LASP's era is coming to an end, but only because of a new and incredibly exciting opportunity🚀
Strongly believing 1+1=3, we are merging @UCL_LASP, @FLAIR_Ox, @whi_rl, @UCL_DARK and AIRL in favour of BOLD — the British Open-ended Learning and Discovery Lab (@BOLD_Lab_AI)
🧵 1/5
4/5
BOLD is our commitment to open-ended learning, discovery, and ambitious fundamental AI research🌍🧠
We are incredibly excited for what comes next, to build cutting-edge research, and to do it together with such an amazing dream-team @bold_lab_ai 🚀✨
AI4Science Workshop: “From In Silico to In Vitro: Evaluating Molecule Generative Models for Hit Generation.”
We answer: Can generative models generate hit-like molecules? We build a multi-stage evaluation and benchmark models across datasets.
👥Osman, Lembo, Bottegoni, Toni
🎉 Excited to share that our team has 4 papers accepted at #NeurIPS2025!
🧠 1 Main Conference paper
📊 1 Datasets & Benchmarks paper
🔬 1 AI4Science Workshop paper
🕸️ 1 New Perspectives on Graph ML Workshop paper
Huge thanks to all collaborators — see you in San Diego! 🚀 #ML#AI
NPGML Workshop: “LGDC: Latent Graph Diffusion via Spectrum-Preserving Coarsening.”
We propose a diffusion model that generates graphs in a latent space built via key properties spectrum-preserving coarsening.
📄Paper: https://t.co/hmMPin1fs0
👥Osman, Jiang, Buffelli, Dong, Toni
An amazing opportunity to become part of our LASP team!
Working on the bridge between generative AI, causality, and network systems!
Join our team to explore the future of AI ✨
⏲️Deadline for applications: 5 January 2026, 1pm UK time
📄 How to apply: https://t.co/L9sd3e5MIP
New PhD opportunity at @UCL_LASP to explore if AI could understand the why beyond global events!
Join us to build smarter, more resilient systems for our future.
In collaboration with Francesca Boem & @mircomusolesi https://t.co/uNYGFpSMii
📢 ⚡️Glad to finally release #NAVIX: we re-implemented MiniGrid in #JAX for lightning-fast RL training🚀
You can now train more than 2048 #PPO agents on #MiniGrid environments in less than *2* minutes! ⚡️⚡️
Say goodbye to long and tedious hparams search! 👋👋
📢Really excited to present our work on #CreditAssignment with #LanguageModels as a #spotlight at the #ICML@London meetup.
📅Happening tomorrow, Fri, 12th Jul, 3pm BST,
📍UCL Cruciform Building, LT1
Come say hi! 👋
📢!
Tomorrow (27th) I will be presenting our
- *Survey on Credit Assignment* and our work on
- *Credit Assignment with Language models* (SPOILER🚩) at @UoE_Agents (@InfAtEd) reading group.
Thank you, @UoE_Agents for the kind invite!
Tune in, if interested! (links below)
Very excited to share this #survey focused on #CreditAssignment in #DeepRL🥳👏
Check out the great work and efforts done by one of our PhD students @edupignatelli along with his outstanding collaborators and supervisor @tonizza82 🤩
📖https://t.co/rZ66JFRupp
Glad to share our latest work "A survey of Temporal Credit Assignment in Deep RL".
We bridge the gap between the fundamental role of Credit Assignment in RL and the little discussion on its precise mathematical nature:
https://t.co/90Xnvq9yMu
🎉Huge congratulations to @SilviaRossi24, our @UCL_LASP PhD alumni, on receiving the ACM @SIGMM Award for her thesis on user interactivity in #VR systems 📹👓
Read more on her research ⬇️
https://t.co/Zx3IO1Rlcs
🎉Join us in celebrating the achievements of our BEng student, Iason Chaimalas (https://t.co/XbTAHdhNFF)
🚀 You can learn more about his work with #BBC#Research & Development on overcoming the cold-start problem in #recommender systems in this article:
https://t.co/dojBTjqj4w
LASP Alumni workshop 🚀
Two days of interesting talks from #VR to #RL and #graph learning, discussion on #PhD life and mentoring
...and it's always nice being back at @UCLEngineering with old friends and colleagues :)
Thanks to @tonizza82 for organising this☺️