@sacha2.bsky.social
Pinned
sacha2.bsky.social
@sacha2.bsky.social
· Nov 13
I might look smart, however, I am absolutely not.
Reposted
Thrilled to present HyperMARL at #NeurIPS2025 in San Diego next week! 🚀 (Amos will present at
@euripsconf.bsky.social too.)
TL;DR: Coupling obs and agent IDs can hurt performance in MARL. Agent-conditioned hypernets cleanly decouple grads and enable specialisation.
📜: arxiv.org/abs/2412.04233
@euripsconf.bsky.social too.)
TL;DR: Coupling obs and agent IDs can hurt performance in MARL. Agent-conditioned hypernets cleanly decouple grads and enable specialisation.
📜: arxiv.org/abs/2412.04233
November 26, 2025 at 4:07 PM
Thrilled to present HyperMARL at #NeurIPS2025 in San Diego next week! 🚀 (Amos will present at
@euripsconf.bsky.social too.)
TL;DR: Coupling obs and agent IDs can hurt performance in MARL. Agent-conditioned hypernets cleanly decouple grads and enable specialisation.
📜: arxiv.org/abs/2412.04233
@euripsconf.bsky.social too.)
TL;DR: Coupling obs and agent IDs can hurt performance in MARL. Agent-conditioned hypernets cleanly decouple grads and enable specialisation.
📜: arxiv.org/abs/2412.04233
Hi! In some parts of the world, people celebrate Thanksgiving this week! Wish you everyone to have nice time.
November 26, 2025 at 4:01 PM
Hi! In some parts of the world, people celebrate Thanksgiving this week! Wish you everyone to have nice time.
Reposted
#VisionLanguage models are increasingly used for a wide range of problems, but seem complex to build. I wrote some code and recorded a tutorial in my lab yesterday to help others demystify how to create these models. #keepbuilding
November 25, 2025 at 5:40 PM
#VisionLanguage models are increasingly used for a wide range of problems, but seem complex to build. I wrote some code and recorded a tutorial in my lab yesterday to help others demystify how to create these models. #keepbuilding
this is sweet direct.mit.edu/isal/isal202...
Volumes | ALIFE 2022: The 2022 Conference on Artificial Life | MIT Press
ALIFE 2025: Ciphers of Life: Proceedings of the Artificial Life Conference 2025 | October 2025 | Kyoto, Japan
direct.mit.edu
November 24, 2025 at 9:27 PM
this is sweet direct.mit.edu/isal/isal202...
Reposted
How do we close the gap between specialist RL and generalist LLM agents?
We're benchmarking it in Pokémon. Join us at the PokeAgent Challenge competition workshop @ NeurIPS 2025.
📍 Dec 7, 8AM
🎮 Track 1: Competitive Pokémon (game-theoretic reasoning)
🗺️ Track 2: Speedrunning (long-horizon planning)
We're benchmarking it in Pokémon. Join us at the PokeAgent Challenge competition workshop @ NeurIPS 2025.
📍 Dec 7, 8AM
🎮 Track 1: Competitive Pokémon (game-theoretic reasoning)
🗺️ Track 2: Speedrunning (long-horizon planning)
November 24, 2025 at 5:50 PM
How do we close the gap between specialist RL and generalist LLM agents?
We're benchmarking it in Pokémon. Join us at the PokeAgent Challenge competition workshop @ NeurIPS 2025.
📍 Dec 7, 8AM
🎮 Track 1: Competitive Pokémon (game-theoretic reasoning)
🗺️ Track 2: Speedrunning (long-horizon planning)
We're benchmarking it in Pokémon. Join us at the PokeAgent Challenge competition workshop @ NeurIPS 2025.
📍 Dec 7, 8AM
🎮 Track 1: Competitive Pokémon (game-theoretic reasoning)
🗺️ Track 2: Speedrunning (long-horizon planning)
I have a suggestion for computer science papers: if a paper contains experiments, the submission (not camera version) should contain code and minimal reproduction instructions for the experiments. Camera version should contain minimal reproduction data samples, code, and the instructions.
November 23, 2025 at 5:30 PM
I have a suggestion for computer science papers: if a paper contains experiments, the submission (not camera version) should contain code and minimal reproduction instructions for the experiments. Camera version should contain minimal reproduction data samples, code, and the instructions.
Reposted
Olmo 3 is notable as a "fully open" LLM - all of the training data is published, plus complete details on how the training process was run. I tried out the 32B thinking model and the 7B instruct models, + thoughts on why transparent training data is so important simonwillison.net/2025/Nov/22/...
Olmo 3 is a fully open LLM
Olmo is the LLM series from Ai2—the Allen institute for AI. Unlike most open weight models these are notable for including the full training data, training process and checkpoints along …
simonwillison.net
November 23, 2025 at 12:17 AM
Olmo 3 is notable as a "fully open" LLM - all of the training data is published, plus complete details on how the training process was run. I tried out the 32B thinking model and the 7B instruct models, + thoughts on why transparent training data is so important simonwillison.net/2025/Nov/22/...
