Neel Guha
Ph.D. Student in Computer Science, admitted Summer 2021
Juris Doctor Student, Law
All Publications
-
Prospector Heads: Generalized Feature Attribution for Large Models & Data.
ArXiv
2024
Abstract
Feature attribution, the ability to localize regions of the input data that are relevant for classification, is an important capability for ML models in scientific and biomedical domains. Current methods for feature attribution, which rely on "explaining" the predictions of end-to-end classifiers, suffer from imprecise feature localization and are inadequate for use with small sample sizes and high-dimensional datasets due to computational challenges. We introduce prospector heads, an efficient and interpretable alternative to explanation-based attribution methods that can be applied to any encoder and any data modality. Prospector heads generalize across modalities through experiments on sequences (text), images (pathology), and graphs (protein structures), outperforming baseline attribution methods by up to 26.3 points in mean localization AUPRC. We also demonstrate how prospector heads enable improved interpretation and discovery of class-specific patterns in input data. Through their high performance, flexibility, and generalizability, prospectors provide a framework for improving trust and transparency for ML models in complex domains.
View details for PubMedID 38947933
View details for PubMedCentralID PMC11213143
-
Understanding Liability Risk from Using Health Care Artificial Intelligence Tools.
The New England journal of medicine
2024; 390 (3): 271-278
View details for DOI 10.1056/NEJMhle2308901
View details for PubMedID 38231630
-
PRIVATE ENFORCEMENT IN THE STATES
UNIVERSITY OF PENNSYLVANIA LAW REVIEW
2023; 172 (1)
View details for Web of Science ID 001208959900003
-
ChatGPT and Physicians' Malpractice Risk.
JAMA health forum
2023; 4 (5): e231938
Abstract
This JAMA Forum discusses the possibilities, limitations, and risks of physician use of large language models (such as ChatGPT) along with the improvements required to improve the accuracy of the technology.
View details for DOI 10.1001/jamahealthforum.2023.1938
View details for PubMedID 37200013
-
Gamesmanship in Modern Discovery Tech
LEGAL TECH AND THE FUTURE OF CIVIL JUSTICE
2023: 112-132
View details for Web of Science ID 001003618200006
-
Don't Use a Cannon to Kill a Fly: An Efficient Cascading Pipeline for Long Documents
ASSOC COMPUTING MACHINERY. 2023: 141-147
View details for DOI 10.1145/3594536.3595142
View details for Web of Science ID 001139079400015
-
Embroid: Unsupervised Prediction Smoothing Can Improve Few-Shot Classification
NEURAL INFORMATION PROCESSING SYSTEMS (NIPS). 2023
View details for Web of Science ID 001220818804023
-
LEGALBENCH: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models
NEURAL INFORMATION PROCESSING SYSTEMS (NIPS). 2023
View details for Web of Science ID 001226352803033