Dylan Slack
dslack@uci.edu

Hello! I am a machine learning researcher. Currently, I'm a research scientist at Google, where I work on Gemini. Previously, I recieved a Ph.D. from UC Irvine advised by Sameer Singh and Hima Lakkaraju. My Ph.D. was generously supported by an HPI fellowship. I interned at AWS in 2020 and Google AI in 2021 👨‍💻.

CV /  Google Scholar /  Github /  Twitter /  Notes

profile photo
Research

Please find my research works here. * Denotes equal contribution.

Learning Goal-Conditioned Representations for Language Reward Models
Vaskar Nath*, Dylan Slack*, Jeff Da, Yuntao Ma, Hugh Zhang, Spencer Whitehead Sean Hendryx
NeurIPS, 2024  
arXiv / code

A Careful Examination of Large Language Model Performance on Grade School Arithmetic
Hugh Zhang, Jeff Da, Dean Lee, Vaughn Robinson, Catherine Wu, Will Song, Tiffany Zhao, Pranav Raja, Dylan Slack, Qin Lyu, Sean Hendryx, Russell Kaplan, Summer Yue
NeurIPS D&B, 2024  
arXiv

TalkToModel: Explaining Machine Learning Models with Interactive Natural Language Conversations
Dylan Slack, Satyapriya Krishna, Hima Lakkaraju*, Sameer Singh*
Nature Machine Intelligence, 2023  
TSRML @ NeurIPS, 2022   Honoral Mention Outstanding Paper
Nature Link / arXiv / bibtex / code / demo

Post Hoc Explanations of Language Models Can Improve Language Models
Satyapriya Krishna, Jiaqi Ma, Dylan Slack, Asma Ghandeharioun, Sameer Singh, Himabindu Lakkaraju,
NeurIPS, 2023
arXiv

TABLET: Learning From Instructions For Tabular Data
Dylan Slack, Sameer Singh
arXiv, 2023
arXiv / bibtex / code / demo

Rethinking Explainability as a Dialogue: A Practitioner's Perspective
Hima Lakkaraju*, Dylan Slack*, Yuxin Chen, Chenhao Tan, and Sameer Singh
HCAI @ NeurIPS, 2022
arXiv / bibtex

SAFER: Data-Efficient and Safe Reinforcement Learning via Skill Acquisition
Dylan Slack, Yinlam Chow, Bo Dai, and Nevan Wichers
DARL @ ICML, 2022
arXiv / bibtex

Active Meta-Learning for Predicting and Selecting Perovskite Crystallization Experiments
Venkateswaran Shekar, Gareth Nicholas, Mansoor Ani Najeeb, Margaret Zeile, Vincent Yu, Xiaorong Wang, Dylan Slack, Zhi Li, Philip Nega, Emory Chan, Alexander Norquist, Joshua Schrier, and Sorelle Friedler
Journal of Chemical Physics, 2022

Reliable Post hoc Explanations: Modeling Uncertainty in Explainability
Dylan Slack, Sophie Hilgard, Sameer Singh, and Hima Lakkaraju
NeurIPS, 2021
arXiv / bibtex / Project Links

Counterfactual Explanations Can Be Manipulated
Dylan Slack, Sophie Hilgard, Hima Lakkaraju, and Sameer Singh
NeurIPS, 2021
arXiv / bibtex / Project Links

On the Lack of Robustness of Neural Text Classifier Interpretations
Muhammad Bilal Zafar, Michele Donini, Dylan Slack, Cedric Archambeau, Sanjiv Das, and Krishnaram Kenthapadi
Findings of ACL, 2021
arXiv / bibtex

Defuse: Training More Robust Models through Creation and Correction of Novel Model Errors
Dylan Slack, Nathalie Rauschmayr, and Krishnaram Kenthapadi
XAI 4 Debugging Workshop, NeurIPS, 2021
Paper Link / Code

Context, Language Modeling, and Multimodal Data in Finance
Sanjiv Ranjan Das, Connor Goggins, John He, George Karypis, Sandeep Krishnamurthy, Mitali Mahajan, Nagpurnanand Prabhala, Dylan Slack, Robert Van Dusen, Shenghua Yue, Sheng Zha, and Shuai Zheng
The Journal of Financial Data Science, 2021
Journal Link / bibtex

Fooling LIME and SHAP: Adversarial Attacks on Post hoc Explanation Methods
Dylan Slack*, Sophie Hilgard*, Emily Jia, Sameer Singh, and Hima Lakkaraju
AIES, 2020   (Oral Presentation)
Work also presented at SafeAI Workshop, AAAI, 2020
code / video / arXiv / bibtex
Press: Deeplearning.ai / Harvard Business Review

Differentially Private Language Models Benefit from Public Pre-training
Gavin Kerrigan*, Dylan Slack*, and Jens Tuyls*
EMNLP PrivateNLP Workshop, 2020
code / arXiv / bibtex

Fairness Warnings and Fair-MAML: Learning Fairly with Minimal Data
Dylan Slack, Sorelle Friedler, Emile Givental,
FAccT, 2020  
Work also presented at HCML Workshop, NeurIPS, 2019
code / video / arXiv / bibtex

Assessing the Local Interpretability of Machine Learning Models
Dylan Slack, Sorelle Friedler, Carlos Scheidegger, and Chitradeep Dutta Roy
HCML Workshop, NeurIPS, 2019
arXiv / bibtex

Recent Talks
bayes-lime-shap

Speaking to the MedAI Group at Stanford

bayes-lime-shap

Presenting Reliable Post hoc Explanations: Modeling Uncertainty in Explainability

cfes-manipulated

Presenting Counterfactual Explanations Can Be Manipulated

aisc-talk

Speaking at AISC, virtually.

facct-talk

Speaking at FAccT, in Barcelona, Spain.


Source modified from here.