Reposted
Introducing 🥚EGGROLL 🥚(Evolution Guided General Optimization via Low-rank Learning)! 🚀 Scaling backprop-free Evolution Strategies (ES) for billion-parameter models at large population sizes
⚡100x Training Throughput
🎯Fast Convergence
🔢Pure Int8 Pretraining of RNN LLMs
⚡100x Training Throughput
🎯Fast Convergence
🔢Pure Int8 Pretraining of RNN LLMs
November 21, 2025 at 5:56 PM
Introducing 🥚EGGROLL 🥚(Evolution Guided General Optimization via Low-rank Learning)! 🚀 Scaling backprop-free Evolution Strategies (ES) for billion-parameter models at large population sizes
⚡100x Training Throughput
🎯Fast Convergence
🔢Pure Int8 Pretraining of RNN LLMs
⚡100x Training Throughput
🎯Fast Convergence
🔢Pure Int8 Pretraining of RNN LLMs
danish is a cool language
Kan naturens egen designproces videreudvikle AI?
En ny lærebog skrevet af ITU-professor @risi.bsky.social m.fl. udforsker feltet 𝙉𝙚𝙪𝙧𝙤𝙚𝙫𝙤𝙡𝙪𝙩𝙞𝙤𝙣, der kan ændre fremtidens kunstige intelligens.
"Grundideen er at efterligne, hvordan intelligens opstod i naturen."
Læs mere 👉 itu.dk/Om-ITU/Press...
En ny lærebog skrevet af ITU-professor @risi.bsky.social m.fl. udforsker feltet 𝙉𝙚𝙪𝙧𝙤𝙚𝙫𝙤𝙡𝙪𝙩𝙞𝙤𝙣, der kan ændre fremtidens kunstige intelligens.
"Grundideen er at efterligne, hvordan intelligens opstod i naturen."
Læs mere 👉 itu.dk/Om-ITU/Press...
November 21, 2025 at 3:15 PM
danish is a cool language
have you used Google Antigravity? can email some questions
November 20, 2025 at 7:40 PM
have you used Google Antigravity? can email some questions
good to know that I am still better at looking for articles than AI
but it's catching up
but it's catching up
Check this out! "Google Scholar Labs helps you answer research questions with AI" share.google/aKBsEebfos7c...
We’re introducing Google Scholar Labs to answer your research questions.
Today, we are introducing Google Scholar Labs, a new feature that explores how generative AI can transform the process of answering detailed scholarly research questions…
share.google
November 20, 2025 at 1:09 PM
good to know that I am still better at looking for articles than AI
but it's catching up
but it's catching up
Reposted
Lovely to see our episode on play and games in utopia/dystopia featured here. We had a great time chatting with Stefano!
🏹The Hunger Games builds its narrative around a fictional game. What can the (fictional) games we play reveal about the values held by society?
🎙️ @gua-le-ni.com chats with @matthewleggatt.bsky.social & @dystopiajunkie.bsky.social on @utopiaanddystopia.bsky.social
🔗 www.um.edu.mt/newspoint/li...
🎙️ @gua-le-ni.com chats with @matthewleggatt.bsky.social & @dystopiajunkie.bsky.social on @utopiaanddystopia.bsky.social
🔗 www.um.edu.mt/newspoint/li...
Utopia, Play, and Fictional Games: A conversation with Prof. Stefano Gualeni
How do fictional games work as world-building tools and what can the games we play reveal?
www.um.edu.mt
November 20, 2025 at 10:58 AM
Lovely to see our episode on play and games in utopia/dystopia featured here. We had a great time chatting with Stefano!
ICLR 2026 Response to LLM-Generated Papers and Reviews – ICLR Blog
blog.iclr.cc
November 20, 2025 at 9:10 AM
UwU
Excited to announce our book “Neuroevolution: Harnessing Creativity in AI Agent Design” by Sebastian Risi, Yujin Tang, Risto Miikkulainen, and myself. We explore decades of work on evolving intelligent agents and shows how neuroevolution can drive creativity in deep learning, RL, LLMs and AI Agents!
November 20, 2025 at 8:56 AM
UwU
November 19, 2025 at 9:22 PM
Reposted
I'm looking for two PhD student for Fall 2026. Both on multi-agent reinforcement learning (MARL).
- Theory of MARL: experience with theory and/or MARL
-Formal methods for MARL: experience with formal methods or MARL (interest in learning the other)
www.khoury.northeastern.edu/programs/com...
- Theory of MARL: experience with theory and/or MARL
-Formal methods for MARL: experience with formal methods or MARL (interest in learning the other)
www.khoury.northeastern.edu/programs/com...
