5 papers
TopoAlign: Topology-Aware Visual Representation Alignment
Xinyuan Yan, Rita Sevastjanova, Mennatallah El-Assady +1
Neural networks encode inputs as high-dimensional vectors, known as representations, that capture how models process data by encoding task-relevant structure and semantics. Represe…
Teaching People LLM's Errors and Getting it Right
Nathan Stringham, Fateme Hashemi Chaleshtori, Xinyuan Yan +3
People use large language models (LLMs) when they should not. This is partly because they see LLMs compose poems and answer intricate questions, so they understandably, but incorre…
Flexible and Probabilistic Topology Tracking with Partial Optimal Transport
Mingzhe Li, Xinyuan Yan, Lin Yan +2
In this paper, we present a flexible and probabilistic framework for tracking topological features in time-varying scalar fields using merge trees and partial optimal transport. Me…
Explainable Mapper: Charting LLM Embedding Spaces Using Perturbation-Based Explanation and Verification Agents
Xinyuan Yan, Rita Sevastjanova, Sinie van der Ben +2
Large language models (LLMs) produce high-dimensional embeddings that capture rich semantic and syntactic relationships between words, sentences, and concepts. Investigating the to…
VISLIX: An XAI Framework for Validating Vision Models with Slice Discovery and Analysis
Xinyuan Yan, Xiwei Xuan, Jorge Piazentin Ono +6
Real-world machine learning models require rigorous evaluation before deployment, especially in safety-critical domains like autonomous driving and surveillance. The evaluation of…