Unraveling The Impact of Training Samples

Daiwei Chen, Jane Zhang, Ramya Korlakai Vinayak

Keywords: Machine Learning, Deep Learning, interpretability, Dataset

Blogpost Url: https://iclr-blogposts.github.io/2024/blog/unraveling-the-impact-of-training-samples/

Abstract: How do we quantify the influence of datasets? Recent works on Data Attribution Methods shed light on this problem. In this blog post, we introduce Data Attribution Methods which leverage robust statistics and surrogate functions, and present their applications like distinguishing the feature selection difference of learning algorithms, detecting data leakage, and assessing model robustness.

Ref Papers: https://arxiv.org/abs/1703.04730, https://arxiv.org/abs/2202.00622, https://arxiv.org/abs/2303.14186, https://arxiv.org/abs/2211.12491

Id Of The Authors Of The Papers: ~Percy_Liang1, ~Aleksander_Madry1,

Conflict Of Interest: No

Submission Number: 43