Timezone: »

Quantifying Uncertainty in Foundation Models via Ensembles
Meiqi Sun · Wilson Yan · Pieter Abbeel · Igor Mordatch
Event URL: https://openreview.net/forum?id=LpBlkATV24M »

As large-scale foundation models begin to have increasing impact in real-world applications, to guarantee reliability and trustworthiness it is important for these models to "know what they don't know": to be capable of quantifying uncertainty about their own outputs. In this work, we propose disagreement of model ensembles as an effective and compute-efficient method to quantify uncertainty. We also conduct a systematic study of uncertainty quantification spanning multiple tasks - a synthetic string task, and natural language arithmetic and question-answering tasks - over a progression of increasingly out of distribution inputs. We find that considering ensemble disagreement results in improved uncertainty prediction over only considering a single model's likelihood. We hope that our investigation and results encourage more research in the area of uncertainty quantification in foundation models and the use of model ensembles.

Author Information

Meiqi Sun (University of California, Berkeley)
Meiqi Sun

Hi! My name is Meiqi Sun, and I am an undergraduate at Berkeley studying Computer Science, Economics, and Statistics. I am also a member of Berkeley Artificial Intelligence Research Lab.

Wilson Yan (UC Berkeley)
Pieter Abbeel (UC Berkeley & Covariant)

Pieter Abbeel is Professor and Director of the Robot Learning Lab at UC Berkeley [2008- ], Co-Director of the Berkeley AI Research (BAIR) Lab, Co-Founder of covariant.ai [2017- ], Co-Founder of Gradescope [2014- ], Advisor to OpenAI, Founding Faculty Partner AI@TheHouse venture fund, Advisor to many AI/Robotics start-ups. He works in machine learning and robotics. In particular his research focuses on making robots learn from people (apprenticeship learning), how to make robots learn through their own trial and error (reinforcement learning), and how to speed up skill acquisition through learning-to-learn (meta-learning). His robots have learned advanced helicopter aerobatics, knot-tying, basic assembly, organizing laundry, locomotion, and vision-based robotic manipulation. He has won numerous awards, including best paper awards at ICML, NIPS and ICRA, early career awards from NSF, Darpa, ONR, AFOSR, Sloan, TR35, IEEE, and the Presidential Early Career Award for Scientists and Engineers (PECASE). Pieter's work is frequently featured in the popular press, including New York Times, BBC, Bloomberg, Wall Street Journal, Wired, Forbes, Tech Review, NPR.

Igor Mordatch (Google)

More from the Same Authors