"Augmenting Model-Based Reinforcement Learning by Encouraging Joint Self-Consistency", Zita Marinho (DeepMind)

3 views
Skip to first unread message

Diogo Pernes

unread,
Mar 22, 2023, 7:06:52 PM3/22/23
to priberam_...@googlegroups.com, isr-...@isr.tecnico.ulisboa.pt, si...@omni.isr.ist.utl.pt
Dear all,

We are pleased to announce that we will be holding the third session of this year's Priberam Machine Learning Seminars on Tuesday, April 4. Our guest speaker will be Zita Marinho, a Research Scientist at DeepMind. Her talk, entitled "Augmenting Model-Based Reinforcement Learning by Encouraging Joint Self-Consistency", will explore ways of augmenting model-based RL by encouraging a learned model and value function to be jointly self-consistent.

The event will occur at 1 PM in Instituto Superior Técnico (room PA2), and we will provide lunch bags for attendees. To learn more about the event and register (which is mandatory if you plan to attend), please follow the link below:


We look forward to seeing you all there!

Kind regards,
Diogo Pernes

 
Priberam Labs
http://labs.priberam.com/

Priberam is hiring!
If you are interested in working with us please consult the available positions at priberam.com/careers.

Image result for priberam logoPRIBERAM SEMINARS

__________________________________________________

Priberam Machine Learning Lunch Seminar
Speaker: Zita Marinho (DeepMind)
Venue: Instituto Superior Técnico (room PA2)
Date: Tuesday, April 4, 2023
Time: 1 PM 
Title:
Model-Value Self-Consistent Updates and Applications
Abstract:
Learned models of the environment provide reinforcement learning agents with flexible ways of making predictions about the environment. Models enable planning, i.e. using more computation to improve value functions or policies, without requiring additional environment interactions. In this talk, we investigate a way of augmenting model-based RL, by additionally encouraging a learned model and value function to be jointly self-consistent. This work covers possible ways to use self-consistency updates both for policy evaluation and control (Farquhar et al 20), as well as a proxy for epistemic uncertainty in exploration (Filos et al. 22).
Short Bio:
Zita Marinho is a Research Scientist at Deepmind, where she is currently working on reinforcement learning. She holds a dual PhD/MSc in Robotics from the Robotics Institute, and from IST University of Lisbon as part of the CMU/Portugal program. She graduated from her MSc. degree in Physics Engineering from Instituto Superior Técnico, Universidade de Lisboa in 2010. Her research interests lie in the intersection of machine learning algorithms and Natural Language Processing. She is particularly interested in studying how agents can interact and learn more effectively from those interactions. She studied during her PhD spectral algorithms for sequence prediction and planning. She was jointly advised by Prof. André Martins at Unbabel/IST, Prof. Geoffrey Gordon at the Machine Learning Department/CMU and Prof. Siddhartha Srinivasa from University of Washington.

Gonçalo Correia

unread,
Apr 3, 2023, 7:13:44 AM4/3/23
to priberam_...@googlegroups.com

Dear all,

Friendly reminder that we are having a seminar tomorrow at 1 PM in room PA2 of Instituto Superior Técnico. Our guest will be Zita Marinho, who is a Research Scientist at DeepMind.

To register,
Gonçalo Correia

 

 
Priberam Labs
http://labs.priberam.com/

Priberam is hiring!
If you are interested in working with us please consult the available positions at priberam.com/careers.

