
I recently completed my PhD at UC Berkeley in the Berkeley Artificial Research (BAIR) Lab and am now at OpenAI.
In my research, I explored machine learning for robotics, advised by Sergey Levine.
I also did undergrad at Berkeley, working with Pieter Abbeel and Sergey Levine on similar topics.
I am generally interested in how to develop intelligent behavior through interaction, especially through offline reinforcement learning and self-supervised reinforcement learning.
If you want to chat, shoot me an email at anair17berkeley.edu. Some of my projects are listed below.

all
Learning on the Job: Self-Rewarding Offline-to-Online Finetuning for Industrial Insertion of Novel Connectors from Vision. Ashvin Nair*, Brian Zhu*, Gokul Narayanan, Eugen Solowjow, Sergey Levine.
[
PDF]
[
Website]
Generalization with Lossy Affordances: Leveraging Broad Offline Data for Learning Visuomotor Tasks. Kuan Fang, Patrick Yin, Ashvin Nair, Homer Walke, Gengchen Yan, Sergey Levine. CoRL 2022.
[
PDF]
[
Website]
Planning to Practice: Efficient Online Fine-Tuning by Composing Goals in Latent Space. Kuan Fang*, Patrick Yin*, Ashvin Nair, Sergey Levine. IROS 2022.
[
PDF]
[
Website]
Bisimulation Makes Analogies in Goal-Conditioned Reinforcement Learning. Philippe Hansen-Estruch, Amy Zhang, Ashvin Nair, Patrick Yin, Sergey Levine. ICML 2022.
[
PDF]
[
Website]
Offline Reinforcement Learning with Implicit Q-Learning. Ilya Kostrikov, Ashvin Nair, Sergey Levine. ICLR 2022.
[
PDF]
[
Website]
Offline Meta-Reinforcement Learning with Online Self-Supervision. Vitchyr Pong, Ashvin Nair, Laura Smith, Catherine Huang, Sergey Levine. ICML 2022.
[
PDF]
[
Website]
What Can I Do Here? Learning New Skills by Imagining Visual Affordances. Alexander Khazatsky*, Ashvin Nair*, Daniel Jing, Sergey Levine. ICRA 2021.
[
PDF]
[
Website]
DisCo RL: Distribution-Conditioned Reinforcement Learning for General Purpose Policies. Soroush Nasiriany*, Vitchyr H. Pong*, Ashvin Nair*, Alexander Khazatsky, Glen Berseth, Sergey Levine. ICRA 2021.
[
PDF]
[
Website]
Accelerating Online Reinforcement Learning with Offline Datasets. Ashvin Nair, Murtaza Dalal, Abhishek Gupta, Sergey Levine.
[
PDF]
[
Website]
Meta-Reinforcement Learning for Robotic Industrial Insertion Tasks. Gerrit Schoettler, Ashvin Nair, Juan Aparicio Ojea, Sergey Levine, Eugen Solowjow. IROS 2020.
[
PDF]
[
Website]

Code for context-conditioned RIG released (available on project website)!
Jan 2020
Deep Reinforcement Learning for Industrial Insertion Tasks with Visual Inputs and Natural Rewards. Gerrit Schoettler*, Ashvin Nair*, Jianlan Luo, Shikhar Bahl, Juan Aparicio Ojea, Eugen Solowjow, Sergey Levine. IROS 2020.
[
PDF]
[
Website]
Skew-Fit: State-Covering Self-Supervised Reinforcement Learning. Vitchyr Pong*, Murtaza Dalal*, Steven Lin*, Ashvin Nair, Shikhar Bahl, Sergey Levine. ICML 2020.
[
PDF]
[
Website]
Contextual Imagined Goals for Self-Supervised Robotic Learning. Ashvin Nair*, Shikhar Bahl*, Alexander Khazatsky*, Vitchyr Pong, Glen Berseth, Sergey Levine. CoRL 2019.
[
PDF]
[
Website]

A couple of preprints of recent work on self-supervised exploration and industrial part insertion released now on arXiv and presented at ICML workshops.
Jul 2019
Residual Reinforcement Learning for Robot Control. Tobias Johannink*, Shikhar Bahl*, Ashvin Nair*, Jianlan Luo, Avinash Kumar, Matthias Loskyll, Juan Aparicio Ojea, Eugen Solowjow, Sergey Levine. ICRA 2019.
[
PDF]
[
Website]

