Ax Jianrui Zhang, Yue Yang, Rohun Tripathi, Winson Han, Ranjay Krishna, Christopher Clark, Yong Jae Lee, Sangho Lee 3/19/2026

Unified Spatio-Temporal Token Scoring for Efficient Video VLMs

Token pruning framework for efficient video vision-language models reducing computational cost via temporal token scoring.

Ax Ling Chen, Jiahui Xu, Binqing Wu, Mingqi Lv, Chaoqun Zhan, Sanjian Chen, Jian Chang 3/19/2026

HighAir: A Hierarchical Graph Neural Network-Based Air Quality Forecasting Method

Graph neural network method for air quality forecasting modeling pollution diffusion between cities and monitoring stations.

Ax Bart Cox, Lydia Y. Chen, J\'er\'emie Decouchant 3/19/2026

Aergia: Leveraging Heterogeneity in Federated Learning Systems

Aergia: federated learning system leveraging client heterogeneity in computing power to reduce training time.

Ax Jun Sun, Wancheng Zhang, Chao Zhou, Zhongjie Mao, Chao Li, Xiao-Jun Wu 3/19/2026

Feature Space Renormalization for Semi-supervised Learning

Feature space renormalization mechanism for semi-supervised learning improving consistency regularization on unlabeled data.

Ax Bruno Laboissiere Camargos Borges, Bruno Machado Pacheco, Danilo Silva 3/19/2026

Soft Dice Confidence: A Near-Optimal Confidence Estimator for Selective Prediction in Semantic Segmentation

Soft Dice Confidence: confidence estimator for selective prediction in semantic segmentation enabling model abstention.

Ax Chuang Liu, Zelin Yao, Xueqi Ma, Mukun Chen, Luzhi Wang, Jia Wu, Wenbin Hu 3/19/2026

Hi-GMAE: Hierarchical Graph Masked Autoencoders

Hi-GMAE: hierarchical graph masked autoencoders for multi-scale self-supervised learning on graph-structured data.

Ax Francesco Montagna, Max Cairney-Leeming, Dhanya Sridhar, Francesco Locatello 3/19/2026

Demystifying amortized causal discovery with transformers

Analyzes transformer-based amortized causal discovery on observational data, bridging supervised learning with identifiability theory.

Ax Ruining Yang, Yi Xu, Yun Fu, Lili Su 3/19/2026

Den-TP: A Density-Balanced Data Curation and Evaluation Framework for Trajectory Prediction

Den-TP: data curation framework addressing long-tail distribution in trajectory prediction datasets for autonomous driving.

Ax Aleksandar Vujinovic, Aleksandar Kovacevic 3/19/2026

ACT-JEPA: Novel Joint-Embedding Predictive Architecture for Efficient Policy Representation Learning

ACT-JEPA: joint-embedding predictive architecture for efficient policy representation learning via self-supervised learning from unlabeled data.

Ax Xuyang Li, Romit Maulik 3/19/2026

SALSA-RL: Stability Analysis in the Latent Space of Actions for Reinforcement Learning

SALSA-RL: stability analysis method for deep reinforcement learning agents enabling interpretability and safety assessments in continuous control.

Ax Yuxiao Wen, Yanjun Han, Zhengyuan Zhou 3/19/2026

Joint Value Estimation and Bidding in Repeated First-Price Auctions

Studies regret minimization in repeated first-price auctions with causal inference for online advertising scenarios.

Ax Ioannis Dimanidis, Tolga Ok, Peyman Mohajerin Esfahani 3/19/2026

Offline Reinforcement Learning via Inverse Optimization

Offline reinforcement learning algorithm leveraging inverse optimization and sub-optimality loss for continuous state/action spaces.

Ax Xiaohong Yang, Minghui Liwang, Liqun Fu, Yuhan Su, Seyyedali Hosseinalipour, Xianbin Wang, Yiguang Hong 3/19/2026

Adaptive UAV-Assisted Hierarchical Federated Learning: Optimizing Energy, Latency, and Resilience for Dynamic Smart IoT

Hierarchical federated learning framework using UAVs as mobile aggregators for distributed IoT systems with limited connectivity.

