UVa ILP Blog

• 2024-08-16 The Mismeasure of Man and Models

  TL;DR: Evaluating Allocational Harms in Large Language Models

• 2022-11-28 CS-Shapley: Class-wise Shapley Values for Data Valuation in Classification

  TL;DR: A new value function for estimating data valuation in classification tasks

• 2022-11-11 Balanced Adversarial Training: Balancing Tradeoffs between Fickleness and Obstinacy in NLP Models

  TL;DR: Balancing Tradeoffs between Fickleness and Obstinacy in NLP Models

• 2021-11-21 EMNLP 2021 Quick Notes

  TL;DR: A quick summary of some EMNLP 2021 papers.

• 2021-11-18 Tradeoffs between Invariance- and Sensitivity-based Adversarial Perturbations

  TL;DR: The fundamental tradeoffs between two types of adversarial perturbations and the underlying explanations.