UCL LASP Logo

LASP

University College London

PI:
Dr. Laura Toni

Department:
Electronic & Electrical Eng.

Location:
Malet Place Eng. Bldg

Learning & Signal Processing (LASP)@UCL

The Learning And Signal Processing group is a research lab led by Dr. Laura Toni. We investigate the fundamental limits of learning and information processing, operating at the intersection of Graph Machine Learning and Deep Reinforcement Learning.

Core Research Areas

  • Graph Generative Models: AI for drug discovery.
  • Autonomous Agents: Deep RL for complex, adaptive environments.
  • Decision Systems: Data-efficient bandits and reasoning.
  • AI4Science: AI applied to drug discovery, transcriptomics, and sustainability.

Latest News

  • Jan 2026: Paper Bures-Wasserstein Flow Matching for Graph Generation accepted at ICLR26!
  • Jan 2026: Paper Reinforcement Learning Using Known Invariances accepted at AISTATS26!
  • Dec 2025: 4 Papers presented at NeurIPS (various tracks/workshops). See you in SD! [View Post]
  • Nov 2025: Aya successfully defended her PhD thesis "Towards Data-Efficient AI". [View Post]

Our Team

Principal Investigators

Avatar

Laura Toni

Associate Professor

Researchers

Avatar

Han Wang

Phd student

Avatar

Keyue Jiang

Phd student

Avatar

Nagham Osman

Phd student

Alumni

Avatar

Alan Guedes

Assistant Professor

Avatar

Aya Kayal

Assistant professor at AUB

Avatar

Eduardo Pignatelli

Lecturer (teaching) UCL

Avatar

Kaige Yang

Reseacher at Amazon

Avatar

Pedro Gomes

Scientist at Fujitsu - London

Avatar

Sephora Madjiheurem

Reseacher at DeepMind

Recent Publications

See the full list: View all publications

Open Research Projects

We are looking for motivated students. Click the links below to see active projects.

Geometric & Graph Generative AI

Developing theoretically grounded models for molecule and 3D generation. Read More →

Fundamentals of RL & Agents

Tackling core challenges in RL like credit assignment and building robust Autonomous Agents. Read More →

Graph ML for Science

Applying Graph Learning to Drug Discovery and Transcriptomics to decode biological interactions. Read More →

AI for Sustainable Power Grids

Decarbonizing the energy sector using Graph Neural Networks to master grid complexity. Read More →

LLM Alignment & Exploration

Scaling Bayesian Optimization to improve post-training exploration in Large Language Models. Read More →

Agents for Optimization

Designing RL agents that learn to navigate the complex landscapes of Combinatorial Optimization. Read More →

Life at LASP

A glimpse into our culture. See more photos

Contact

  • l.toni@ucl.ac.uk
  • Gower Street, London, WC1E 7JE
  • Malet Place Engineering Building