve this task, we propose a new sequence-to-sequence AudioVisual ASR TrAnsformeR (AVATAR) which is trained end-to-end from spectrograms and full-frame RGB. To prevent the audio stream from dominating training, we propose different word-masking strategies, thereby encouraging our model to pay attention to the visual stream. We demonstrate the contribution of the visual modality on the How2 AV-ASR benchmark, especially in the presence of simulated noise, and show that our model outperforms all other prior work by a large margin. Finally, we also create a new, real-world test bed for AV-ASR called VisSpeech, which demonstrates the contribution of the visual modality under challenging audio conditions.

Structured Video Tokens @ Ego4D PNR Temporal Localization Challenge 2022

• Authors: Elad Ben-Avraham, Roei Herzig, Karttikeya Mangalam, Amir Bar, Anna Rohrbach, Leonid Karlinsky, Trevor Darrell, Amir Globerson
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07689
• Pdf link: https://arxiv.org/pdf/2206.07689
• Abstract This technical report describes the SViT approach for the Ego4D Point of No Return (PNR) Temporal Localization Challenge. We propose a learning framework StructureViT (SViT for short), which demonstrates how utilizing the structure of a small number of images only available during training can improve a video model. SViT relies on two key insights. First, as both images and videos contain structured information, we enrich a transformer model with a set of \emph{object tokens} that can be used across images and videos. Second, the scene representations of individual frames in video should "align" with those of still images. This is achieved via a "Frame-Clip Consistency" loss, which ensures the flow of structured information between images and videos. SViT obtains strong performance on the challenge test set with 0.656 absolute temporal localization error.

A Simple Data Mixing Prior for Improving Self-Supervised Learning

• Authors: Sucheng Ren, Huiyu Wang, Zhengqi Gao, Shengfeng He, Alan Yuille, Yuyin Zhou, Cihang Xie
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07692
• Pdf link: https://arxiv.org/pdf/2206.07692
• Abstract Data mixing (e.g., Mixup, Cutmix, ResizeMix) is an essential component for advancing recognition models. In this paper, we focus on studying its effectiveness in the self-supervised setting. By noticing the mixed images that share the same source images are intrinsically related to each other, we hereby propose SDMP, short for $\textbf{S}$imple $\textbf{D}$ata $\textbf{M}$ixing $\textbf{P}$rior, to capture this straightforward yet essential prior, and position such mixed images as additional $\textbf{positive pairs}$ to facilitate self-supervised representation learning. Our experiments verify that the proposed SDMP enables data mixing to help a set of self-supervised learning frameworks (e.g., MoCo) achieve better accuracy and out-of-distribution robustness. More notably, our SDMP is the first method that successfully leverages data mixing to improve (rather than hurt) the performance of Vision Transformers in the self-supervised setting. Code is publicly available at https://github.com/OliverRensu/SDMP

Masked Siamese ConvNets

• Authors: Li Jing, Jiachen Zhu, Yann LeCun
• Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
• Arxiv link: https://arxiv.org/abs/2206.07700
• Pdf link: https://arxiv.org/pdf/2206.07700
• Abstract Self-supervised learning has shown superior performances over supervised methods on various vision benchmarks. The siamese network, which encourages embeddings to be invariant to distortions, is one of the most successful self-supervised visual representation learning approaches. Among all the augmentation methods, masking is the most general and straightforward method that has the potential to be applied to all kinds of input and requires the least amount of domain knowledge. However, masked siamese networks require particular inductive bias and practically only work well with Vision Transformers. This work empirically studies the problems behind masked siamese networks with ConvNets. We propose several empirical designs to overcome these problems gradually. Our method performs competitively on low-shot image classification and outperforms previous methods on object detection benchmarks. We discuss several remaining issues and hope this work can provide useful data points for future general-purpose self-supervised learning.

