Research

Quantum Computing

Papers:

Generating k EPR pairs from an n-party resource state (with Sergey Bravyi, Mario Szegedy and Ronald de Wolf)

Currently research questions

Why are there so many different measures for entanglement?

What is the deal with multipartite entanglement and why is it difficult to pin down?

Machine Learning

Lets think. So LLMs are going good. Transformers are being applied for solving Linear Algebra and undergrad math. Graph neural networks are doing good on CLRS benchmark. AlphaTensor found a more optimal algorithm for matrix multiplication. SayCan is a good step for robotics. Great. What problems cannot be solved by scaling data and model? I think I will work on causality. For AI to be creative thinking like coming up with new ideas. How to make an enterpreneurial AI that will find problems and work on them? Maybe the big thing we need is just AI that can extremely cheaply build roads, lay pipelines and electricity lines in developing or under-developed places and also monitor and supply maintanence for those constructions?