About
Aleksandra Faust is a Director of Research at Google DeepMind. Her research is centered around safe and scalable autonomous systems for social good, including reinforcement learning, planning, and control for robotics, autonomous driving, and digital assistants. Previously, Aleksandra co-founded Reinforcement Learning Research in Google Brain, founded Task and Motion Planning research in Robotics at Google, and machine learning for self-driving car planning and controls in Waymo, and was a senior researcher in Sandia National Laboratories. She earned a Ph.D. in Computer Science at the University of New Mexico with distinction, and a Master's in Computer Science from the University of Illinois at Urbana-Champaign. Aleksandra won the IEEE RAS Early Career Award for Industry, the Tom L. Popejoy Award for the best doctoral dissertation at the University of New Mexico in the period of 2011-2014, and was named Distinguished Alumna by the University of New Mexico School of Engineering. Her work has been featured in the New York Times, PC Magazine, ZdNet, VentureBeat, and was awarded Best Paper in Service Robotics at ICRA 2018, Best Paper in Reinforcement Learning for Real Life (RL4RL) at ICML 2019, Best Paper of IEEE Computer Architecture Letters in 2020, and IEEE Micro Top Picks 2023 Honorable Mention.
Latest News
All newsSep 2024 - “Many-Shot In-Context Learning," selected for Spotlight at NeurIPS 2024.
July 2024 - “Stop Regressing: The Unreasonable Effectiveness of Classification in Deep Reinforcement Learning,” selected for Oral at ICML 2024 - 1.5% acceptance rate.
July 2024 - “Levels of AGI: Operationalizing Progress on the Path to AGI,” Positional paper selected for Spotlight at ICML 2024 - 3.5% acceptance rate.
May 2024 - “A Real-World WebAgent with Planning, Long Context Understanding, and Program Synthesis,” selected for Oral at ICLR 2024 – 1.17% acceptance rate.
March 2024 - Named as 50 Women in Robotics you need to know about, Women in Robotics.
Sep 2023 - Program Chair for AutoML 2023
Jan 2023 - Selected for IEEE RA-L Senior Editor for Robot Learning
Jan 2023 - “Automatic Domain-Specific SoC Design for Autonomous Unmanned Aerial Vehicles,” S Krishnan, et al., MICRO 2022, receives IEEE Micro Top Picks 2023 Honorable Mention
Dec 2022 - CoRL Keynote
Oct 2022 - IROS Keynote
Jan 2021 - Elected to serve on the IEEE Robotics and Automation Society (RAS) Administrative board.
Dec 2020 - The Sky Is Not the Limit: A Visual Performance Model for Cyber-Physical Co-Design in Autonomous Machines, selected as Best of CAL in 2020.
Dec 2020 - Program Chair for CoRL 2021
Current projects
Full list of publicationsGemini Self-Improvement, Project Lead (2023 - present)
Led a ~20 person global research team to develop methods for autonomous and continuous base model improvement, resulting in the first Gemini model to outperform GPT4 [T6, W28]. Landed in Gemini Ultra 1.0, Gemini 1.5 Pro, and Cloud Next and Google I/O 2024.
Autonomous Agents Research, PI / Technical Lead (2020 - present)
Spearheaded a systematic inquiry into autonomy at scale, observing that autonomy concepts are valid beyond autonomous driving and robotics. Developed methodology where reinforcement, self-supervision, and learned curriculum work together to train agents across domains and applications by building up from simple to more complex skills.
Developed theoretical foundations for multi-task generalization [C20], and a learned curriculum with generative environments [C20, C26] to enable generalization of a very high number of tasks (~10120 tasks). Reduced Google Assistant Automation operational cost by 30%, delivering Google Password Manager, Food Ordering, and Shopping automation on 600+ domains.
To enable truly generalist agents on the web, addressing context window and action space limitations, developed foundation models for understanding HTML (C37) and self-improving code-as-actions text (C39) and multimodal (C38) web agents. Launched in Bard Extensions for @Chome and @YouTube.
To pave the steps toward the foundation sequential decision making models, helped develop methods that combine imitation learning and RL (C34), and stabilize RL training (C40), and brought the ideas from levels of autonomy to AGI to enable structured analysis of the AGI (C41).
Past Projects
Reinforcement Learning On-Edge (2017 - 2023)
- Accelerator / Neural network co-design.
Srivatsan Krishnan, Natasha Jaques, Shayegan Omidshafiei, Dan Zhang, Izzeddin Gur, Vijay Janapa Reddi, Aleksandra Faust, “Multi-Agent Reinforcement Learning for Microprocessor Design Space Exploration,” MLSys@NeurIPS 2022. Oral.