Ax Cheng Zhen, Prayoga, Nischal Aryal, Arash Termehchy, Garrett Biwer, Lubna Alzamil 3/19/2026

Learning Over Dirty Data with Minimal Repairs

Proposes minimal repair concept showing imputing all missing values unnecessary; identifies critical missing data subsets for accurate ML models.

Ax Zihao Guo, Shuqing Shi, Richard Willis, Tristan Tomilin, Joel Z. Leibo, Yali Du 3/19/2026

SocialJax: An Evaluation Suite for Multi-agent Reinforcement Learning in Sequential Social Dilemmas

SocialJax: evaluation suite for multi-agent reinforcement learning in sequential social dilemmas, measuring agent generalization.

Ax Deshani Geethika Poddenige, Sachith Seneviratne, Asela Hevapathige, Damith Senanayake, Mahesan Niranjan, PN Suganthan, Saman Halgamuge 3/19/2026

Arch-VQ: Discrete Architecture Representation Learning with Autoregressive Priors

Proposes Arch-VQ for learning discrete neural architecture representations using autoregressive priors instead of continuous VAE mapping.

Ax Alireza Aghabagherloo, Aydin Abadi, Sumanta Sarkar, Vishnu Asutosh Dasu, Bart Preneel 3/19/2026

Impact of Data Duplication on Deep Neural Network-Based Image Classifiers: Robust vs. Standard Models

Studies impact of duplicated training data on deep neural network image classifiers, comparing robust vs. standard models against adversarial attacks.

Ax Simone Piaggesi, Riccardo Guidotti, Fosca Giannotti, Dino Pedreschi 3/19/2026

Explanations Go Linear: Post-hoc Explainability for Tabular Data with Interpretable Meta-Encoding

arXiv paper: ILLUME method for post-hoc explainability of tabular ML models with interpretable meta-encoding.

Ax Jenni Lampainen, Kaisa Joki, Napsu Karmitsa, Marko M. M\"akel\"a 3/19/2026

Clust-Splitter - an Efficient Nonsmooth Optimization-Based Algorithm for Clustering Large Datasets

arXiv paper: Clust-Splitter algorithm for efficient clustering on large datasets using nonsmooth optimization.

Ax Arjun Prakash, Naicheng He, Denizalp Goktas, Jacob Makar-Limanov, Amy Greenwald 3/19/2026

Bi-Level Policy Optimization with Nystr\"om Hypergradients

arXiv paper: Bi-level policy optimization with Nyström hypergradients for actor-critic reinforcement learning algorithms.

Ax Yasi Zhang, Tianyu Chen, Zhendong Wang, Ying Nian Wu, Mingyuan Zhou, Oscar Leong 3/19/2026

Score Distillation Beyond Acceleration: Generative Modeling from Corrupted Data

arXiv paper: Restoration Score Distillation framework for learning generative models from corrupted data. Novel ML research.

Ax Aobo Liang, Yan Sun, Xiaohou Shi, Ke Li 3/19/2026

Time Tracker: Mixture-of-Experts-Enhanced Foundation Time Series Forecasting Model with Decoupled Training Pipelines

Foundation model for time series forecasting using mixture-of-experts architecture with decoupled training to handle diverse temporal patterns and multi-variable correlations.

Ax Peter Holderrieth, Ezra Erives 3/19/2026

An Introduction to Flow Matching and Diffusion Models

Tutorial on diffusion and flow-based generative models covering mathematical foundations, ODEs, SDEs, and core algorithms for image, video, and multi-modal generation.

Ax Yihong Guo, Yu Yang, Pan Xu, Anqi Liu 3/19/2026

MOBODY: Model Based Off-Dynamics Offline Reinforcement Learning

Offline reinforcement learning method for mismatched dynamics leveraging model-based approaches to explore high-reward states.

Ax Luca Stradiotti, Dario Pesenti, Stefano Teso, Jesse Davis 3/19/2026

Knowing What You Cannot Explain: Learning to Reject Low-Quality Explanations

Learning to Reject framework extending ML models to abstain from predictions and explanations with low quality.

Ax Kazuki Irie, Samuel J. Gershman 3/19/2026

Fast weight programming and linear transformers: from machine learning to neurobiology

Fast weight programmers with 2D matrix hidden states connecting RNNs, language modeling, and neurobiology.

