Yufeng Zhang

Generative Adversarial Imitation Learning with Neural Networks: Global Optimality and Convergence Rate

2020 · arXiv (Cornell University)

Generative adversarial imitation learning (GAIL) demonstrates tremendous success in practice, especially when combined with neural networks. Different from reinforcement learning, GAIL learns both policy and reward function from expert (human) demonstration. Despite its empirical success, …

Kullback-Leibler Divergence-Based Out-of-Distribution Detection with Flow-Based Generative Models

2020 · arXiv (Cornell University)

Recent research has revealed that deep generative models including flow-based models and Variational Autoencoders may assign higher likelihoods to out-of-distribution (OOD) data than in-distribution (ID) data. However, we cannot sample OOD data from the model. …

Verification of message-passing uninterpreted programs

2024 · Science of Computer Programming

Message-passing programs involve several processes with channel-based communications to deal with tasks concurrently. The complex computations and communications between processes make the verification of message-passing programs hard. By regarding the functions in programs as uninterpreted …

$\mathbf{(N,K)}$-Puzzle: A Cost-Efficient Testbed for Benchmarking Reinforcement Learning Algorithms in Generative Language Model

2024 · arXiv (Cornell University)

Recent advances in reinforcement learning (RL) algorithms aim to enhance the performance of language models at scale. Yet, there is a noticeable absence of a cost-effective and standardized testbed tailored to evaluating and comparing these …

Image Coding and Compression Method Based on Semantics of Pixels for Semantic Communication

2024

Semantic image transmission is promising for both image reconstruction or downstream task execution. However, it is still challenging to simultaneously achieve pleasing visual clarity and superior downstream tasks. To tackle this issue, we propose a …

Seed-CTS: Unleashing the Power of Tree Search for Superior Performance in Competitive Coding Tasks

2024 · arXiv (Cornell University)

Competition-level code generation tasks pose significant challenges for current state-of-the-art large language models (LLMs). For example, on the LiveCodeBench-Hard dataset, models such as O1-Mini and O1-Preview achieve pass@1 rates of only 0.366 and 0.143, respectively. …

Multi-Granularity Content-Aware Network with Semantic Integration for Unsupervised Anomaly Detection

2025 · Applied Sciences

Unsupervised anomaly detection has been widely applied to industrial scenarios. Recently, transformer-based methods have also been developed and have produced good performance. Although the global dependencies in anomaly images are considered, the typical patch partition …