Srivatsan Krishnan, Zishen Wan, Kshitij Bhardwaj, Paul Whatmough, Aleksandra Faust, Sabrina Neuman, Gu-Yeon Wei, David Brooks, Vijay Janapa Reddi, “Automatic Domain-Specific SoC Design for Autonomous Unmanned Aerial Vehicles,” MICRO 2022
Srivatsan Krishnan , Zishen Wan, Kshitij Bhardwaj, Ninad Jadhav,. Aleksandra Faust, Vijay Janapa Reddi, "Roofline Model for UAVs: A Visual Performance Model for Guiding Compute System Design in Autonomous Drones," IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS), 2022
Srivatsan Krishnan, Zishen Wan, Kshitij Bhardwaj, Paul Whatmough, Aleksandra Faust, Gu-Yeon Wei, David Brooks, Vijay Janapa Reddi, "The Sky Is Not the Limit: A Visual Performance Model for Cyber-Physical Co-Design in Autonomous Machines", CAL 2020. (Best of CAL in 2020)
Srivatsan Krishnan, Sharad Chitlangia, Maximilian Lam, Zishen Lam, Aleksandra Faust, Vijay Janapa Reddi, "Quantized Reinforcement Learning (QUARL)" 1st Workshop on Resource-Constrained Machine Learning. Medium (Arxiv, GitHub)
- TinyRL - autonomous navigation on nano drone.
Tiny Robot Learning: Challenges and Directions for Machine Learning in Resource-Constrained Robots, Sabrina M. Neuman, Brian Plancher, Bart Duisterhof, Srivatsan Krishnan, Colby Banbury, Mark Mazumder, Shvetank Prakash, Jason Jabbour, Aleksandra Faust, Guido de Croon, Vijay Janapa Reddi, IEEE AICAS special session on Low Power Autonomous Systems 2022.
Tiny Robot Learning (tinyRL) for Source Seeking on a Nano Quadcopter, Bardienus P. Duisterhof, Srivatsan Krishnan, Jonathan J. Cruz, Colby R. Banbury, William Fu, Aleksandra Faust, Guido C. H. E. de Croon, Vijay Janapa Reddi, ICRA, 2021
Learning to Seek: Autonomous Source Seeking with Deep Reinforcement Learning Onboard a Nano Drone Microcontroller, Bardienus P. Duisterhof, Srivatsan Krishnan, Jonathan J. Cruz, Colby R. Banbury, William Fu, Aleksandra Faust, Guido C. H. E. de Croon, Vijay Janapa Reddi. BitCraze blog (Arxiv, Video, GitHub), 2019
- Hardware in the loop benchmarking for nano-drones
Behzad Boroujerdian, Hasan Genc, Srivatsan Krishnan, Bardienus Pieter Duisterhof, Brian Plancher, Kayvan Mansoorshahi, Marcelino Almeida, Aleksandra Faust, Vijay Janapa Reddi, "The Role of Compute in Autonomous Micro Aerial Vehicles: Optimizing for Flight Time and Energy Efficiency," ACM Transactions on Computer Systems (TOCS), 2022
Srivatsan Krishnan, Behzad Borojerdian, William Fu, Aleksandra Faust, Vijay Janapa Reddi, "Air Learning: a deep reinforcement learning gym for autonomous aerial robot visual navigation" Mach Learn 110, 2501–2540 (2021). https://doi.org/10.1007/s10994-021-06006-6
Behzad Boroujerdian, Hasan Genc, Srivatsan Krishnan, Wenzhi Cui, Aleksandra Faust, Vijay Janapa Reddi, "MAVBench: Micro Aerial Vehicle Benchmarking,” 51st IEEE/ACM International Symposium on Microarchitecture (MICRO), 2018 21% acceptance rate. (Pdf, BibTex, Video)
Learning to Learn for Reinforcement Learning (AutoRL) (2017 - 2023)
Jack Parker-Holder, Raghu Rajan, Xingyou Song, André Biedenkapp, Yingjie Miao, Theresa Eimer, Baohe Zhang, Vu Nguyen, Roberto Calandra, Aleksandra Faust, Frank Hutter, Marius Lindauer, "Automated Reinforcement Learning (AutoRL): A Survey and Open Problems," JAIR 2022.
Yingjie Miao, Xingyou Song, John D Co-Reyes, Daiyi Peng, Summer Yue, Eugene Brevdo, Aleksandra Faust, "Differentiable Architecture Search for Reinforcement Learning" AutoML, 2022. (19% acceptance rate).
