Location
MC 6334
Candidate
Matheus Zambianco| Applied Mathematics, University of Waterloo
Title
New aspects of General Relativity and Quantum Information: Modelling Stellar Collapse with Quantum Energy Teleportation and using Transformers for Quantum Error Correction
Abstract
General Relativity is currently the most successful theory for our understanding of gravity and spacetime. Proposed by Albert Einstein more than 100 years ago, its predictions are still being proven accurate by cutting-edge experiments, exemplified by the detection of gravitational waves by LIGO in 2015 and the black hole picture obtained in 2022. However, there still is much to be understood when Quantum Physics comes into play. For years, many physicists have tried to obtain a complete theory of Quantum Gravity, but we still do not have one.
The field of Relativistic Quantum Information has emerged as a concrete approach for achieving a better understanding of nature in scenarios where both General Relativity and Quantum Physics are relevant. Using tools from Quantum Information, protocols such as Quantum Energy Teleportation (QET) and Entanglement Harvesting have emerged as new and interesting physics. In the case of QET, the possibility of generating spacetime regions with negative energy densities poses interesting questions, namely 1) can a QET protocol be used to induce or prevent a stellar collapse and 2) can a QET protocol be used to generate a spacetime with a stable wormhole or a warp drive? In this seminar, we will present research projects based on these questions.
On the other hand, the application of Machine Learning techniques in Physics has drawn a lot of interest lately. Particularly, the Transformer is widely recognized for its success in language modeling tasks, as exemplified by the large language model GPT-4. It turns out that quantum error-correcting codes have a structure that can be framed as a possible problem to be solved by Transformers. Thus, we will also present a research project to use Transformers for Quantum Error Correction.