Debugging Deep Model-based Reinforcement Learning Systems
Things I have learned in 3 years of a young, and generally tricky research field.
Nathan is a robotics researcher, non-professional athlete, writer, mental-health advocate and a team player.
Featured Posts:
Research Directions:
Built with Webflow
About me
Hello! I am a PhD Candidate at the University of California, Berkeley, Department of Electrical Engineering and Computer Sciences, advised by Professor Kristofer Pister in the Berkeley Autonomous Microsystems Lab, and pseudo-advised by Roberto Calandra at Facebook AI Research!
Hello! I am a PhD Candidate at the University of California, Berkeley studying the intersection of robotics and machine learning. I am a member of the Department of Electrical Engineering and Computer Sciences, advised by Professor Kristofer Pister in the Berkeley Autonomous Microsystems Lab, and pseudo-advised by Roberto Calandra at Facebook AI Research! I am actively involved in outreach and inclusion efforts and an advocate for mental health -- he is the EEGSA wellness chair and founder of the UC Berkeley Equal Access to Application Assistance program.
Prior to UC Berkeley, I was a proud member of Cornell Electrical and Computer Engineering 2017 where I learned to do research with the Lab of Plasma Studies and the SonicMEMs Lab. I bring my research foundation in hardware, models, and physics to the data-driven world of machine learning. At Cornell, I was a part of Cornell Lightweight Rowing. I did an internship with Tesla Motors Battery Engineering in 2015.
I am happy to be a product of The Ocean State.
Nathan Lambert is a PhD Candidate at the University of California, Berkeley working at the intersection of machine learning and robotics. He is a member Department of Electrical Engineering and Computer Sciences, advised by Professor Kristofer Pister in the Berkeley Autonomous Microsystems Lab. Nathan has worked extensively with Roberto Calandra at Facebook AI Research and is joining DeepMind Robotics remotely for the summer of 2021. During his Ph.D., he was awarded the UC Berkeley EECS Demetri Angelakos Memorial Achievement Award for Altruism.
- I am always looking to work with and promote under-represented groups in STEM fields. If you relate to this, please email me directly if you're curious in my work or trying to find a project in AI at UC Berkeley (I have mentored and advised multiple students through teaching at Berkeley).
- Online mentor at Polygence (currently only taking scholarship-based students from under-represented groups).
- Creator of Democratizing Automation newsletter & blog on making the future of AI and robotics equitable.
I like to try and have fun between my many projects. You can find me on Strava, I also happen to be a brand ambassador for Picky Bars. I actively track my health, cook, and read (recipe and book pages in construction).
News
- Summer 2021: I will be interning with Martin Riedmiller's team at DeepMind.
- April 2021 Update: Research Foundations.
- 27 April 2021: I was awarded the UC Berkeley EECS Demetri Angelakos Memorial Achievement Award for Altruism. [link]
- 25 April 2021: I had another paper on sociotechnics in AI published in IEEE Transactions on Technology and Society. [link]
- 21 April 2021: We released an open-source library for model-based reinforcement learning. [link]
- March 2021 Update: Building!
- 5 March 2021: I was on the TalkRL Podcast covering my research and interests. [link]
- 4 March 2021: I gave a talk on Model-based RL at Cornell Robotics Seminar. [video] [slides]
- 18 Feb 2021: Some paper updates: 1 was accepted to IEEE Symposium on Technology and Society (here) and 1 more to AISTATS 2021 (here).
- Feb 2021 Update: Was lucky to be home!
- 1 Feb 2021: I made a new mental health page for researchers and Berkeley students.
- 26 Jan 2021: I appeared on the Software Engineering Daily Podcast talking about reinforcement learning and robotics!
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Scroll for some of my content
I think about...
Robotics
Intelligent & novel devices to interact with the physical world.
Machine Learning
The science of using data to decide in the presence of uncertainty.
Society
Making sure the stakeholders of automation are in the conversation.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Blogos
All grad students (should) study graphic design
People judge your papers by their cover. You can trick them into believing your science with pretty pictures.
Exploitation Exploration (in MBRL)
A few lessons from model-based reinforcement learning how exploration can happen through exploitation of some metric.
Papers
Axes for Sociotechnical Inquiry in AI Research
We present a concise set of directions for understanding the societal risks of new directions of AI research.
MBRL-Lib: A Modular Library for Model-based Reinforcement Learning
An open-source PyTorch repository designed from the bottom up for model-based reinforcement learning research.
On the Importance of Hyperparameter Optimization for Model-based Reinforcement Learning
We showed that advancements in AutoML when paired with common deep RL tasks, MBRL algorithms perform so well they break the simulator.
AI Development for the Public Interest: From Abstraction Traps to Sociotechnical Risks
We study three developing subfields of AI research and their growing relationship with the sociotechnical: AI Safety, Fair Machine Learning, and Human-in-the-loop Autonomy.