Chelsea Finn

Chelsea Finn

Auto-generated profile, based on publicly available information

Chelsea Finn

Stanford University

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 March 2021 in IEEE Robotics and Automation Letters
Verified
Journal Article
Closed access
BTBrijen Thananjeyan
Brijen Thananjeyan
ABAshwin Balakrishna
Ashwin Balakrishna
SNSuraj Nair
Suraj Nair
MLMichael Luo
Michael Luo
KSKrishnan Srinivasan
Krishnan Srinivasan
KGKen Goldberg
Ken Goldberg
Safety remains a central obstacle preventing widespread use of RL in the real world: learning new tasks in uncertain environments requires extensive exploration, but safety requires limiting exploration. We propose Recovery RL, an algorithm which navigates this tradeoff by (1) leveraging offline data to learn about constraint violating zones before policy learning and (2) separating the goals of improving task performance and constraint satisfaction across two policies: a task policy that only optimizes the task reward and a recovery policy that guides the agent to safety when constraint violation is likely. We evaluate Recovery RL on 6 simulation domains, including two contact-rich manipulation tasks and an image-based navigation task, and an image-based obstacle avoidance task on a physical robot. We compare Recovery RL to 5 prior safe RL methods which jointly optimize for task performance and safety via constrained optimization or reward shaping and find that Recovery RL outperforms the next best prior method across all domains. Results suggest that Recovery RL trades off constraint violations and task successes 2–20 times more efficiently in simulation domains and 3 times more efficiently in physical experiments. See https://tinyurl.com/rl-recovery for videos and supplementary material.

1 January 2024
Verified
Conference Paper
Open access
PDF file
RPRyan Park
Ryan Park
RRRafael Rafailov
Rafael Rafailov
SEStefano Ermon
Stefano Ermon
CFChelsea Finn
Chelsea Finn
Ryan Park, Rafael Rafailov, Stefano Ermon, Chelsea Finn. Findings of the Association for Computational Linguistics ACL 2024. 2024.

10 July 2023
Verified
Conference Paper
Open access
AKAviral Kumar
Aviral Kumar
ASAnikait Singh
Anikait Singh
FEFrederik Ebert
Frederik Ebert
MNMitsuhiko Nakamoto
Mitsuhiko Nakamoto
YYYanlai Yang
Yanlai Yang
SLSergey Levine
Sergey Levine

1 April 2015 in Pattern Recognition
Verified
Journal Article
Closed access
JDJames Duyck
James Duyck
CFChelsea Finn
Chelsea Finn
AHAndy Hutcheon
Andy Hutcheon
PVPablo Vera
Pablo Vera
JSJoaquin Salas
Joaquin Salas
SRSai Ravela
Sai Ravela

8 August 2023
Verified
Conference Paper
Open access
PHPeter Henderson
Peter Henderson
EMEric Mitchell
Eric Mitchell
CMChristopher Manning
Christopher Manning
DJDan Jurafsky
Dan Jurafsky
CFChelsea Finn
Chelsea Finn
A growing ecosystem of large, open-source foundation models has reduced the labeled data and technical expertise necessary to apply machine learning to many new problems. Yet foundation models pose a clear dual-use risk, indiscriminately reducing the costs of building both harmful and beneficial machine learning systems. Policy tools such as restricted model access and export controls are the primary methods currently used to mitigate such dual-use risks. In this work, we review potential safe-release strategies and argue that both policymakers and AI researchers would benefit from fundamentally new technologies enabling more precise control over the downstream usage of open-source foundation models. We propose one such approach: the task blocking paradigm, in which foundation models are trained with an additional mechanism to impede adaptation to harmful tasks without sacrificing performance on desirable tasks. We call the resulting models self-destructing models, inspired by mechanisms that prevent adversaries from using tools for harmful purposes. We present an algorithm for training self-destructing models leveraging techniques from meta-learning and adversarial learning, which we call meta-learned adversarial censoring (MLAC). In a small-scale experiment, we show MLAC can largely prevent a BERT-style model from being re-purposed to perform gender identification without harming the model’s ability to perform profession classification.

12 November 2020
Verified
Book Chapter
Closed access
PDF file
KSKarl Schmeckpeper
Karl Schmeckpeper
AXAnnie Xie
Annie Xie
OROleh Rybkin
Oleh Rybkin
STStephen Tian
Stephen Tian
KDKostas Daniilidis
Kostas Daniilidis
CFChelsea Finn
Chelsea Finn