Our work on visual RL with imagined goals (RIG) is released now on arXiv and was presented at ICML workshops on goal specification and lifelong learning!
Jul 2018
Visual Reinforcement Learning with Imagined Goals. Ashvin Nair*, Vitchyr Pong*, Murtaza Dalal, Shikhar Bahl, Steven Lin, Sergey Levine. NeurIPS 2018 (Spotlight).
[
PDF]
[
Website]

I presented recent work on combining demonstrations and RL at the Deep RL symposium and robot learning workshop at NIPS 2017.
Dec 2017
Overcoming Exploration in Reinforcement Learning with Demonstrations. Ashvin Nair, Bob McGrew, Marcin Andrychowicz, Wojciech Zaremba, Pieter Abbeel. ICRA 2018.
[
PDF]
[
Website]

Released our ICRA 2017 paper on arXiv along with data for the Baxter projects! Links to the data can be found on the project websites.
Mar 2017

"Vision-Based Rope Manipulation" accepted to ICRA 2017.
Jan 2017

Finalist for the Computing Research Association (CRA) Outstanding Undergraduate Researcher Award!
Dec 2016

Our work was featured in a MIT Technology Review article: "AI Begins to Understand the 3-D World".
Dec 2016

I presented our rope manipulation work at the "Intuitive Physics" workshop at NIPS 2016.
Dec 2016
Combining Self-Supervised Learning and Imitation for Vision-Based Rope Manipulation. Ashvin Nair*, Dian Chen*, Pulkit Agrawal*, Phillip Isola, Pieter Abbeel, Jitendra Malik, Sergey Levine. ICRA 2017.
[
PDF]
[
Website]

"Learning to Poke by Poking" accepted as Oral Presentation to NIPS 2016!
Aug 2016
Learning to Poke by Poking: Experiential Learning of Intuitive Physics. Pulkit Agrawal*, Ashvin Nair*, Pieter Abbeel, Jitendra Malik, Sergey Levine. NIPS 2016 (Oral).
[
PDF]
[
Website]

Robots!
robotics
/
vision
/
python
/
ros
Aug 2015

Wanderlust: Starting Out with Computer Vision
vision
/
web-dev
/
python
/
opencv
/
hack
/
flask
/
android
/
scrapy
/
mongodb
Dec 2014

Scrape All The Things
python
/
scrapy
Dec 2014

GradeSeer: Student Grade Prediction
web-dev
/
machine-learning
/
rails
/
angularjs
/
mongodb
/
python
Nov 2014

CalCentral Student Developing
web-dev
/
rails
/
angularjs
/
maps
Nov 2014

Intro to Machine Learning
machine-learning
/
python
/
neural-nets
Nov 2014

Mr Miyagi: On-demand Carwashes
hack
/
maps
/
rails
/
angularjs
Nov 2014

SplitWithMe: Payments App in Vegas
web-dev
/
hack
/
rails
/
angularjs
/
venmo
Nov 2014

JamWithMe: In-Browser Music Collaboration
web-dev
/
hack
/
rails
/
js
/
firebase
Jan 2014
news

Code for context-conditioned RIG released (available on project website)!
Jan 2020

A couple of preprints of recent work on self-supervised exploration and industrial part insertion released now on arXiv and presented at ICML workshops.
Jul 2019

Our work on visual RL with imagined goals (RIG) is released now on arXiv and was presented at ICML workshops on goal specification and lifelong learning!
Jul 2018

I presented recent work on combining demonstrations and RL at the Deep RL symposium and robot learning workshop at NIPS 2017.
Dec 2017

Released our ICRA 2017 paper on arXiv along with data for the Baxter projects! Links to the data can be found on the project websites.
Mar 2017

"Vision-Based Rope Manipulation" accepted to ICRA 2017.
Jan 2017

Finalist for the Computing Research Association (CRA) Outstanding Undergraduate Researcher Award!
Dec 2016

Our work was featured in a MIT Technology Review article: "AI Begins to Understand the 3-D World".
Dec 2016

I presented our rope manipulation work at the "Intuitive Physics" workshop at NIPS 2016.
Dec 2016

"Learning to Poke by Poking" accepted as Oral Presentation to NIPS 2016!
Aug 2016