Ax Yuxiang Ji, Ziyu Ma, Yong Wang, Guanhua Chen, Xiangxiang Chu, Liaoni Wu 3/19/2026

Tree Search for LLM Agent Reinforcement Learning

Tree-based group relative policy optimization for LLM agents addressing sparse supervision in multi-turn tasks.

Ax Aayush Mishra, Daniel Khashabi, Anqi Liu 3/19/2026

IA2: Alignment with ICL Activations Improves Supervised Fine-Tuning

Method aligning supervised fine-tuning with in-context learning activations to improve LLM generalization and calibration.

Ax Qiushui Xu, Yuhao Huang, Yushu Jiang, Lei Song, Jinyu Wang, Wenliang Zheng, Jiang Bian 3/19/2026

In-Context Compositional Q-Learning for Offline Reinforcement Learning

Offline reinforcement learning using in-context learning with linear Transformers for compositional Q-function estimation.

Ax Arpit Garg, Hemanth Saratchandran, Ravi Garg, Simon Lucey 3/19/2026

Stable Forgetting: Bounded Parameter-Efficient Unlearning in Foundation Models

Parameter-efficient unlearning method for foundation models addressing privacy/safety with bounded weight growth.

Ax Ziyan Wang, Zheng Wang, Xingwei Qu, Qi Cheng, Jie Fu, Shengpu Tang, Minjia Zhang, Xiaoming Huo 3/19/2026

Slow-Fast Policy Optimization: Reposition-Before-Update for LLM Reasoning

Slow-Fast Policy Optimization framework for improving LLM reasoning via reinforcement learning with stable gradient updates.

Ax Hyunsik Yoo, Ting-Wei Li, SeongKu Kang, Zhining Liu, Charlie Xu, Qilin Qi, Hanghang Tong 3/19/2026

Continual Low-Rank Adapters for LLM-based Generative Recommender Systems

Proposes continual low-rank adapters for LLM-based recommender systems handling evolving users and preferences without catastrophic forgetting.

Ax Yanan Zhao, Feng Ji, Jingyang Dai, Jiaze Ma, Keyue Jiang, Kai Zhao, Wee Peng Tay 3/19/2026

Adaptive Multi-view Graph Contrastive Learning via Fractional-order Neural Diffusion Networks

Presents augmentation-free graph contrastive learning via fractional-order neural diffusion networks for multi-scale structure learning.

Ax Jorge Paz-Ruza, Jo\~ao Gama, Amparo Alonso-Betanzos, Bertha Guijarro-Berdi\~nas 3/19/2026

A robust methodology for long-term sustainability evaluation of Machine Learning models

Proposes standardized methodology for evaluating long-term sustainability and efficiency of ML models addressing Green AI gaps.

Ax Nathan Breslow, Aayush Mishra, Mahler Revsine, Michael C. Schatz, Anqi Liu, Daniel Khashabi 3/19/2026

Genomic Next-Token Predictors are In-Context Learners

Demonstrates in-context learning emerges organically in genomic sequence models trained with next-token prediction on DNA sequences.

Ax Leo Elmecker-Plakolm, Pierre Fasterling, Philip Sosnin, Calvin Tsay, Matthew Wicker 3/19/2026

Provably Safe Model Updates

Develops methods for provably safe ML model updates preventing catastrophic forgetting and alignment drift in dynamic environments.

Ax Zhongjian Qiao, Rui Yang, Jiafei Lyu, Chenjia Bai, Xiu Li, Siyang Gao, Shuang Qiu 3/19/2026

Efficient Cross-Domain Offline Reinforcement Learning with Dynamics- and Value-Aligned Data Filtering

Proposes data filtering method for cross-domain offline RL addressing dynamics misalignment between source and target domains.

Ax Vedant Shah, Johan Obando-Ceron, Vineet Jain, Brian Bartoldson, Bhavya Kailkhura, Sarthak Mittal, Glen Berseth, Pablo Samuel Castro, Yoshua Bengio, Nikolay Malkin, Moksh Jain, Siddarth Venkatraman, Aaron Courville 3/19/2026

A Comedy of Estimators: On KL Regularization in RL Training of LLMs

Analyzes KL regularization estimators in RL training of LLMs, comparing bias-variance tradeoffs of different approximation methods.

