Stanford Seminar - Learning-enabled Adaptation to Evolving Conditions for Robotics
May 3, 2024
Somrita Banerjee, Stanford University
With advancements in machine learning and artificial intelligence, a new generation of “learning-enabled” robots is emerging, which are better suited to operating autonomously in unstructured, uncertain, and unforgiving environments. To achieve these goals, robots must be able to adapt to evolving conditions that are different from those seen during training or expected during deployment. In this talk I will first talk about adapting to novel instantiations, i.e., different task instances with shared structure, through parameter adaptation. Such adaptation is done passively, by augmenting physics-based models with learned models, with our key contribution being that the interpretability of physical parameters is retained, allowing us to monitor adaptation. Second, I will talk about a framework for active adaptation where the model monitors its own performance and curates a diverse subset of uncertain inputs to be used for periodic fine-tuning of the model, improving performance over the full data lifecycle.
About about the speaker: https://stanfordasl.github.io//people/somrita-banerjee/
More about the course can be found here: https://stanfordasl.github.io/robotics_seminar/
View the entire AA289 Stanford Robotics and Autonomous Systems Seminar playlist: https://www.youtube.com/playlist?list=PLoROMvodv4rMeercb-kvGLUrOq4HR6BZD
► Check out the entire catalog of courses and programs available through Stanford Online: https://online.stanford.edu/explore
Видео Stanford Seminar - Learning-enabled Adaptation to Evolving Conditions for Robotics канала Stanford Online
Somrita Banerjee, Stanford University
With advancements in machine learning and artificial intelligence, a new generation of “learning-enabled” robots is emerging, which are better suited to operating autonomously in unstructured, uncertain, and unforgiving environments. To achieve these goals, robots must be able to adapt to evolving conditions that are different from those seen during training or expected during deployment. In this talk I will first talk about adapting to novel instantiations, i.e., different task instances with shared structure, through parameter adaptation. Such adaptation is done passively, by augmenting physics-based models with learned models, with our key contribution being that the interpretability of physical parameters is retained, allowing us to monitor adaptation. Second, I will talk about a framework for active adaptation where the model monitors its own performance and curates a diverse subset of uncertain inputs to be used for periodic fine-tuning of the model, improving performance over the full data lifecycle.
About about the speaker: https://stanfordasl.github.io//people/somrita-banerjee/
More about the course can be found here: https://stanfordasl.github.io/robotics_seminar/
View the entire AA289 Stanford Robotics and Autonomous Systems Seminar playlist: https://www.youtube.com/playlist?list=PLoROMvodv4rMeercb-kvGLUrOq4HR6BZD
► Check out the entire catalog of courses and programs available through Stanford Online: https://online.stanford.edu/explore
Видео Stanford Seminar - Learning-enabled Adaptation to Evolving Conditions for Robotics канала Stanford Online
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
![Stanford Seminar - Fight over the Law of Software APIs & stories from Electronic Frontier Foundation](https://i.ytimg.com/vi/j4m2rVTQ098/default.jpg)
![Stanford CS224W: Machine Learning with Graphs | 2021 | Lecture 17.3 - Cluster GCN: Scaling up GNNs](https://i.ytimg.com/vi/RJkR8Ig6dXI/default.jpg)
![Stanford Seminar - Intelligence Augmentation through the Lens of Interactive Data Visualization](https://i.ytimg.com/vi/wZc1Rq5FtXc/default.jpg)
![Stanford Seminar - Get in touch: Tactile perception for human-robot systems](https://i.ytimg.com/vi/sHedq5u4gF8/default.jpg)
![Learner Spotlight: Andrew Pelosi](https://i.ytimg.com/vi/fUTtXiEuQX0/default.jpg)
![Antonio Del Santo talks about his experience in Stanford's Digital Health Product Development Course](https://i.ytimg.com/vi/KTuhZAMKCWY/default.jpg)
![Stanford Seminar - Human-AI Interaction Under Societal Disagreement](https://i.ytimg.com/vi/Y3Kj_2ZNRfw/default.jpg)
![Student Spotlight: Daphne Wallbridge talks about the Creativity and Design Thinking Program](https://i.ytimg.com/vi/oDhlQ_gbODg/default.jpg)
![Webinar - Big Breaches: What We Learned From the World’s Most Disruptive Cybersecurity Attacks](https://i.ytimg.com/vi/lK7XV13wE30/default.jpg)
![Stanford Seminar - Interview with Amy Chang of Accompany](https://i.ytimg.com/vi/uXc7_TohfYY/default.jpg)
![Stanford Webinar - Natural Language Understanding Student Project Showcase - Plus AI Program Q&A](https://i.ytimg.com/vi/asKwoOG6nHM/default.jpg)
![Building Energy Efficiency: Technology, Policy & Policy (bee.stanford.edu)](https://i.ytimg.com/vi/hLJgz5OYs48/default.jpg)
![Stanford CS224W: Machine Learning with Graphs | 2021 | Lecture 10.2 - Knowledge Graph Completion](https://i.ytimg.com/vi/xop5tC9T5xM/default.jpg)
![Stanford Seminar - Towards Shape Changing Displays and Shape Changing Robots](https://i.ytimg.com/vi/3eCkGvvMGA4/default.jpg)
![Stanford Workshop - Innovation at Work Prototyping Preview](https://i.ytimg.com/vi/n_-54F8K9wI/default.jpg)
![Stanford Seminar - Improving Computational Efficiency for Powered Descent Guidance](https://i.ytimg.com/vi/U-c2XG_gG9E/default.jpg)
![Stanford Seminar - Incorporating Sample Efficient Monitoring into Learned Autonomy](https://i.ytimg.com/vi/Xig_Kn31R2M/default.jpg)
![Stanford CS224W: ML with Graphs | 2021 | Lecture 16.4 - Robustness of Graph Neural Networks](https://i.ytimg.com/vi/uBmtTL3EitI/default.jpg)
![Network and Information Security: Reflections on the Past and the Look Ahead](https://i.ytimg.com/vi/Tg0JjeVvQmw/default.jpg)
![Decarbonizing the Grid: Technology and Policy (webcast)](https://i.ytimg.com/vi/8uSaLB6UDiQ/default.jpg)
![Stanford Seminar - ML Explainability Part 3 I Post hoc Explanation Methods](https://i.ytimg.com/vi/_6n8r523QP8/default.jpg)