Keyword: autonomous driving

Mean-Semivariance Policy Optimization via Risk-Averse Reinforcement Learning

• Authors: Xiaoteng Ma, Shuai Ma, Li Xia, Qianchuan Zhao
• Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI)
• Arxiv link: https://arxiv.org/abs/2206.07376
• Pdf link: https://arxiv.org/pdf/2206.07376
• Abstract Keeping risk under control is often more crucial than maximizing expected reward in real-world decision-making situations, such as finance, robotics, autonomous driving, etc. The most natural choice of risk measures is variance, while it penalizes the upside volatility as much as the downside part. Instead, the (downside) semivariance, which captures negative deviation of a random variable under its mean, is more suitable for risk-averse proposes. This paper aims at optimizing the mean-semivariance (MSV) criterion in reinforcement learning w.r.t. steady rewards. Since semivariance is time-inconsistent and does not satisfy the standard Bellman equation, the traditional dynamic programming methods are inapplicable to MSV problems directly. To tackle this challenge, we resort to the Perturbation Analysis (PA) theory and establish the performance difference formula for MSV. We reveal that the MSV problem can be solved by iteratively solving a sequence of RL problems with a policy-dependent reward function. Further, we propose two on-policy algorithms based on the policy gradient theory and the trust region method. Finally, we conduct diverse experiments from simple bandit problems to continuous control tasks in MuJoCo, which demonstrate the effectiveness of our proposed methods.

AI Ethics Issues in Real World: Evidence from AI Incident Database

• Authors: Mengyi Wei, Zhixuan Zhou
• Subjects: Artificial Intelligence (cs.AI); Computers and Society (cs.CY)
• Arxiv link: https://arxiv.org/abs/2206.07635
• Pdf link: https://arxiv.org/pdf/2206.07635
• Abstract With the powerful performance of Artificial Intelligence (AI) also comes prevalent ethical issues. Though governments and corporations have curated multiple AI ethics guidelines to curb unethical behavior of AI, the effect has been limited, probably due to the vagueness of the guidelines. In this paper, we take a closer look at how AI ethics issues take place in real world, in order to have a more in-depth and nuanced understanding of different ethical issues as well as their social impact. With a content analysis of AI Incident Database, which is an effort to prevent repeated real world AI failures by cataloging incidents, we identified 13 application areas which often see unethical use of AI, with intelligent service robots, language/vision models and autonomous driving taking the lead. Ethical issues appear in 8 different forms, from inappropriate use and racial discrimination, to physical safety and unfair algorithm. With this taxonomy of AI ethics issues, we aim to provide AI practitioners with a practical guideline when trying to deploy AI applications ethically.

Waymo Open Dataset: Panoramic Video Panoptic Segmentation

• Authors: Jieru Mei, Alex Zihao Zhu, Xinchen Yan, Hang Yan, Siyuan Qiao, Yukun Zhu, Liang-Chieh Chen, Henrik Kretzschmar, Dragomir Anguelov
• Subjects: Computer Vision and Pattern Recognition (cs.CV)
• Arxiv link: https://arxiv.org/abs/2206.07704
• Pdf link: https://arxiv.org/pdf/2206.07704
• Abstract Panoptic image segmentation is the computer vision task of finding groups of pixels in an image and assigning semantic classes and object instance identifiers to them. Research in image segmentation has become increasingly popular due to its critical applications in robotics and autonomous driving. The research community thereby relies on publicly available benchmark dataset to advance the state-of-the-art in computer vision. Due to the high costs of densely labeling the images, however, there is a shortage of publicly available ground truth labels that are suitable for panoptic segmentation. The high labeling costs also make it challenging to extend existing datasets to the video domain and to multi-camera setups. We therefore present the Waymo Open Dataset: Panoramic Video Panoptic Segmentation Dataset, a large-scale dataset that offers high-quality panoptic segmentation labels for autonomous driving. We generate our dataset using the publicly available Waymo Open Dataset, leveraging the diverse set of camera images. Our labels are consistent over time for video processing and consistent across multiple cameras mounted on the vehicles for full panoramic scene understanding. Specifically, we offer labels for 28 semantic categories and 2,860 temporal sequences that were captured by five cameras mounted on autonomous vehicles driving in three different geographical locations, leading to a total of 100k labeled camera images. To the best of our knowledge, this makes our dataset an order of magnitude larger than existing datasets that offer video panoptic segmentation labels. We further propose a new benchmark for Panoramic Video Panoptic Segmentation and establish a number of strong baselines based on the DeepLab family of models. We will make the benchmark and the code publicly available. Find the dataset at https://waymo.com/open.