Ax Jianhong Pang, Ruoxi Cheng, Ziyi Ye, Xingjun Ma, Zuxuan Wu, Xuanjing Huang, Yu-Gang Jiang 4/9/2026

Steering the Verifiability of Multimodal AI Hallucinations

Studies multimodal LLM hallucinations, distinguishing obvious from elusive types; proposes steering hallucination verifiability.

Ax Dzianis Piatrashyn, Nikita Kotelevskii, Kirill Grishchenkov, Nikita Glazkov, Ivan Nasonov, Ilya Makarov, Timothy Baldwin, Preslav Nakov, Roman Vashurin, Maxim Panov 4/9/2026

ReDAct: Uncertainty-Aware Deferral for LLM Agents

ReDAct: uncertainty-aware deferral mechanism for LLM agents to mitigate hallucination errors in sequential decision-making by deferring to humans when uncertain.

Ax Artemy Rubtsov, Sergey Samsonov, Vladimir Ulyanov, Alexey Naumov 4/9/2026

Gaussian Approximation for Asynchronous Q-learning

Convergence rate analysis for asynchronous Q-learning with polynomial stepsize under high-dimensional central limit theorem conditions.

Ax Roberto Vercellino (National Laboratory of the Rockies), Jared Willard (National Laboratory of the Rockies), Gustavo Campos (National Laboratory of the Rockies), Weslley da Silva Pereira (National Laboratory of the Rockies), Olivia Hull (National Laboratory of the Rockies), Matthew Selensky (National Laboratory of the Rockies), Juliane Mueller (National Laboratory of the Rockies) 4/9/2026

Measurement of Generative AI Workload Power Profiles for Whole-Facility Data Center Infrastructure Planning

Methodology for measuring generative AI power consumption across data centers to address proprietary data gaps and infrastructure planning.

Ax Shaowei Liu, Xuanchi Ren, Tianchang Shen, Huan Ling, Saurabh Gupta, Shenlong Wang, Sanja Fidler, Jun Gao 4/9/2026

MoRight: Motion Control Done Right

arXiv paper on motion-controlled video generation with disentangled control and motion causality for physical scene dynamics.

Ax Ziqiao Ma, Xueyang Yu, Haoyu Zhen, Yuncong Yang, Joyce Chai, Chuang Gan 4/9/2026

Fast Spatial Memory with Elastic Test-Time Training

Elastic Test-Time Training method addressing catastrophic forgetting in long-context 3D reconstruction with plastic inference-time updates.