Guy
16 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
SALLIE introduces a lightweight, modal-agnostic runtime detection framework that effectively safeguards LLMs and VLMs against both textual and visual jailbreaks and prompt injections without performance degradation.
The paper introduces Security Level 5 (SL5), a new, highly stringent security standard for AI systems designed to withstand attacks from state-level, top-tier cyber adversaries.
The paper argues that agent security must be treated as a systems problem, requiring the enforcement of security invariants at the system level rather than solely relying on improving the underlying AI model's robustness.
The paper introduces partial multi-neuron relaxation, a novel verification technique that selectively computes tight linear bounds for a small subset of neurons to improve the efficiency and tightness of neural network safety verification.
The paper introduces partial multi-neuron relaxation, a novel verification technique that selectively computes tight linear bounds for a small subset of neurons to improve the efficiency and tightness of neural network safety verification.
The paper introduces ProbMoE, a probabilistic routing framework that tackles the non-differentiability of top-$k$ routing in Mixture-of-Experts (MoE) models, achieving strong performance with improved expert utilization.
The paper proposes a novel probabilistic globally constrained decoding (P-GCD) method that efficiently constructs proposals for locally constrained decoding, significantly improving convergence speed and performance compared to existing approaches.
The paper introduces ProbMoE, a probabilistic routing framework that tackles the non-differentiability of top-$k$ routing in Mixture-of-Experts (MoE) models, achieving strong performance with improved expert utilization.
The paper proposes a novel probabilistic globally constrained decoding (P-GCD) method that efficiently constructs proposals for locally constrained decoding, significantly improving convergence speed and performance compared to existing approaches.
The authors introduce the VibrantForests framework for mapping forest attributes and providing a foundation for effective forest and wildfire planning.
This paper proposes a Real-Bogus classification framework using injection-driven, weakly supervised training without human-labeled data, achieving strong performance and calibrated uncertainties.
A multimodal Mixture-of-Experts framework was developed for asthma detection using vocal biomarkers and clinical data, achieving better performance than unimodal and bimodal approaches.
This paper proposes AURA, a meta-learning framework for efficient online adaptation of high-dimensional models in non-stationary environments.
The paper presents a model-based reinforcement learning framework for assistive agents in shared-control settings that addresses the challenges of proactive assistance and human legibility.
This paper investigates the impact of incongruence between text and audio on speech emotion recognition in voice-interactive large language models.
The paper introduces MaxIM, a framework for incremental summarization that optimizes memory representation for downstream task utility.
Papers
MaxIM: Maximally Informative Incremental Summarization via Reinforcement Learning
Jihwan Jeong, Guy Tennenholtz, Yinlam Chow, Chih-Wei Hsu +1 more
The paper introduces MaxIM, a framework for incremental summarization that optimizes memory representation for downstream task utility.