Barcelona, Catalonia, Spain
Scaling emergent communication to realistic deep network communities and tasks
Consistency of factual confidence in LLMs
Studied multi-agent reinforcement learning, focusing on the effect of communication topologies on an agent's capacity to innovate.
Implemented multimodal auto-encoders to predict cognitive scores from MRI brain imagery https://github.com/mahautm/INT_fMRI_processing
Studied VR in diagnosis on a 15 volunteers pilot study presented to investors Designed and coded User Interface to facilitate immersion Analysed learning curve and performance data for use of scientific director