WhiRL

595 posts

WhiRL banner
WhiRL

WhiRL

@whi_rl

Whiteson Research Lab @CompSciOxford. Reinforcement learning and deep learning with a focus on multi-agent learning and meta-learning

Oxford, England Katılım Mayıs 2017
205 Takip Edilen4.7K Takipçiler
WhiRL retweetledi
Roberta Raileanu
Roberta Raileanu@robertarail·
How can agents get better at algorithm discovery? Meta-meta-learning is one answer, aka improving the agents themselves at inventing generalizable algorithms. DiscoBench provides a way to procedurally generate algorithm discovery tasks at scale, which can be used for meta-meta-learning. Kudos to @AlexDGoldie and team for the release!
Alex Goldie@AlexDGoldie

1/ 🪩 Automating the discovery of new algorithms could unlock significant breakthroughs in ML research. But optimising agents for this research has been limited by too few tasks to learn from! Introducing DiscoGen, a procedural generator of algorithm discovery tasks 🧵

English
1
15
88
12.2K
WhiRL retweetledi
Alex Goldie
Alex Goldie@AlexDGoldie·
1/ 🪩 Automating the discovery of new algorithms could unlock significant breakthroughs in ML research. But optimising agents for this research has been limited by too few tasks to learn from! Introducing DiscoGen, a procedural generator of algorithm discovery tasks 🧵
Alex Goldie tweet media
English
3
41
144
34.7K
WhiRL
WhiRL@whi_rl·
👀 MARL can unlock self-driving in unseen cities without needing any prior human demonstrations! 🤯 Performance improvements in terms of driving success rate *and* human likeness! Check out our new work, led by @zilinwang4ai and @saeedrmd 👇
Zilin Wang@zilinwang4ai

1/ 🚗 🌏 What if an autonomous vehicle could move to a new city without collecting a single human demonstration in that city? I am so excited to introduce our new work: Learning to Drive in New Cities Without Human Demonstrations.

English
0
0
8
557
WhiRL retweetledi
Matthew Jackson
Matthew Jackson@JacksonMattT·
Unifloral has been accepted as an Oral at NeurIPS 2025! Immensely grateful to my @FLAIR_Ox co-authors @uljadb99 and @JarekLiesen for pouring months of effort into this project. There’s a ton of low-hanging fruit in offline RL… If you’re looking for a project, check it out!
Matthew Jackson tweet media
Matthew Jackson@JacksonMattT

🌹 Today we're releasing Unifloral, our new library for Offline Reinforcement Learning! We make research easy: ⚛️ Single-file 🤏 Minimal ⚡️ End-to-end Jax Best of all, we unify prior methods into one algorithm - a single hyperparameter space for research! ⤵️

English
3
20
179
33.2K
WhiRL retweetledi
Alex Goldie
Alex Goldie@AlexDGoldie·
🥳 It’s an honour to have been awarded the Outstanding Paper for Scientific Understanding in RL at RLC for our work, ‘How Should We Meta-Learn RL Algorithms?’ Thank you to the organisers @RL_Conference for putting on a great conference, and congratulations to the other winners!
Alex Goldie tweet mediaAlex Goldie tweet media
English
3
21
224
22.1K
WhiRL
WhiRL@whi_rl·
The best of RL research, brought to Offline RL! 🚀 TL;DR 1. CleanRL-style implementations ⚡️ 2. Rainbow-style algorithm unification 🦾 3. Rliable-style evaluation protocol 🔬 Check out our paper + library!
Matthew Jackson@JacksonMattT

🌹 Today we're releasing Unifloral, our new library for Offline Reinforcement Learning! We make research easy: ⚛️ Single-file 🤏 Minimal ⚡️ End-to-end Jax Best of all, we unify prior methods into one algorithm - a single hyperparameter space for research! ⤵️