Ax Zhehao Huang, Baijiong Lin, Jingyuan Zhang, Jingying Wang, Yuhang Liu, Ning Lu, Tao Li, Xiaolin Huang 3/19/2026

VL-RouterBench: A Benchmark for Vision-Language Model Routing

VL-RouterBench benchmark for evaluating vision-language model routing systems with quality-cost tradeoff assessment at scale.

Ax Yuxuan Li, Harshith Reddy Kethireddy, Srijita Das 3/19/2026

Evaluating Feature Dependent Noise in Preference-based Reinforcement Learning

Evaluates feature-dependent noise in preference-based reinforcement learning with realistic noise patterns correlated to observations.

Ax Zhengyang Zhao, Lu Ma, Yizhen Jiang, Xiaochen Ma, Zimo Meng, Chengyu Shen, Lexiang Tang, Haoze Sun, Peng Pei, Wentao Zhang 3/19/2026

GIFT: Reconciling Post-Training Objectives via Finite-Temperature Gibbs Initialization

Proposes GIFT method reconciling SFT and RL post-training for Large Reasoning Models via Gibbs initialization to prevent distributional collapse.

Ax Ming Li 3/19/2026

Global Optimization By Gradient From Hierarchical Score-Matching Spaces

Solves constrained optimization problems via gradient-based methods using hierarchical score-matching spaces to overcome local optima.

Ax Wei Chen, Xingyu Guo, Shuang Li, Zhao Zhang, Yan Zhong, Fuzhen Zhuang, Deqing wang 3/19/2026

Learning Adaptive Distribution Alignment with Neural Characteristic Function for Graph Domain Adaptation

Proposes neural characteristic function approach for graph domain adaptation addressing distributional shifts without manual feature design.

Ax Jialin Yu, Mo\"ise Blanchard 3/19/2026

Distribution-Free Sequential Prediction with Abstentions

Studies sequential prediction with option to abstain in semi-adversarial settings mixing adversarial and stochastic instances.

Ax Danning Jing, Xinhai Chen, Xifeng Pu, Jie Hu, Chao Huang, Xuguang Chen, Qinglin Wang, Jie Liu 3/19/2026

A Deep Surrogate Model for Robust and Generalizable Long-Term Blast Wave Prediction

Creates deep surrogate model for blast wave prediction that generalizes to out-of-distribution urban scenarios using machine learning.

Ax Nazal Mohamed, Ayush Mohanty, Nagi Gebraeel 3/19/2026

Federated Causal Representation Learning in State-Space Systems for Decentralized Counterfactual Reasoning

Develops federated causal representation learning for decentralized counterfactual reasoning across coupled industrial systems while preserving data privacy.

Ax Harry Amad, Mihaela van der Schaar 3/19/2026

Hyperparameter Trajectory Inference with Conditional Lagrangian Optimal Transport

Proposes Hyperparameter Trajectory Inference to adjust neural network hyperparameters post-deployment without full retraining using optimal transport.

Ax Huihan Liu, Changyeon Kim, Bo Liu, Minghuan Liu, Yuke Zhu 3/19/2026

Pretrained Vision-Language-Action Models are Surprisingly Resistant to Forgetting in Continual Learning

Studies how pretrained Vision-Language-Action models resist catastrophic forgetting during continual learning in robot policy training.

Ax Hengshuai Yao, Xing Chen, Ahmed Murtadha, Guan Wang 3/19/2026

Thin Keys, Full Values: Reducing KV Cache via Low-Dimensional Attention Selection

Reduces transformer KV cache by using low-dimensional keys for attention selection while maintaining high-dimensional values, achieving O(log N) dimensional compression.

Ax Fengxiang Nie, Yasuhiro Suzuki 3/19/2026

JAWS: Enhancing Long-term Rollout of Neural PDE Solvers via Spatially-Adaptive Jacobian Regularization

JAWS improves neural PDE solvers' long-term rollouts using spatially-adaptive Jacobian regularization to prevent spectral blow-up and unphysical divergence.