- Learning RL loss functions
Juan Jose Garau-Luis, Yingjie Miao, John D. Co-Reyes, Aaron Parisi, Jie Tan, Esteban Real, Aleksandra Faust, "Multi-Objective Evolution for Generalizable Policy Gradient Algorithms,"Generalizable Policy Learning in the Physical World Workshop @ ICLR 2022.
John D. Co-Reyes, Yingjie Miao, Daiyi Peng, Esteban Real, Sergey Levine, Quoc Le, Honglak Lee, Aleksandra Faust, "Evolving Reinforcement Learning Algorithms," Oral at ICLR 2021 and DeepRL@NeurIPS 2020. (<2% acceptance rate, mentioned in Google AI Year in Review, Analytics India Magazine).
- Automated curriculum learning
Abdus Salam Azad, Izzeddin Gur, Pieter Abbeel, Ion Stoica, Aleksandra Faust, CLUTR: Curriculum Learning via Unsupervised Task Representation Learning, CoRR 2022
Izzeddin Gur, Natasha Jaques, Yingjie Miao, Jongwook Choi, Manoj Tiwari, Honglak Lee, Aleksandra Faust, "Environment Generation for Zero-Shot Compositional Reinforcement Learning," Advances in Neural Information Processing Systems (NeurIPS), 2021
Izzeddin Gur, Ofir Nachum, Aleksandra Faust, "Targeted Environment Design from Offline Data," Deep RL @ NeurIPS, 2021
- Learning rewards
2019 - Evolving Rewards to Automate Reinforcement Learning, Aleksandra Faust, Anthony Francis, Dar Mehta, 6th AutoML@ICML. (BibTex) 58% acceptance rate, Mentioned in Jeff Dean's keynote. Press: [1].
2019 - Learning Navigation Behaviors End to End with AutoRL, Hao-Tien Chiang, Aleksandra Faust, Marek Fiser, Antony Francis, IEEE Robotics and Automation Letters, vol. 4, no. 2, pp. 2007-2014, April 2019. Blog, Press: [1], [2], [3], [4], [5]. #1 or #2 (out of ~1600) most downloaded paper in Feb-Nov 2020. (BibTex, Video)
Generalization in RL (2019 - 2023)
- Learning compositional tasks
Sungryull Sohn, Hyunjae Woo, Jongwook Choi, lyubing Qiang, Izzeddin Gur, Aleksandra Faust, Honglak Lee, "Fast Inference and Transfer of Compositional Task Structures for Few-shot Task Generalization," Oral @ Uncertainty in Artificial Intelligence (UAI) 2022. (5% acceptance rate)
Michael Lim, Andy Zeng, Brian Ichter, Maryam Bandari, Erwin Coumans, Claire Tomlin, Stefan Schaal, Aleksandra Faust, "Multi-Task Learning with Sequence-Conditioned Transporter Networks," ICRA 2022.
Izzeddin Gur, Natasha Jaques, Yingjie Miao, Jongwook Choi, Manoj Tiwari, Honglak Lee, Aleksandra Faust, "Environment Generation for Zero-Shot Compositional Reinforcement Learning," Advances in Neural Information Processing Systems (NeurIPS), 2021
Izzeddin Gur, Ofir Nachum, Aleksandra Faust, "Targeted Environment Design from Offline Data," Deep RL @ NeurIPS, 2021
Izzeddin Gur, Ofir Nachum, Yingjie Miao, Mustafa Safdari, Austin Huang, Sharan Narang, Aakanksha Chowdhery, Noah Fiedel, Aleksandra Faust, Understanding HTML with Large Language Models, CoRR 2022
Su Wang, Ceslee Montgomery, Jordi Orbay, Vighnesh Birodkar, Aleksandra Faust, Izzeddin Gur, Natasha Jaques, Austin Waters, Jason Baldridge, Peter Anderson, "Less is More: Generating Grounded Navigation Instructions from Landmarks," CVPR 2022
Izzeddin Gur, Ulrich Rueckert, Aleksandra Faust, Dilek Hakkani-Tur, "Learning to Navigate the Web," ICLR 2019. ZdNet article; Medium, Top 10%, 30% acceptance rate. (Pdf, BibTex)
Safe reinforcement learning (2018 - 2020)
2020 - Safe Policy Learning for Continuous Control, Yinlam Chow, Ofir Nachum, Aleksandra Faust, Mohammad Ghavamzadeh, Edgar Duenez-Guzman, CoRL. (Arxiv) Mentioned in Google AI Year in Review. RL4RL@ICML 2019. Best paper award.
