Technion - Reinforcement Learning Research Labs

Official account covering research performed in the various #ReinforcementLearning labs at the @TechnionLive

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Deep networks combined with finite memory poses a problem for bandit-based algorithms.

In this blog post, Ofir explains their method for overcoming the drift in the learned features and avoiding catastrophic forgetting!

https://t.co/63bLKQ3Hqh

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Interested in learning how to detect reward deterioration?
Check out the blog post by Ido Greenberg!

https://t.co/0bYH6MB91T

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Check out @carmelrabinovi1 's new work! Constrastive domain randomization for sim-to-real transfer!

[Carmel Rabinovitz @CarmelRabinovi1]

Excited to share our new #ICRA2021 paper on “Contrastive Domain Randomization” (CDR) with my amazing collaborators @nikogrupen, and @AvivTamar1:

Website: https://t.co/VM329oos1Y
Paper: https://t.co/z18zNg1VPg
Code: https://t.co/9AOaRiFXhI

#robotics #ContrastiveLearning #AI

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Action Redundancy in Reinforcement Learning - UAI
Nir Baram, @guytenn and @MannorShie

[Technion - Reinforcement Learning Research Labs @Technion_RL]

It's been a while without updates. Let's cover some of the recent publications from the various RL groups at the @TechnionLive !

Thread.

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Bandits with Partially Observable Confounded Data - UAI
@guytenn, @ShalitUri, @MannorShie and Yonathan Efroni

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Awesome blog post by Orr, Omer and Raveh!
Interested in RL and robotics? Certainly worth checking this out!

[Orr Krupnik @orrkrup]

What happens when you set out to reproduce a "reproducible robotics benchmark"?
Omer and Raveh tried it out with #REPLAB! Their findings are detailed in the new blog post at
@Technion_RL.

TL;DR: it takes some work, but it's definitely possible!

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Super excited for the first Israeli RL day! 🙌

[Orr Krupnik @orrkrup]

The first Israeli RL Day will take place on Apr. 7th! Program is packed with great speakers, plus keynotes by @AvivTamar1, @MannorShie and Yishay Mansour.

Schedule details at https://t.co/92m5iNK0xJ.

#IRLD2021

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Work by @BeloStav, https://t.co/vTgjA2iHHy, @TZahavy, @tesslerc, @MannorShie
To be featured in a special issue on RL for Real Life @Springer @MLJ_Social, by (@yuxili99, @AlborzGr, @LihongLi20, @CsabaSzepesvari, Tao Wang). https://t.co/8LwDXEwg2a

[Technion - Reinforcement Learning Research Labs @Technion_RL]

Interested in #RealWorld applications of #ReinforcementLearning?
Check out our recent work, “Inverse Reinforcement Learning in Contextual MDPs” where we learn from clinicians how to treat patients with Sepsis!

Paper: https://t.co/sPhpF8wxYw
Code: https://t.co/MIAxLQMdEz

[Technion - Reinforcement Learning Research Labs @Technion_RL]

To evaluate our algorithm we use the MIMIC-III data set https://t.co/J5blyIYSVu @MIT_IMES. We focus on Sepsis treatment, which recently emerged as a benchmark for RL algorithms, and achieve new SOTA results.