Image result for priberam logoPRIBERAM SEMINARS

__________________________________________________

Priberam Machine Learning Lunch Seminar
Speaker: Zita Marinho (DeepMind)
Venue: Instituto Superior Técnico (room PA2)
Date: Tuesday, April 4, 2023
Time: 1 PM 
Title:
Model-Value Self-Consistent Updates and Applications
Abstract:
Learned models of the environment provide reinforcement learning agents with flexible ways of making predictions about the environment. Models enable planning, i.e. using more computation to improve value functions or policies, without requiring additional environment interactions. In this talk, we investigate a way of augmenting model-based RL, by additionally encouraging a learned model and value function to be jointly self-consistent. This work covers possible ways to use self-consistency updates both for policy evaluation and control (Farquhar et al 20), as well as a proxy for epistemic uncertainty in exploration (Filos et al. 22).
Short Bio:
Zita Marinho is a Research Scientist at Deepmind, where she is currently working on reinforcement learning. She holds a dual PhD/MSc in Robotics from the Robotics Institute, and from IST University of Lisbon as part of the CMU/Portugal program. She graduated from her MSc. degree in Physics Engineering from Instituto Superior Técnico, Universidade de Lisboa in 2010. Her research interests lie in the intersection of machine learning algorithms and Natural Language Processing. She is particularly interested in studying how agents can interact and learn more effectively from those interactions. She studied during her PhD spectral algorithms for sequence prediction and planning. She was jointly advised by Prof. André Martins at Unbabel/IST, Prof. Geoffrey Gordon at the Machine Learning Department/CMU and Prof. Siddhartha Srinivasa from University of Washington.


From: Gonçalo Correia
Sent: Monday, April 3, 2023 10:28:08 AM
To: priberam_...@googlegroups.com; isr-...@isr.tecnico.ulisboa.pt; si...@omni.isr.ist.utl.pt
Subject: Re: "Augmenting Model-Based Reinforcement Learning by Encouraging Joint Self-Consistency", Zita Marinho (DeepMind)
 

Dear all,


Friendly reminder that we are having a seminar tomorrow at 1 PM in room PA2 of Instituto Superior Técnico. Our guest will be Zita Marinho, who is a Research Scientist at DeepMind.


To register,

Gonçalo Correia


Priberam Labs
http://labs.priberam.com/

Priberam is hiring!
If you are interested in working with us please consult the available positions at priberam.com/careers.

Image result for priberam logoPRIBERAM SEMINARS

__________________________________________________

Priberam Machine Learning Lunch Seminar
Speaker: Zita Marinho (DeepMind)
Venue: Instituto Superior Técnico (room PA2)
Date: Tuesday, April 4, 2023
Time: 1 PM 
Title:
Model-Value Self-Consistent Updates and Applications
Abstract:
Learned models of the environment provide reinforcement learning agents with flexible ways of making predictions about the environment. Models enable planning, i.e. using more computation to improve value functions or policies, without requiring additional environment interactions. In this talk, we investigate a way of augmenting model-based RL, by additionally encouraging a learned model and value function to be jointly self-consistent. This work covers possible ways to use self-consistency updates both for policy evaluation and control (Farquhar et al 20), as well as a proxy for epistemic uncertainty in exploration (Filos et al. 22).
Short Bio:
Zita Marinho is a Research Scientist at Deepmind, where she is currently working on reinforcement learning. She holds a dual PhD/MSc in Robotics from the Robotics Institute, and from IST University of Lisbon as part of the CMU/Portugal program. She graduated from her MSc. degree in Physics Engineering from Instituto Superior Técnico, Universidade de Lisboa in 2010. Her research interests lie in the intersection of machine learning algorithms and Natural Language Processing. She is particularly interested in studying how agents can interact and learn more effectively from those interactions. She studied during her PhD spectral algorithms for sequence prediction and planning. She was jointly advised by Prof. André Martins at Unbabel/IST, Prof. Geoffrey Gordon at the Machine Learning Department/CMU and Prof. Siddhartha Srinivasa from University of Washington.



From: Diogo Pernes
Sent: Wednesday, March 22, 2023 11:06:48 PM
To: priberam_...@googlegroups.com; isr-...@isr.tecnico.ulisboa.pt; si...@omni.isr.ist.utl.pt
Subject: "Augmenting Model-Based Reinforcement Learning by Encouraging Joint Self-Consistency", Zita Marinho (DeepMind)
 
Reply all
Reply to author
Forward
0 new messages