Events

ATOMIC: Attention Reallocation to Mitigate Irrelevant Context in Small Language Models (Mar 3)

Published: | 12:28 pm | Posted in: Events

Speaker: Bolin Shen Date: March 3, 11:45 – 12:45 pm Abstract: Small language models have recently gained increasing attention due to their strong cost efficiency, computational efficiency, and competitive performance across a wide range of reasoning tasks. However, when reasoning inputs contain irrelevant context, SLMS are substantially more vulnerable to distraction than large language models, […]

Robust Machine Learning on the Edge (Feb 27)

Published: | 9:09 am | Posted in: Events

Speaker: Stratis Ioannidis Date: Feb 27, 11:15 – 12:15 pm Abstract: Adversarial robustness, i.e., the ability of a machine learning (ML) algorithm to maintain its predictive power under input perturbations, is an important property for many safety-critical applications. It is even more important in edge deployments of ML algorithms, where inference is performed in a […]

ReAD: Reinforcement-Guided Capability Distillation (Feb 24)

Published: | 10:58 am | Posted in: Events

Speaker: Xueqi Cheng Date: Feb 24, 11:45 – 12:45 pm Abstract: Knowledge distillation (KD) compresses a large model into a smaller one that preserves the capabilities needed for a downstream task, yet existing methods assume that capabilities can be optimized independently…

Deep Learning for 3D Scene Modeling & AI-Enhanced Healthcare (Feb 20)

Published: | 9:53 am | Posted in: Events

Speaker: Andy Duan Date: Feb 20, 11:45 – 12:45 pm Abstract: 3D scene modeling is fundamental in many applications including Virtual Reality, Augmented Reality, Autonomous Driving, Robotics, Telehealth, etc. In this talk, I will first discuss some of our recent works in 3D scene modeling including: 1) PanoDepth: a deep learning based omnidirectional depth estimation […]

Silhouette: Leveraging Consistency Mechanisms to Detect Bugs in Persistent Memory-Based File Systems (Dec 5)

Published: | 1:30 pm | Posted in: Events

Speaker: An-I Andy Wang Date: Dec 5, 2:15 – 3:05 pm Abstract: The emergence of persistent memory (PM), with its non-volatile and byte-addressable characteristics, has led to a novel storage programming paradigm. However, PM programs need to flush stores from CPU caches and correctly order them to avoid inconsistencies after a crash. As a result, […]