Sheng Liu

Affiliations

New York University

papers

1,597

total citations

papers (16)

Early-Learning Regularization Prevents Memorization of Noisy Labels

NEURIPS 2020arXiv

681

citations

In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steering

ICML 2024arXiv

224

citations

Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews

ICML 2024arXiv

183

citations

Adaptive Early-Learning Correction for Segmentation From Noisy Annotations

CVPR 2022arXiv

121

citations

Are All Losses Created Equal: A Neural Collapse Perspective

NEURIPS 2022arXiv

citations

MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine

ICLR 2025arXiv

citations

On Learning Contrastive Representations for Learning With Noisy Labels

CVPR 2022arXiv

citations

Multiple Instance Learning via Iterative Self-Paced Supervised Contrastive Learning

CVPR 2023arXiv

citations

Reducing Hallucinations in Large Vision-Language Models via Latent Space Steering

ICLR 2025

citations

Convolutional Normalization: Improving Deep Convolutional Network Robustness and Training

NEURIPS 2021arXiv

citations

PADDLES: Phase-Amplitude Spectrum Disentangled Early Stopping for Learning with Noisy Labels

ICCV 2023arXiv

citations

OTLRM: Orthogonal Learning-based Low-Rank Metric for Multi-Dimensional Inverse Problems

AAAI 2025arXiv

citations

OLMD: Orientation-aware Long-term Motion Decoupling for Continuous Sign Language Recognition

AAAI 2025arXiv

citations

Sheng Liu

Affiliations

papers (16)

Early-Learning Regularization Prevents Memorization of Noisy Labels

In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steering

Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews

Adaptive Early-Learning Correction for Segmentation From Noisy Annotations

Are All Losses Created Equal: A Neural Collapse Perspective

MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine

On Learning Contrastive Representations for Learning With Noisy Labels

Multiple Instance Learning via Iterative Self-Paced Supervised Contrastive Learning

Reducing Hallucinations in Large Vision-Language Models via Latent Space Steering

Convolutional Normalization: Improving Deep Convolutional Network Robustness and Training

LEMaRT: Label-Efficient Masked Region Transform for Image Harmonization

Video Shadow Detection via Spatio-Temporal Interpolation Consistency Training

Depth-Guided Sparse Structure-From-Motion for Movies and TV Shows

PADDLES: Phase-Amplitude Spectrum Disentangled Early Stopping for Learning with Noisy Labels

OTLRM: Orthogonal Learning-based Low-Rank Metric for Multi-Dimensional Inverse Problems

OLMD: Orientation-aware Long-term Motion Decoupling for Continuous Sign Language Recognition

papers (16)

Early-Learning Regularization Prevents Memorization of Noisy Labels

In-context Vectors: Making In Context Learning More Effective and Controllable Through Latent Space Steering

Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviews

Adaptive Early-Learning Correction for Segmentation From Noisy Annotations

Are All Losses Created Equal: A Neural Collapse Perspective

MedTrinity-25M: A Large-scale Multimodal Dataset with Multigranular Annotations for Medicine

On Learning Contrastive Representations for Learning With Noisy Labels

Multiple Instance Learning via Iterative Self-Paced Supervised Contrastive Learning

Reducing Hallucinations in Large Vision-Language Models via Latent Space Steering

Convolutional Normalization: Improving Deep Convolutional Network Robustness and Training

LEMaRT: Label-Efficient Masked Region Transform for Image Harmonization

Video Shadow Detection via Spatio-Temporal Interpolation Consistency Training

Depth-Guided Sparse Structure-From-Motion for Movies and TV Shows

PADDLES: Phase-Amplitude Spectrum Disentangled Early Stopping for Learning with Noisy Labels

OTLRM: Orthogonal Learning-based Low-Rank Metric for Multi-Dimensional Inverse Problems

OLMD: Orientation-aware Long-term Motion Decoupling for Continuous Sign Language Recognition