Ax Gwanwoo Song, Kwanyoung Park, Youngwoon Lee 3/17/2026

Chunk-Guided Q-Learning

Chunk-Guided Q-Learning algorithm for offline reinforcement learning balancing bootstrapping error and policy flexibility over long horizons.

Ax Maria Rosaria Briglia, Simone Facchiano, Paolo Cursi, Alessio Sampieri, Emanuele Rodol\`a, Guido Maria D'Amely di Melendugno, Luca Franco, Fabio Galasso, Iacopo Masi 3/17/2026

Not All Latent Spaces Are Flat: Hyperbolic Concept Control

Hyperbolic control mechanism using parallel transport to steer text-to-image models away from unsafe content generation.

Ax Georgii Serbin, Kirill Koshkin, Zhongao Sun, Anastasiya Bistrigova, C. C. Korikov 3/17/2026

SVD Contextual Sparsity Predictors for Fast LLM Inference

Framework for accelerating LLM inference using contextual sparsity predictors for ReGLU-based feed-forward networks with minimal accuracy loss.

Ax Xiaowen Jiang, Andrei Semenov, Sebastian U. Stich 3/17/2026

Enhancing LLM Training via Spectral Clipping

Spectral clipping optimization technique for LLM training that addresses spectral norm instability and gradient noise issues in standard optimizers.

Ax Shiyuan Li, Yixin Liu, Yu Zheng, Xiaofeng Cao, Shirui Pan, Heng Tao Shen 3/17/2026

Towards One-for-All Anomaly Detection for Tabular Data

OFA-TAD proposes generalist one-for-all anomaly detection for tabular data with cross-domain generalization, replacing dataset-specific training approaches.

Ax Jan Kobiolka, Christian Frey, Arlind Kadra, Gresa Shala, Josif Grabocka 3/17/2026

Learning to Order: Task Sequencing as In-Context Optimization

Demonstrates deep neural networks can meta-learn task sequencing from few demonstrations, enabling generalization to new sequencing problems without task-specific training.

Ax Ian Osband 3/17/2026

Delightful Policy Gradient

Delightful policy gradient method that addresses variance issues in policy gradient updates by accounting for action likelihood under current policy.

Ax Yu Hao (Beijing University of Posts and Telecommunications), Qiuyu Wang (Beijing University of Posts and Telecommunications), Cheng Yang (Beijing University of Posts and Telecommunications), Yawen Li (Beijing University of Posts and Telecommunications), Zhiqiang Zhang (Ant Group), Chuan Shi (Beijing University of Posts and Telecommunications) 3/17/2026

GNNVerifier: Graph-based Verifier for LLM Task Planning

GNNVerifier uses graph neural networks to verify and correct task plans generated by LLMs in autonomous agent systems, reducing hallucinations.