Academic paper review is a critical yet time-consuming task within the research community. With the increasing volume of academic publications, automating the review process has become a significant challenge. The primary issue lies in generating comprehensive, accurate, and reasoning-consistent review comments that align with human reviewers' judgments. In this paper, we address this challenge by proposing ReviewAgents, a framework that leverages large language models (LLMs) to generate academic paper reviews. We first introduce a novel dataset, Review-CoT, consisting of 142k review comments, designed for training LLM agents. This dataset emulates the structured reasoning process of human reviewers-summarizing the paper, referencing relevant works, identifying strengths and weaknesses, and generating a review conclusion. Building upon this, we train LLM reviewer agents capable of structured reasoning using a relevant-paper-aware training method. Furthermore, we construct ReviewAgents, a multi-role, multi-LLM agent review framework, to enhance the review comment generation process. Additionally, we propose ReviewBench, a benchmark for evaluating the review comments generated by LLMs. Our experimental results on ReviewBench demonstrate that while existing LLMs exhibit a certain degree of potential for automating the review process, there remains a gap when compared to human-generated reviews. Moreover, our ReviewAgents framework further narrows this gap, outperforming advanced LLMs in generating review comments.

Substantial quantity and high quality are the golden rules of making a good training dataset with sample privacy protection equally important. Generating synthetic samples that resemble high-quality private data while ensuring Differential Privacy (DP), a formal privacy guarantee, promises scalability and practicality. However, existing methods relying on pre-trained models for data synthesis %that avoid fine-tuning large pre-trained generative models often struggle in data-deficient scenarios, suffering from limited sample size, inevitable generation noise and existing pre-trained model bias. To address these challenges, we propose a novel contrAstive private data Synthesis via Weighted multiple Pre-trained language models (PLM) framework, named as WASP. WASP utilizes limited private samples for more accurate private data distribution estimation via a Top-Q voting mechanism, and leverages low-quality synthetic samples for contrastive generation via collaboration among dynamically weighted multiple pre-trained models.Extensive experiments on 6 well-developed datasets with 6 open-source and 3 closed-source PLMs demonstrate the superiority of WASP in improving model performance over diverse downstream tasks. Code is available at https://anonymous.4open.science/r/WASP.

Continual learning involves training neural networks incrementally for new tasks while retaining the knowledge of previous tasks. However, efficiently fine-tuning the model for sequential tasks with minimal computational resources remains a challenge. In this paper, we propose Task Incremental Continual Learning (TI-CL) of audio classifiers with both parameterefficient and compute-efficient Audio Spectrogram Transformers (AST). To reduce the trainable parameters without performance degradation for TI-CL, we compare several Parameter Efficient Transfer (PET) methods and propose AST with Convolutional Adapters for TI-CL, which has less than 5% of trainable parameters of the fully fine-tuned counterparts. To reduce the computational complexity, we introduce a novel FrequencyTime factorized Attention (FTA) method that replaces the traditional self-attention in transformers for audio spectrograms.

Recently Transformer and Convolution neural network (CNN) based models have shown promising results in Automatic Speech Recognition (ASR), outperforming Recurrent neural networks (RNNs). Transformer models are good at capturing content-based global interactions, while CNNs exploit local features effectively. In this work, we achieve the best of both worlds by studying how to combine convolution neural networks and transformers to model both local and global dependencies of an audio sequence in a parameter-efficient way. To this regard, we propose the convolution-augmented transformer for speech recognition, named Conformer. Conformer significantly outperforms the previous Transformer and CNN based models achieving state-of-the-art accuracies.

This paper is a review of the state-of-the-art emphasizing convolutional neural nets (CNNs) and recurrent neural nets (RNNs) and their applications to key domains including vision, speech recognition, and more. This figure explains how information flows through a CNN on the forward pass. First filters are convolved with sections of the image and regularized with ReLu. This information is max pooled, which selects the maximum from a set of values. This process is repeated for multiple layers.

The field of natural language processing (NLP) has been transformed by massive pre-trained language models. They form the basis of all state-of-the-art systems across a wide range of tasks and have shown an impressive ability to generate fluent text and perform few-shot learning. At the same time, these models are hard to understand and give rise to new ethical and scalability challenges. In this course, students will learn the fundamentals about the modeling, theory, ethics, and systems aspects of large language models, as well as gain hands-on experience working with them. Where: Class will by default be in person at 200-002 (History Corner).

OpenAI's ChatGPT is the new cool AI in town and has taken the world by storm. We've all seen countless Twitter threads, medium articles, etc., that highlight the different ways ChatGPT can be used. Some developers have already started to build applications, plugins, services, etc., that leverage ChatGPT. While the exact workings of ChatGPT aren't yet known since OpenAI hasn't released a paper or open-sourced their code yet. We trained this model using Reinforcement Learning from Human Feedback (RLHF), using the same methods as InstructGPT, but with slight differences in the data collection setup.

Original Abstract This paper presents an end-to-end semi-supervised object detection approach, in contrast to previous more complex multi-stage methods. The end-to-end training gradually improves pseudo-label qualities during the curriculum and the more and more accurate pseudo-labels in turn benefit object detection training. We also propose two simple yet effective techniques within this framework: a soft teacher mechanism where the classification loss of each unlabeled bounding box is weighed by the classification score produced by the teacher network; a box jittering approach to select reliable pseudo boxes for the learning of box regression. On the COCO benchmark, the proposed approach outperforms previous methods by a large margin under various labelling ratios, i.e. 1%, 5% and 10%. Moreover, our approach proves to perform also well when the amount of labelled data is relatively large.

When browsing through new data science papers, from time to time you encounter clever new ideas. And even though they sometimes aren't broadly adopted yet, they can yield great potential. I believe the paper I'll talk about today is one of those papers: "Emergent Complexity and Zero-shot Transfer via Unsupervised Environment Design" by Michael Dennis, Natasha Jaques et al. from the Google Brain team. The title is a mouthful, but in short, the paper talks about an automated way to generate increasingly challenging environments for RL models. You can find the paper here.

The idea behind this paper is that we often don't understand why a particular network works and others don't. What are the hyperparameters that should be tuned? What are the things that need to be changed in order to improve the accuracy (without modifying the architecture)? This paper investigates several approaches and develops insight on why ConvNets are still the king in computer vision? So, without further ado let's jump right into this amazing paper.