Furong Huang (University of Maryland): “Rethinking Test-Time Thinking: From Token-Level Rewards to Robust Generative Agents”
Abstract:
We present a unified perspective on test-time thinking as a lens for improving generative AI agents through finer-grained reward modeling, data-centric reasoning, and robust alignment. Beginning with GenARM, we introduce an inductive bias for denser, token-level reward modeling that guides generation during decoding, enabling token-level alignment without retraining. While GenARM targets reward design, ThinkLite-VL focuses on the data side of reasoning. It proposes a self-improvement framework that selects the most informative samples via MCTS-guided search, yielding stronger visual reasoning with fewer labels. Taking this a step further, MORSE-500 moves beyond selection to creation: it programmatically generates targeted, controllable multimodal data to systematically probe and stress-test models’ reasoning abilities. We then interrogate a central assumption in inference-time alignment: Does Thinking More Always Help? Our findings reveal that increased reasoning steps can degrade performance–not due to better or worse reasoning per se, but due to rising variance in outputs, challenging the naive scaling paradigm. Finally, AegisLLM applies test-time thinking in the service of security, using an agentic, multi-perspective framework to defend against jailbreaks, prompt injections, and unlearning attacks–all at inference time. Together, these works chart a path toward generative agents that are not only more capable, but more data-efficient, introspective, and robust in real-world deployment.
Biography:
Furong Huang is an Associate Professor of the Department of Computer Science at the University of Maryland. Specializing in trustworthy machine learning, Security in AI, AI for sequential decision-making, and generative AI, Dr. Huang focuses on applying principles to solve practical challenges in contemporary computing to develop efficient, robust, scalable, sustainable, ethical, and responsible machine learning algorithms. She is recognized for her contributions with awards including best paper awards, the MIT Technology Review Innovators Under 35 Asia Pacific, the MLconf Industry Impact Research Award, the NSF CRII Award, the Microsoft Accelerate Foundation Models Research award, the Adobe Faculty Research Award, three JP Morgan Faculty Research Awards and Finalist of AI in Research – AI researcher of the year for Women in AI Awards North America.
Zoom: https://upenn.zoom.us/j/94857956796
