Talks

I will discuss the computation of second-order terms in the entanglement entropy and subregion complexity for a spherical entangling region in the AdS black hole background relative to pure AdS. I will suggest an extension of the conjectured relationship between subregion complexity and Fisher information into a relation that is reminiscent of the first law of thermodynamics. By analogy, entanglement and complexity play the roles of heat and work, respectively. Time permitting, I will also discuss the computation of third- and fourth-order terms in the relative entropy. 

This is practice for a talk at Berkeley, so it will involve explaining stuff many people here probably already know.  I'll try to summarize what I've learned about 3 dimensional N=4 supersymmetric quantum field theories, their twists and how these manifest in terms of interesting objects in mathematics.  If nothing else, hopefully there will be some comedy value in my attempt.  

In this talk, I will discuss some interesting connections between Hamiltonian complexity, error correction, and quantum circuits. First, motivated by the Quantum PCP Conjecture, I will describe a construction of a family of local Hamiltonians where the complexity of ground states — even when subject to large amounts of noise — is superpolynomial (under plausible complexity assumptions). The construction is simple, making use of the well-known Feynman-Kitaev circuit Hamiltonian construction.

Next, I will describe how this complexity result can be repurposed to construct local approximate error correcting codes with (nearly) linear distance and (nearly) constant rate. By contrast, it remains an open problem to construct quantum low-density parity check (QLDPC) stabilizer codes that achieve similarly good parameters.

An interesting component of our code construction is the Brueckmann-Terhal spacetime circuit Hamiltonian construction, and a novel analysis of its spectral gap. Our codes are ground spaces of non-commuting  — but frustration-free -- local Hamiltonians. We believe that our results present additional evidence that the study of approximate error correction and non-stabilizer codes may be much richer than previously thought.

This talk will have more questions than answers, and is based on two papers:

·  https://arxiv.org/abs/1802.07419 (joint with Chinmay Nirkhe and Umesh Vazirani)

·  https://arxiv.org/abs/1811.00277 (joint work with Thom Bohdanowicz, Elizabeth Crosson, and Chinmay Nirkhe)

Gravity theories naturally allow for edge states generated by non-trivial boundary-condition preserving diffeomorphisms. I present a specific set of boundary conditions inspired by near horizon physics, show that it leads to soft hair excitations of black hole solutions and discuss implications for black hole entropy.

So far artificial neural networks have been applied to discover phase diagrams in many different physical models. However, none of these studies have revealed any fundamentally new physics. A major problem is that these neural networks are mainly considered as black box algorithms. On the journey to detect new physics it is important to interpret what artificial neural networks learn. On the one hand this allows us to judge whether to trust the results, and on the other hand this can give us insight to possible new physics. In this talk I will 
discuss applications to different models where we successfully interpreted what was learned by the neural networks.