remosasso

PhD @QMUL, AI Developer @ xDNA

[remosasso @remosasso]

Very excited to announce that our paper Posterior Sampling for Deep Reinforcement Learning (PSDRL) has been accepted at #ICML2023!

⌨️Code: https://t.co/BYEwwpOa3m
📄 Paper: https://t.co/YxuNCtsNyh

🧵Thread 👇

[European Digital Media Observatory @EDMO_EUI]

🚀 Enhancing the efficiency of #factchecking through #AI: @remosasso explains how, funded by @EMIF_Fund, the #Aletheia project develops several AI systems that find relevant sources that can extract different claims and cross-reference relevant statements. #EDMOeu
#xDNA

[remosasso @remosasso]

@GoogleDeepMind @LFC @SuleyOzturk

[Ayla @ayla_pollmann]

🧠 Exploring the Link Between Adversity and Brain Development 🌱

Our latest #preprint investigates the impact of adverse adolescent experiences on brain connectivity using the Adolescent Brain Cognitive Development study (ABCD). 🧩

🧵🪡

https://t.co/C4urdiEGr7

[remosasso @remosasso]

Numer 835!

[remosasso @remosasso]

Come join us today from 2PM - 3.30PM at ICML 2023, Exhibit Hall 1 to chat about our latest work "Posterior Sampling for Deep Reinforcement Learning"!

We are the first to scale PSRL in its original model-based formulation, and it works really well!

👇👇👇

[remosasso @remosasso]

Very excited to announce that our paper Posterior Sampling for Deep Reinforcement Learning (PSDRL) has been accepted at #ICML2023!

⌨️Code: https://t.co/BYEwwpOa3m
📄 Paper: https://t.co/YxuNCtsNyh

🧵Thread 👇

[remosasso @remosasso]

I will be in Honolulu to present our PSDRL paper at ICML together with @Michelangelo755 from 22nd July - 30th July!

DM me if you would like to meet up and have a chat!

#ICML2023

[remosasso @remosasso]

Very excited to announce that our paper Posterior Sampling for Deep Reinforcement Learning (PSDRL) has been accepted at #ICML2023!

⌨️Code: https://t.co/BYEwwpOa3m
📄 Paper: https://t.co/YxuNCtsNyh

🧵Thread 👇

[remosasso @remosasso]

Huge thanks to my collaborator @Michelangelo755 and supervisor Paulo Rauber (https://t.co/7m7XnTd6sa) for their invaluable support in this project!

[remosasso @remosasso]

Very excited to announce that our paper Posterior Sampling for Deep Reinforcement Learning (PSDRL) has been accepted at #ICML2023!

⌨️Code: https://t.co/BYEwwpOa3m
📄 Paper: https://t.co/YxuNCtsNyh

🧵Thread 👇

[remosasso @remosasso]

🔎By acting greedily w.r.t. sampled models, the exploration of the agent is naturally driven through uncertainty over models of the environments.

The diversity among decoded latent trajectories of different sampled models starting at the same initial state can be seen below: