Hello, and welcome to my home on the web! I’m Petar, a Staff Research Scientist at Google DeepMind, Affiliated Lecturer at the University of Cambridge, and an Associate of Clare Hall, Cambridge. I hold a PhD in Computer Science from the University of Cambridge (Trinity College), obtained under the supervision of Pietro Liò. My research concerns geometric deep learning—devising neural network architectures that respect the invariances and symmetries in data (a topic I’ve co-written a proto-book about). For my contributions, I am recognised as an ELLIS Scholar in the Geometric Deep Learning Program. Particularly, I focus on graph representation learning and its applications in algorithmic reasoning (featured in VentureBeat). I am the first author of Graph Attention Networks—a popular convolutional layer for graphs—and Deep Graph Infomax—a popular self-supervised learning pipeline for graphs (featured in ZDNet). My research has been used in substantially improving travel-time predictions in Google Maps (featured in the CNBC, Endgadget, VentureBeat, CNET, the Verge and ZDNet), and guiding intuition of mathematicians towards new top-tier theorems and conjectures (featured in Nature, Science, Quanta Magazine, New Scientist, The Independent, Sky News, The Sunday Times, la Repubblica and The Conversation).
Our survey on Scientific discovery in the age of AIhas been published in Nature!
Want to know more about GNNs? Check out my latest survey titledEverything is Connected: Graph Neural Networksnow accepted in Current Opinion in Structural Biology!
We successfully trained One GNN to simultaneously execute thirty algorithmsnow accepted to LoG 2022 (as an oral)!
Proud to share that we have been awarded aBest Paper Award for “Expander Graph Propagation”at the NeurIPS'22 GLFrontiers workshop! EGP is also accepted to LoG 2022.