TMLR

Training Dynamics of the Cooldown Stage in Warmup-Stable-Decay Learning Rate Scheduler

Aleksandr Dremov, Alexander Hägele, Atli Kosson, Martin Jaggi

Comments Published in TMLR. Review: https://openreview.net/forum?id=ZnSYEcZod3

Journal ref Transactions on Machine Learning Research (TMLR), 2025

Probabilities of Chat LLMs Are Miscalibrated but Still Predict Correctness on Multiple-Choice Q&A

Benjamin Plaut, Nguyen X. Khanh, Tu Trinh

机构 * Center for Human-Compatible AI（兼容人类的人工智能中心） ； UC Berkeley（伯克利大学） 纠错

Comments Published in Transactions on Machine Learning Research (TMLR)

Understanding In-Context Learning of Linear Models in Transformers Through an Adversarial Lens

Usman Anwar, Johannes Von Oswald, Louis Kirsch, David Krueger, Spencer Frei

机构 * University of Cambridge（剑桥大学） ； Google, Paradigms of Intelligence Team（谷歌，Paradigms of Intelligence团队） ； Google DeepMind（谷歌DeepMind） ； Mila & Université de Montréal（Mila与蒙特利尔大学） ； UC Davis（加州大学戴维斯分校） 纠错

Comments Transactions on Machine Learning Research (TMLR) 2025, with Featured Certification

Set-Based Training for Neural Network Verification

Lukas Koller, Tobias Ladner, Matthias Althoff

Comments published at Transactions on Machine Learning Research (TMLR)

Generalized Compressed Sensing for Image Reconstruction with Diffusion Probabilistic Models

Ling-Qi Zhang, Zahra Kadkhodaie, Eero P. Simoncelli, David H. Brainard

机构 * Janelia Research Campus, Howard Hughes Medical Institute（贾勒尼亚研究campus，霍华德·休斯医学研究所） ； Flatiron Institute, Simons Foundation（Flatiron研究所，Simons基金会） ； New York University（纽约大学） ； University of Pennsylvania（宾夕法尼亚大学） 纠错

Comments Transactions on Machine Learning Research (2025)

LumiNet: Perception-Driven Knowledge Distillation via Statistical Logit Calibration

Md. Ismail Hossain, M M Lutfe Elahi, Sameera Ramasinghe, Ali Cheraghian, Fuad Rahman, Nabeel Mohammed, Shafin Rahman

机构 * Apurba-NSU R&D Lab, North South University, Bangladesh（阿普尔巴-NSU研发实验室，北南大学，孟加拉国） ； Pluralis Research（普拉尔西斯研究） ； Data61-CSIRO, Australia & Australian National University, Australia（Data61-CSIRO，澳大利亚及澳大利亚国立大学，澳大利亚） ； Apurba Technologies, Sunnyvale, CA 94085, USA（阿普尔巴技术，硅谷，加州94085，美国） 纠错

Comments Accepted at Transactions on Machine Learning Research (TMLR), August 2025

Associative memory inspires improvements for in-context learning using a novel attention residual stream architecture

Thomas F Burns, Tomoki Fukai, Christopher J Earls

机构 * SciAI Center, Cornell University, USA（SciAI中心，康奈尔大学，美国） ； Neural Coding and Brain Computing Unit, OIST, Japan（神经编码与脑计算单元，海洋研究所，日本） ； SciAI Center, Center for Applied Mathematics, School of Civil and Environmental Engineering, Cornell University, USA（SciAI中心，应用数学中心，土木与环境工程学院，康奈尔大学，美国） 纠错

Comments 35 pages, 14 figures, 6 tables; accepted and published in TMLR

Model Tensor Planning

An T. Le, Khai Nguyen, Minh Nhat Vu, João Carvalho, Jan Peters

机构 * Intelligent Autonomous Systems, Department of Computer Science, Technical University of Darmstadt（德意志技术大学达姆施塔特分校计算机科学系智能自主系统研究中心） ； Systems AI for Robot Learning, German Research Center for AI (DFKI)（德国人工智能研究中心（DFKI）机器人学习系统AI研究中心） ； Hessian Center for Artificial Intelligence (hessian.AI)（黑森人工智能中心（hessian.AI）） ； Centre for Cognitive Science, Technical University of Darmstadt（德意志技术大学达姆施塔特分校认知科学中心） ； Automation & Control Institute, TU Wien（维也纳技术大学自动化与控制研究所） ； Austrian Institute of Technology (AIT), Vienna（奥地利技术研究所（AIT）） ； VinRobotics and VinUniversity（VinRobotics和Vin大学） 纠错

Comments 24 pages, 9 figures. Accepted to TMLR

Between Linear and Sinusoidal: Rethinking the Time Encoder in Dynamic Graph Learning

Hsing-Huan Chung, Shravan Chaudhari, Xing Han, Yoav Wald, Suchi Saria, Joydeep Ghosh

Comments Accepted to TMLR 7/2025

Are Domain Generalization Benchmarks with Accuracy on the Line Misspecified?

