Get our free extension to see links to code for papers anywhere online!

Chrome logo  Add to Chrome

Firefox logo Add to Firefox

TalkToModel: Understanding Machine Learning Models With Open Ended Dialogues


Jul 08, 2022
Dylan Slack, Satyapriya Krishna, Himabindu Lakkaraju, Sameer Singh

* Pre-print; comments welcome! Reach out to [email protected] 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

SAFER: Data-Efficient and Safe Reinforcement Learning via Skill Acquisition


Feb 10, 2022
Dylan Slack, Yinlam Chow, Bo Dai, Nevan Wichers


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Rethinking Explainability as a Dialogue: A Practitioner's Perspective


Feb 03, 2022
Himabindu Lakkaraju, Dylan Slack, Yuxin Chen, Chenhao Tan, Sameer Singh


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Feature Attributions and Counterfactual Explanations Can Be Manipulated


Jun 25, 2021
Dylan Slack, Sophie Hilgard, Sameer Singh, Himabindu Lakkaraju

* arXiv admin note: text overlap with arXiv:2106.02666 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

On the Lack of Robust Interpretability of Neural Text Classifiers


Jun 08, 2021
Muhammad Bilal Zafar, Michele Donini, Dylan Slack, Cédric Archambeau, Sanjiv Das, Krishnaram Kenthapadi

* Appearing at ACL Findings 2021 

   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Counterfactual Explanations Can Be Manipulated


Jun 04, 2021
Dylan Slack, Sophie Hilgard, Himabindu Lakkaraju, Sameer Singh


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Defuse: Harnessing Unrestricted Adversarial Examples for Debugging Models Beyond Test Accuracy


Feb 11, 2021
Dylan Slack, Nathalie Rauschmayr, Krishnaram Kenthapadi


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

Differentially Private Language Models Benefit from Public Pre-training


Sep 13, 2020
Gavin Kerrigan, Dylan Slack, Jens Tuyls


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email

How Much Should I Trust You? Modeling Uncertainty of Black Box Explanations


Aug 11, 2020
Dylan Slack, Sophie Hilgard, Sameer Singh, Himabindu Lakkaraju


   Access Paper or Ask Questions

  • Share via Twitter
  • Share via Facebook
  • Share via LinkedIn
  • Share via Whatsapp
  • Share via Messenger
  • Share via Email
1
2
>>