PhD in Computer Science - Khoury College of Computer Sciences
The PhD in Computer Science program will prepare you with advanced knowledge, industry opportunities, and research experience to be a leader in the field.
www.khoury.northeastern.edu
November 19, 2025 at 2:30 PM
I'm looking for two PhD student for Fall 2026. Both on multi-agent reinforcement learning (MARL).
- Theory of MARL: experience with theory and/or MARL
-Formal methods for MARL: experience with formal methods or MARL (interest in learning the other)
www.khoury.northeastern.edu/programs/com...
- Theory of MARL: experience with theory and/or MARL
-Formal methods for MARL: experience with formal methods or MARL (interest in learning the other)
www.khoury.northeastern.edu/programs/com...
I would love Gemini 3, however, the sheer amount of ads (or affiliated posts) that has been spammed for like a week, leaves a bitter taste.
Don't get me wrong.
Don't get me wrong.
November 19, 2025 at 12:14 PM
I would love Gemini 3, however, the sheer amount of ads (or affiliated posts) that has been spammed for like a week, leaves a bitter taste.
Don't get me wrong.
Don't get me wrong.
A lot of words about games evaluation for different LLMs, maybe some strategic insight.
game-arena.ai/search
cc @sharky6000.bsky.social
game-arena.ai/search
cc @sharky6000.bsky.social
Game Arena: LLM Board Game Tournaments
A new benchmark for evaluating LLM reasoning and instruction following in board game environments.
game-arena.ai
November 18, 2025 at 3:16 PM
A lot of words about games evaluation for different LLMs, maybe some strategic insight.
game-arena.ai/search
cc @sharky6000.bsky.social
game-arena.ai/search
cc @sharky6000.bsky.social
can anyone buy me this book? it's just 40 euros, i will give you back double in some time hehe
I share more on how cooperation leads to larger, collective intelligences in my new book “What Is Intelligence?” (out via @mitpress.bsky.social & Antikythera): bit.ly/3H1p8F6
What Is Intelligence?: Lessons from AI About Evolution, Computing, and Minds (Antikythera): Aguera y Arcas, Blaise: 9780262049955: Amazon.com: Books
What Is Intelligence?: Lessons from AI About Evolution, Computing, and Minds (Antikythera) [Aguera y Arcas, Blaise] on Amazon.com. *FREE* shipping on qualifying offers. What Is Intelligence?: Lessons ...
bit.ly
November 17, 2025 at 7:26 PM
can anyone buy me this book? it's just 40 euros, i will give you back double in some time hehe
interesting what to write in the research proposal to get an interview
November 17, 2025 at 2:59 PM
interesting what to write in the research proposal to get an interview
November 17, 2025 at 8:53 AM
Reposted
Argyrios Deligkas, Gregory Gutin, Mark Jones, Philip R. Neary, Anders Yeo
Public Goods Games in Directed Networks with Constraints on Sharing
https://arxiv.org/abs/2511.11475
Public Goods Games in Directed Networks with Constraints on Sharing
https://arxiv.org/abs/2511.11475
November 17, 2025 at 5:35 AM
Argyrios Deligkas, Gregory Gutin, Mark Jones, Philip R. Neary, Anders Yeo
Public Goods Games in Directed Networks with Constraints on Sharing
https://arxiv.org/abs/2511.11475
Public Goods Games in Directed Networks with Constraints on Sharing
https://arxiv.org/abs/2511.11475
Reposted
Piotr Faliszewski, Stanislaw Kazmierowski, Grzegorz Lisowski, Ildiko Schlotter, Paolo Turrini
Computing Equilibrium Nominations in Presidential Elections
https://arxiv.org/abs/2511.11365
Computing Equilibrium Nominations in Presidential Elections
https://arxiv.org/abs/2511.11365
November 17, 2025 at 5:36 AM
Piotr Faliszewski, Stanislaw Kazmierowski, Grzegorz Lisowski, Ildiko Schlotter, Paolo Turrini
Computing Equilibrium Nominations in Presidential Elections
https://arxiv.org/abs/2511.11365
Computing Equilibrium Nominations in Presidential Elections
https://arxiv.org/abs/2511.11365
Reposted
Erwan Christian Escudie, Matthia Sabatelli, Olivier Buffet, Jilles Steeve Dibangoye
{\epsilon}-Optimally Solving Two-Player Zero-Sum POSGs
https://arxiv.org/abs/2511.11282
{\epsilon}-Optimally Solving Two-Player Zero-Sum POSGs
https://arxiv.org/abs/2511.11282
November 17, 2025 at 5:37 AM
Erwan Christian Escudie, Matthia Sabatelli, Olivier Buffet, Jilles Steeve Dibangoye
{\epsilon}-Optimally Solving Two-Player Zero-Sum POSGs
https://arxiv.org/abs/2511.11282
{\epsilon}-Optimally Solving Two-Player Zero-Sum POSGs
https://arxiv.org/abs/2511.11282