Causal-LLaVA: Causal Disentanglement for Mitigating Hallucination in Multimodal Large Language Models
机构 * Zhejiang University（浙江大学）
Comments 21 pages, 19 figures, Submitted to NeurIPS 2025
期刊&会议
NeurIPS
Conference on Neural Information Processing Systems · 会议 · Machine Learning
Are Vision Language Models Ready for Clinical Diagnosis? A 3D Medical Benchmark for Tumor-centric Visual Question Answering
Comments NeurIPS 2025 datasets&benchmarks track submission
KerZOO: Kernel Function Informed Zeroth-Order Optimization for Accurate and Accelerated LLM Fine-Tuning
机构 * Stevens Institute of Technology（史蒂文斯理工学院） ； University of Georgia（佐治亚大学）
Comments 9 pages, 6 figures, Neurips
Symmetries in Overparametrized Neural Networks: A Mean-Field View
Comments Final version. 59 pages. 10 figures
Journal ref Advances in Neural Information Processing Systems 37 (NeurIPS 2024)
ACSE-Eval: Can LLMs threat model real-world cloud infrastructure?
Comments Submitted to the 39th Annual Conference on Neural Information Processing Systems
MEDMKG: Benchmarking Medical Knowledge Exploitation with Multimodal Knowledge Graph
机构 * Pennsylvania State University（宾夕法尼亚州立大学） ； Renmin Hospital of Wuhan University（武汉大学仁民医院） ； Stony Brook University（石溪大学）
Comments Submitted to Neurips 2025
Erasing Undesirable Concepts in Diffusion Models with Adversarial Preservation
Comments Erasing Concepts, Generative Unlearning, NeurIPS 2024. arXiv admin note: text overlap with arXiv:2403.12326
Revisiting Hyperparameter Tuning with Differential Privacy
Comments ML Safety Workshop of NeurIPS'22 Accepted Paper
Make-An-Agent: A Generalizable Policy Network Generator with Behavior-Prompted Diffusion
机构 * University of Maryland, College Park（马里兰大学学院公园分校） ； Tsinghua University, IIIS（清华大学智能系统研究所） ； University of California, San Diego（加州大学圣地亚哥分校）
Comments Annual Conference on Neural Information Processing Systems 38
Directional Non-Commutative Monoidal Structures for Compositional Embeddings in Machine Learning
机构 * Swansea University（斯旺西大学）
Comments 11 pages submitted to NeurIPS 2025
Deep Policy Gradient Methods Without Batch Updates, Target Networks, or Replay Buffers
Comments In The Thirty-eighth Annual Conference on Neural Information Processing Systems. Source code at https://github.com/gauthamvasan/avg and companion video at https://youtu.be/cwwuN6Hyew0
Graph Neural Networks Do Not Always Oversmooth
机构 * RWTH Aachen University（亚琛工业大学） ； Department of Physics, RWTH Aachen University（物理系，亚琛工业大学） ； Institute for Advanced Simulation (IAS-6), Computational and Systems Neuroscience, Jülich Research Centre（先进模拟研究所（IAS-6），计算与系统神经科学，耶鲁研究中心）
Journal ref Advances in Neural Information Processing Systems 37 (NeurIPS 2024)
Inference via Interpolation: Contrastive Representations Provably Enable Planning and Inference
机构 * Princeton University（普林斯顿大学） ； UC Berkeley（加州大学伯克利分校） ； Carnegie Mellon University（卡内基梅隆大学）
Comments Code: https://github.com/vivekmyers/contrastive_planning
Journal ref Neural information processing systems (2024)
Decoupling Classifier for Boosting Few-shot Object Detection and Instance Segmentation
机构 * Tencent YouTu Lab（腾讯优图实验室） ； CATL（宁德时代）
Comments Accepted by NeurIPS 2022
BeamClean: Language Aware Embedding Reconstruction
Comments 9 pages, 5 figures, under review at NeurIPS, 2025
Concept-Level Explainability for Auditing & Steering LLM Responses
机构 * Department of Computer Science ETH Zurich（计算机科学系 苏黎世联邦理工学院）
Comments 9 pages, 7 figures, Submission to Neurips 2025
WaLRUS: Wavelets for Long-range Representation Using SSMs
Comments 15 pages, 8 figures. Submitted to Neurips 2025
A Self-Improving Coding Agent
机构 * University of Bristol（布里斯托大学） ； iGent AI
Comments Submitted as a preprint to NeurIPS 2025
PEAC: Unsupervised Pre-training for Cross-Embodiment Reinforcement Learning
机构 * Department of Computer Science & Technology, Institute for AI, BNRist Center, Tsinghua-Bosch Joint ML Center, THBI Lab, Tsinghua University（计算机科学与技术系、人工智能研究所、BNRist中心、清华-博世联合机器学习中心、THBI实验室、清华大学）
Comments NeurIPS24
Journal ref NeurIPS 2024
Scalable Density-based Clustering with Random Projections
机构 * University of Auckland（奥克兰大学）
Comments Appear in NeurIPS 2024 with the new title "Scalable DBSCAN with Random Projections"
Out-of-distribution generalisation is hard: evidence from ARC-like tasks
Comments Submission to NeurIPS 2025
Improving the Data-efficiency of Reinforcement Learning by Warm-starting with LLM
Comments 31 pages (9 for the main paper), 27 figures, NeurIPS 25 submission
Auditing Fairness by Betting
机构 * Machine Learning Department, Carnegie Mellon University（卡内基梅隆大学机器学习系） ； Carnegie Mellon University（卡内基梅隆大学） ； Department of Statistics and Data Science, Carnegie Mellon University（卡内基梅隆大学统计与数据科学系）
Comments Accepted to NeurIPS 2023. 28 pages, 5 figures
Understanding and Improving Adversarial Collaborative Filtering for Robust Recommendation
Comments To appear in NeurIPS 2024
Journal ref Advances in Neural Information Processing Systems, 37 (2024), 120381-120417
Construction and Application of Materials Knowledge Graph in Multidisciplinary Materials Science via Large Language Model
机构 * School of Computer Science and Engineering, University of New South Wales（新南威尔士大学计算机科学与工程学院） ； GreenDynamics（绿动力） ； Department of Materials Science and Engineering, City University of Hong Kong（香港城市大学材料科学与工程系） ； Department of Linguistics and Translation, City University of Hong Kong（香港城市大学语言学与翻译系） ； School of Photovoltaic and Renewable Energy Engineering, University of New South Wales（新南威尔士大学光伏与可再生能源工程学院） ； College of Design & Innovation, Tongji University（同济大学设计与创新学院）
Comments Accepted by 38th Conference on Neural Information Processing Systems (NeurIPS 2024)
MetaUAS: Universal Anomaly Segmentation with One-Prompt Meta-Learning
机构 * Tencent YouTu Lab（腾讯优图实验室）
Comments Accepted by NeurIPS 2024
Optimal-state Dynamics Estimation for Physics-based Human Motion Capture from Videos
机构 * Department of Electrical Engineering, Linköping University（_linköping大学电气工程系） ； Delft Center for Systems and Control, Delft University of Technology（代尔夫特理工大学系统与控制中心）
Comments 17 pages, 7 figure, NeurIPS 2024
Steering Large Language Models using Conceptors: Improving Addition-Based Activation Engineering
机构 * University of Groningen（格罗宁根大学）
Comments Presented at the MINT workshop at NeurIPS 2024. v4: fix sign in equation 10
EffiLearner: Enhancing Efficiency of Generated Code via Self-Optimization
Comments Accepted by NeurIPS 2024
EffiBench: Benchmarking the Efficiency of Automatically Generated Code
Comments Camera Ready for NeurIPS 2024