English
0
1
16
1.4K
WhiRL retweetledi
Matthew Jackson
Matthew Jackson@JacksonMattT·
🌹 Today we're releasing Unifloral, our new library for Offline Reinforcement Learning! We make research easy: ⚛️ Single-file 🤏 Minimal ⚡️ End-to-end Jax Best of all, we unify prior methods into one algorithm - a single hyperparameter space for research! ⤵️
English
4
37
172
94.3K
WhiRL retweetledi
Luisa Zintgraf
Luisa Zintgraf@luisa_zintgraf·
🎉 Our Meta-RL survey is now published in Foundations and Trends in Machine Learning! A deep dive into how agents can learn to learn 🤖🧠 Huge kudos to @jakeABeck & @ristovuorio for leading the charge, and to co-authors Evan Liu, Zheng Xiong, @chelseabfinn & @shimon8282!
English
2
10
57
5.2K
WhiRL retweetledi
Jacob Beck
Jacob Beck@jakeABeck·
Big news—our survey paper “A Tutorial on Meta-Reinforcement Learning” is officially published! Meta-RL = learning how to adapt through interaction. It embraces The Bitter Lesson: don’t hardcode agents—train them to adapt on their own arxiv.org/abs/2301.08028 🧵⬇️
English
2
77
337
21.1K
WhiRL retweetledi
Jacob Beck
Jacob Beck@jakeABeck·
🎉🚨 Big news! Our research, Metalic: Meta-Learning In-Context with Protein Language Models, 🧬 won a competition! #NeurIPS2024🤖📚 We advance in-context learning and protein fitness prediction with this paradigm: ✨ Pre-training 🔥 Learning to in-context learn🔥 ✨ Fine-tuning
Jacob Beck tweet media
English
2
4
9
1K
WhiRL
WhiRL@whi_rl·
Check out this fantastic showing by @jakeABeck and @AlexDGoldie at #ICML2024! 🎙️🔥 They dove deep into the future of automated RL, meta-learning, and LLMs 🤖🔮
Jacob Beck@jakeABeck

Missed this provocative panel? I was honored to share the stage at #ICML2024 with @pcastr, @XingyouSong, and my colleague @AlexDGoldie! We discussed future perspectives on automated RL, meta-learning, and LLMs 🤖 Catch the discussion here: icml.cc/virtual/2024/w… at 7:16:00 🎙️

English
0
1
11
1K
WhiRL retweetledi
Alex Goldie
Alex Goldie@AlexDGoldie·
Attending #ICML2024 was amazing and full of firsts: My first time presenting a poster, first time giving a talk at a conference and first time sitting on a panel! Many thanks to the @AutoRL_Workshop organisers for preparing a great workshop about AutoRL!
Alex Goldie tweet mediaAlex Goldie tweet mediaAlex Goldie tweet media
English
4
6
55
5.4K
WhiRL retweetledi
Zheng Xiong
Zheng Xiong@xiongzheng0316·
How to make a generalist robot more efficient? We propose knowledge decoupling as a key principle, and learn a universal morphology controller with 10x smaller size and 100x less FLOPs at inference time. Come to our #ICML2024 poster #217 at 11:30 on July 25 to chat more!
Zheng Xiong tweet mediaZheng Xiong tweet media
English
0
2
5
469
WhiRL retweetledi
Alex Goldie
Alex Goldie@AlexDGoldie·
1/ 🤖 Learned optimization offers huge potential to automate machine learning! So why doesn't it work well in RL (and how did we fix it)?! I'm excited to share OPEN, our @AutoRL_Workshop spotlight paper exploring this question! 🧵
English
1
27
115
25.1K
WhiRL retweetledi
Matthew Jackson
Matthew Jackson@JacksonMattT·
Exciting updates to Policy-Guided Diffusion! 🎉 PGD was accepted at @RL_Conference - see you in Amherst! 📈 For those building on PGD, we just released WandB logs with agent and diffusion model training: api.wandb.ai/links/flair/jo…
Matthew Jackson@JacksonMattT

🎮 Introducing the new and improved Policy-Guided Diffusion! Vastly more accurate trajectory generation than autoregressive models, with strong gains in offline RL performance! Plus a ton of new theory and results since our NeurIPS workshop paper... Check it out ⤵️

English
2
26
122
12.2K
WhiRL
WhiRL@whi_rl·
Excited to share that our work Bayesian Exploration Networks (BEN) has been accepted at ICML 🍾! BEN is the first model-free Bayesian RL approach that can learn Bayes-optimal policies 🙀 Congrats to @mattiefoxcs and collaborators! arxiv.org/pdf/2308.13049
English
0
0
5
750

Keşfet

@AlexDGoldie @zilinwang4ai @saeedrmd @FLAIR_Ox @uljadb99 @JarekLiesen @RL_Conference @jakeABeck