Yi-Chen (Howard) Lo comments

Results 89 comments of


                                            Yi-Chen (Howard) Lo

How to Read a Paper

### [NeurIPS 2019 Reviewer Guildelines](https://nips.cc/Conferences/2019/PaperInformation/ReviewerGuidelines#Examples) I found that maybe try to follow this guideline to summarize the paper can be helpful for developing the research taste.

On Accurate Evaluation of GANs for Language Generation

### Summary This paper presents a systematic re-evaluation for a number of recently purposed GAN-based text generation models. They argued that N-gram based metrics (e.g., BLEU, Self-BLEU) is not sensitive...

On Accurate Evaluation of GANs for Language Generation

@xpqiu Hi Xipeng Actually, your paper is currently in my paper reading pending list! Ha! I will soon get to your paper and write a summary about it. It would...

When and Why are Pre-trained Word Embeddings Useful for Neural Machine Translation?

### Summary 1. Pre-training the word embeddings in the source and/or target languages helps to increase BLEU scores. 2. Pre-training source language embeddings gains much improvement, indicating that better encoding...

Neural-Symbolic VQA: Disentangling Reasoning from Vision and Language Understanding

### Summary ![](http://nsvqa.csail.mit.edu/images/model.jpg) Neural-symbolic VQA (NS-VQA) model has three components: - A scene parser (de-renderer) that segments an input image (a-b) and recovers a structural scene representation (c). - A...

Reaching Human-Level Performance in Automatic Grammatical Error Correction: An Empirical Study

### Summary - Present fluency boost learning and inference mechanism for seq2seq model. (See figure 2) - Fluency boost learning: Generates more error-corrected sentence pairs from model's n-best list, enabling...

A Simple Neural Network Module For Relational Reasoning

### Summary This paper purposes a Relation Networks (RN) to solve relational problems, and achieves SOTA on VQA challenge on CLEVER dataset, text-based QA on bAbI dataset, and complex reasoning...

‘Lighter’ Can Still Be Dark: Modeling Comparative Color Descriptions

### Summary This paper purposes a new paradigm of learning to ground comparative adjectives within the realm of color description: given a reference RGB color and a comparative term (e.g.,...

Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning

### Model uncertainty Model or epistemic uncertainty captures uncertainty in the model parameters. It is higher in regions of no or little training data and lower in regions of more...

Deep Reinforcement Learning with a Natural Language Action Space

### Summary This paper purposes a *Deep Reinforcement Relevance Network* (DQN-based) learn to understand the texts from both text-based state and action (learn the "relevance" between state and action) via...