2019 - Comparison of Deep Reinforcement Learning Policies to Formal Methods for Moving Obstacle Avoidance, Arpit Garg, Hao-Tien Lewis Chiang, Satomi Sugaya, Aleksandra Faust, Lydia Tapia, to appear at IROS
2018 - Resilient Computing with Reinforcement Learning on a Dynamical System: Case Study in Sorting, Aleksandra Faust, James B. Aimone, Conrad D. James, Lydia Tapia, 57th IEEE Conference on Decision and Control ZdNet Article. (Pdf, BibTex)
Self-supervision in planning (2019 - 2021)
- Critical Probabilistic Roadmaps
2021 -"Avoidance Critical Probabilistic Roadmaps for Motion Planning in Dynamic Environments," Felipe Felix Arias, Brian Andrew Ichter, Aleksandra Faust, Nancy M. Amato, ICRA, 2021.
2020 - Learned Critical Probabilistic Roadmaps for Robotic Motion Planning,Brian Ichter, Edward Schmerling, Tsang-Wei Edward Lee, Aleksandra Faust, ICRA (Video)
2020 - Model-based Reinforcement Learning for Multiagent Goal Alignment, Rose E. Wang, J. Chase Kew, Dennis Lee, Brian Ichter, Tsang-Wei Edward, Lee, Tingnan Zhang, Jie Tan, Aleksandra Faust, CoRL, website, video. Mentioned in Google AI Year in Review.
2020 - Neural Collision Clearance Estimator for Batched Robot Motion Planning, J. Chase Kew, Brian Ichter, Maryam Bandari, Tsang-Wei Edward Lee, Aleksandra Faust, WAFR. (Arxiv)
2019 - RL-RRT: End-to-End Kinodynamic Robot Motion Planning, Hao-Tien Lewis Chiang, Jasmine Hsu, Marek Fiser, Lydia Tapia, Aleksandra Faust, IEEE Robotics and Automation Letters, vol. 4, no. 4, pp. 4298-4305, Oct. 2019. (Video) #17 out of ~1600 most downloaded paper in Sep 2019.
- Swept Volume Estimation
2020 - Fast Deep Swept Volume Estimator, Hao-Tien Lewis Chiang, John E. G. Baxter, Satomi Sugaya, Mohammad R. Yousefi, Aleksandra Faust, Lydia Tapia, The International Journal of Robotics Research (IJRR).
2018 - Fast Swept Volume Estimation with Deep Learning, Hao-Tien Chiang, Aleksandra Faust, Satomi Sugaya and Lydia Tapia, WAFR Mentioned in Looking Back at Google’s Research Efforts in 2018. (Pdf, BibTex)
2018 - A Deep Neural Network for Swept Volume Prediction Between Configurations, Hao-Tien Chiang, Aleksandra Faust, Lydia Tapia, 3rd MLPC at ICRA, (Pdf, Bibtex).
Learning complex skills with hierarchical planning (2017 - 2022)
2022 - Michael Lim, Andy Zeng, Brian Ichter, Maryam Bandari, Erwin Coumans, Claire Tomlin, Stefan Schaal, Aleksandra Faust, "Multi-Task Learning with Sequence-Conditioned Transporter Networks," ICRA 2022
2021- Visual Navigation Among Humans with OptimalControl as a Supervisor, RA-L 2021, Arxiv, Website, Code. Press: VentureBeat, Techxplore.
2020 - Zero-shot Imitation Learning from Demonstrations for Legged Robot Visual Navigation, Xinlei Pan, Tingnan Zhang, Brian Ichter, Aleksandra Faust, Jie Tan, Sehoon Ha, ICRA 2022 (Arxiv)
2020 - Long-Range Indoor Navigation with PRM-RL, Anthony Francis, Aleksandra Faust, Hao-Tien Lewis Chiang, Jasmine Hsu, J. Chase Kew, Marek Fiser, Tsang-Wei Edward Lee, T-RO 2020. (Citation, Arxiv, Video) Blog, Press: [1], [2], [3], [4], [5].
2018 - PRM-RL: Long-range Robotic Navigation Tasks by Combining Reinforcement Learning and Sampling-based Planning, Aleksandra Faust, Oscar Ramirez, Marek Fiser, Kenneth Oslund, Anthony Francis, James Davidson, Lydia Tapia, ICRA. Best paper in Service Robotics; Mentioned in Looking Back at Google’s Research Efforts in 2018, Blog. (pdf, Bibtex, Video)
2018 - FollowNet: Towards Robot Navigation by Following Natural Language Directions with Deep Reinforcement Learning, Pararth Shah, Marek Fiser, Aleksandra Faust, J. Chase Kew, Dilek Hakkani-Tur, 3rd MLPC at ICRA, May 2018 (Pdf, BibTex)