Olawale Salaudeen, Nicole Chiou, Shiny Weng, Sanmi Koyejo

Comments Published in TMLR 08/25

Gradient Inversion Attack on Graph Neural Networks

Divya Anand Sinha, Ruijie Du, Yezi Liu, Athina Markopolou, Yanning Shen

机构 * University of California, Irvine（加州大学欧文分校） 纠错

Comments https://openreview.net/forum?id=a0mLrqkWyx

Journal ref Transactions on Machine Learning Research, 2025, 2835-8856

Robust and Efficient Fine-tuning of LLMs with Bayesian Reparameterization of Low-Rank Adaptation

Ayan Sengupta, Vaibhav Seth, Arinjay Pathak, Aastha Verma, Natraj Raman, Sriram Gopalakrishnan, Niladri Chatterjee, Tanmoy Chakraborty

机构 * Indian Institute of Technology Delhi（印度理工学院德里） 纠错

Comments The paper is accepted in TMLR'25

Transformers trained on proteins can learn to attend to Euclidean distance

Isaac Ellmen, Constantin Schneider, Matthew I. J. Raybould, Charlotte M. Deane

Journal ref Transactions on Machine Learning Research (2025) https://openreview.net/forum?id=mU59bDyqqv

LLaVA-Video: Video Instruction Tuning With Synthetic Data

Yuanhan Zhang, Jinming Wu, Wei Li, Bo Li, Zejun Ma, Ziwei Liu, Chunyuan Li

机构 * S-Lab, Nanyang Technological University（南洋理工大学S实验室） ； BUPT（北京邮电大学） ； ByteDance（字节跳动） 纠错

Comments Project page: https://llava-vl.github.io/blog/2024-09-30-llava-video/; Accepted at TMLR

Loss Landscape Degeneracy and Stagewise Development in Transformers

Jesse Hoogland, George Wang, Matthew Farrugia-Roberts, Liam Carroll, Susan Wei, Daniel Murfet

机构 * Timaeus ； Department of Computer Science, University of Oxford（计算机科学系，牛津大学） ； Department of Econometrics and Business Statistics, Monash University（计量经济学与商业统计系，墨尔本大学） ； School of Mathematics and Statistics, the University of Melbourne（数学与统计学系，墨尔本大学） 纠错

Comments To appear, TMLR. Material on essential dynamics from v1 of this preprint has been removed and developed in arXiv:2501.17745

Convergence Properties of Natural Gradient Descent for Minimizing KL Divergence

Adwait Datar, Nihat Ay

Comments replaced earlier draft with the camera-ready version published at tmlr

A Survey on Large Language Model Acceleration based on KV Cache Management

Haoyang Li, Yiming Li, Anxin Tian, Tianhao Tang, Zhanchao Xu, Xuejia Chen, Nicole Hu, Wei Dong, Qing Li, Lei Chen

机构 * Department of Computing, The Hong Kong Polytechnic University（计算系，香港理工大学） ； Department of Computer Science and Engineering（计算机科学与工程系） ； Department of Computer Science and Technology（计算机科学与技术系） ； Department of Computing and Data Science（计算与数据科学系） 纠错

Comments Accepted to TMLR 2025. The revised version incorporates more papers and has been further polished

Scaling and Distilling Transformer Models for sEMG

Nicholas Mehlman, Jean-Christophe Gagnon-Audet, Michael Shvartsman, Kelvin Niu, Alexander H. Miller, Shagun Sodhani

机构 * Viterbi School of Engineering, University of Southern California（南加州大学工程学院） ； Meta FAIR 纠错

Comments Accepted at TMLR 2025 (https://openreview.net/forum?id=hFPWThwUiZ), 11 pages

Long-Term Fairness Inquiries and Pursuits in Machine Learning: A Survey of Notions, Methods, and Challenges

Usman Gohar, Zeyu Tang, Jialu Wang, Kun Zhang, Peter L. Spirtes, Yang Liu, Lu Cheng

机构 * Iowa State University（爱荷华州立大学） ； Carnegie Mellon University（卡内基梅隆大学） ； University of California, Santa Cruz（加州大学圣克鲁兹分校） ； University of Illinois Chicago（伊利诺伊大学芝加哥分校） 纠错

Comments Accepted in TMLR

Does equivariance matter at scale?

