38 papers
Robust Context-Aware Detection of Malicious Instructions in Text
Buzhao Liu, Xinhang Ma, Yevgeniy Vorobeychik
The remarkable instruction-following ability of modern LLMs has enabled their practical use as the minds of agents that can autonomously complete increasingly complex tasks. Therei…
COSMOS: Model-Agnostic Personalized Federated Learning with Clustered Server Models and Pseudo-Label-Only Communication
Ben Rachmut, Luise Ge, William Yeoh +2
Federated learning (FL) in heterogeneous environments remains challenging because client models often differ in both architecture and data distribution. While recent approaches att…
Bootstrap Flow-Map Tree Sampling Enables Online Feedback Driven Search
Binglin Ji, Anindya Sarkar, Hengchang Lu +2
In many scientific and engineering domains, maximizing discovery within a limited sampling budget demands strategic, observation-guided exploration. While generative models have en…
Sequentially-Controlled Interactive Multi-Particle Flow-Maps for Online Feedback-Driven Search
Binglin Ji, Anindya Sarkar, Hengchang Lu +2
While generative models have enabled training-free reward alignment, current methods typically excel in local exploration within narrow regions of the underlying distribution. Thes…
PAPA: Online Personalized Active Preference Alignment
Anindya Sarkar, Nasik Muhammad Nafi, Isaac Lyngaas +2
Diffusion models are highly effective at modeling complex data distributions, including images and text. However, in applications like personalized recommender systems, the objecti…
AutoDojo: Adaptive Black-Box Attacks Reveal the Limits of IPI Defenses and Task-Specification Effects in LLM Agents
Xinhang Ma, Taoran Li, Chaowei Xiao +3
Indirect prompt injection (IPI) is a major security threat to LLM-powered agents. Thus, a growing body of work have proposed a variety of defensive approaches against IPI. These ca…