Johann Brehmer, Sönke Behrends, Pim de Haan, Taco Cohen

机构 * Qualcomm AI Research（高通人工智能研究） 纠错

Comments Version published in TMLR

Text-to-Image Generation Via Energy-Based CLIP

Roy Ganz, Michael Elad

机构 * Electrical Engineering Department Technion（技术学院电子工程系） ； Computer Science Department Technion（技术学院计算机科学系） 纠错

Comments Accepted to TMLR

MOCK: an Algorithm for Learning Nonparametric Differential Equations via Multivariate Occupation Kernel Functions

Victor Rielly, Kamel Lahouel, Ethan Lew, Nicholas Fisher, Vicky Haney, Michael Wells, Bruno Jedynak

机构 * Department of Mathematics（数学系） ； Portland State University（波特兰州立大学） 纠错

Comments 29 pages, 6 figures Accepted at Transactions in Machine Learning Research (TMLR)

Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilities

Zora Che, Stephen Casper, Robert Kirk, Anirudh Satheesh, Stewart Slocum, Lev E McKinney, Rohit Gandikota, Aidan Ewart, Domenic Rosati, Zichu Wu, Zikui Cai, Bilal Chughtai, Yarin Gal, Furong Huang, Dylan Hadfield-Menell

机构 * University of Maryland（马里兰大学） ； MIT CSAIL（麻省理工学院计算机科学与人工智能实验室） ； UK AI Security Institute（英国人工智能安全研究所） ； University of Toronto（多伦多大学） ； Northeastern University（东北大学） ； Haize Labs（Haize实验室） ； Dalhousie University（达尔豪斯大学） ； University of Waterloo（滑铁卢大学） ； University of Oxford（牛津大学） ； MIT（麻省理工学院） 纠错

Comments Accepted to TMLR

Unmasking Trees for Tabular Data

Calvin McCarter

机构 * BigHat Biosciences（BigHat生物科学公司） 纠错

Comments Published in Transactions on Machine Learning Research (7/2025)

How to Leverage Predictive Uncertainty Estimates for Reducing Catastrophic Forgetting in Online Continual Learning

Giuseppe Serra, Ben Werner, Florian Buettner

机构 * Goethe University Frankfurt（弗赖堡大学） ； German Cancer Consortium (DKTK)（德国癌症研究中心（DKTK）） ； German Cancer Research Center (DKFZ)（德国癌症研究中心（DKFZ）） 纠错

Comments Published in Transactions on Machine Learning Research (March 2025)

Knockout: A simple way to handle missing inputs

Minh Nguyen, Batuhan K. Karaman, Heejong Kim, Alan Q. Wang, Fengbei Liu, Mert R. Sabuncu

机构 * Cornell University（康奈尔大学） ； Weill Cornell Medicine（韦尔医学院） 纠错

Comments Accepted at TMLR

Revisiting Data Augmentation for Ultrasound Images

Adam Tupper, Christian Gagné

机构 * Institut Intelligence et Données (IID)（智能与数据研究所） ； Université Laval（拉瓦尔大学） ； Mila 纠错

Comments Published in the Transacations of Machine Learning Research (TMLR, 2025), see https://openreview.net/forum?id=iGcxlTLIL5 . For the associated source code see https://github.com/adamtupper/ultrasound-augmentation

Unified Triplet-Level Hallucination Evaluation for Large Vision-Language Models

Junjie Wu, Tsz Ting Chung, Kai Chen, Dit-Yan Yeung

机构 * The Hong Kong University of Science and Technology（香港科学与技术大学） 纠错

Comments Accepted by TMLR 2025. Project Page: https://kaichen1998.github.io/projects/tri-he/

Tracing Facts or just Copies? A critical investigation of the Competitions of Mechanisms in Large Language Models

Dante Campregher, Yanxu Chen, Sander Hoffman, Maria Heuss

机构 * University of Amsterdam（阿姆斯特丹大学） 纠错

Comments 18 Pages, 13 figures

Journal ref Transactions on Machine Learning Research (TMLR), 2025

Tackling the Abstraction and Reasoning Corpus with Vision Transformers: the Importance of 2D Representation, Positions, and Objects

Wenhao Li, Yudong Xu, Scott Sanner, Elias Boutros Khalil

机构 * University of Toronto（多伦多大学） ； Vector Institute for Artificial Intelligence（人工智能向量研究所） 纠错

Journal ref Transactions on Machine Learning Research (TMLR), 07/2025, https://openreview.net/forum?id=Al72Fp0rCg