Talks

May 23 (Tue)


Time: 19:00 - 19:40 (40min)

Title: de Sitter self-organization and information problem

Abstract: Quantum fluctuations in the de Sitter stage of early universe can be used to explain the origin of unnatural parameters in the universe -- Higgs mass. But this mechanism is also subject to the dS version of information problem.


Time: 20:00 - 20:40 (40min)

Title: Langevin-type equation to describe nonequilibrium dynamics of strongly-coupled system

Abstract: Traditional thermodynamics typically focuses on energy transformation in a macroscopic system coupled to an equilibrium environment. The thermodynamic properties of an open system are insensitive to the details of the system-environment coupling. However, in microscopic open systems, thermodynamic observables should be modified due to the non-negligible effects of the interaction Hamiltonian, which can be comparable to the energy scale of the microscopic system. These modifications, known as strong-coupling effects, have been extensively studied. But their impact on systems far from equilibrium has been hardly investigated due to the limitation of the conventional formalism based on the Langevin equation, which describes only the dynamics of weakly-coupled systems. In this presentation, we derive a generalized equation of motion that describes the classical diffusive dynamics with strong coupling effects. In the equation of motion, the strong-coupling effects are indicated by an additional mean force and interaction-dependent damping coefficients. We demonstate that the conventional Langevin equation is restored when the bath is homogeneous and the length scale of the system-bath interaction is short. Our findings are validated though analytical and numerical analyses of solvable examples. This formalism may pave the way to understanding the thermodynamics of strongly coupled systems where traditional thermodynamics cannot be applied.


May 24 (Wed)


Time: 10:00 - 10:40 (40min)

Title: Towards the Physics of Artificial Intelligence

Abstract: In this informal talk, I hope to discuss some of the contributions theoretical physicists are making to the current frontiers in artificial intelligence and machine learning. While there are ever increasing intersections between physics and AI, this talk will focus on "Physics for AI" where ways of thinking and tools in physics provide a novel perspective in the development of AI. I will start with my highly personal biased view on current trends in artificial intelligence based on large scale deep neural networks. The plan is to deliver high level important ideas and contributions on these trends stemming from physicists and physics minded researchers.


Time: 11:00 - 11:40 (40min)

Title: A supersymmetric model of information loss and recovery

Abstract: TBA


Time: 19:00 - 19:40 (40min)

Title: Mirror descent of Hopfield model

Abstract: Mirror descent is an elegant optimization technique that leverages a dual space of parametric models to perform gradient descent. While originally developed for convex optimization, it has increasingly been applied in the field of machine learning. In this study, we propose a novel approach for utilizing mirror descent to initialize the parameters of neural networks. Specifically, we demonstrate that by using the Hopfield model as a prototype for neural networks, mirror descent can effectively train the model with significantly improved performance compared to traditional gradient descent methods that rely on random parameter initialization. Our findings highlight the potential of mirror descent as a promising initialization technique for enhancing the optimization of machine learning models


Time: 20:00 - 20:40 (40min)

Title: Quantum Simulations of Integrable Models

Abstract: One of the most promising applications of quantum computers is the quantum simulations of physical systems. In this talk I will discuss why we are interested in integrable models in this context and what we might hope to achieve there, based on my experience in a recent paper arXiv: 2208.00576 [quant-ph].

May 25 (Thu)


Time: 10:00 - 10:40 (40min)

Title: Beyond the Thermodynamic Second Law

Abstract: Since the late 1990s, thermodynamics has been reformulated in terms of a stochastic trajectory. This has led to a renaissance in thermodynamics, in which the thermodynamic second law is generalized to the fluctuation theorems (FT) by using the definition of stochastic entropy, providing a theoretical framework for general nonequilibrium processes. Moreover, since 2015, there has been a new wave of studies investigating fundamental thermodynamic inequalities beyond the thermodynamic second law, which have set new horizons for our understanding of thermodynamic processes. These inequalities are thermodynamic uncertainty relations (TUR) and thermodynamic speed limit (TSL), which are tradeoff relations associated with thermodynamic cost. These FT and tradeoff relations are major pillars of stochastic thermodynamics. In this talk, I will introduce the basic concepts of stochastic entropy production and the meaning of these major theories.


Time: 11:00 - 11:40 (40min)

Title: TARGET SPACE ENTANGLEMENT ENTROPY AND FINITENESS

Abstract: I will discuss a notion of entanglement among internal degrees of freedom in field theories of matrices which appear in gauge gravity duality and the origins of finiteness of this quantity.


Time: 10:00 - 10:40 (40min)

Title: Eigenstate thermalization in the presence of SU(2) symmetry

Abstract: TBA


Time: 20:00 - 20:40 (40min)

Title: Bulk-edge correspondence for non-unitary TQFTs/RCFTs

Abstract: TBA

May 26 (Fri)


Time: 19:00 - 19:40 (40min)

Title: Lower-dimensional gravity models and the SYK model

Abstract: I will discuss recent developments in our understanding of lower-dimensional gravity models. In particular, I will highlight how the SYK model in a double-scaling limit has a very similar structure as Jackiw-Teitelboim (JT) gravity. Based in part on upcoming work with A. Blommaert and S. Yao.


Time: 20:00 - 20:40 (40min)

Title: On spectrum of BPS strings in 6d supergravities

Abstract: TBA

May 27 (Sat)


Time: 10:00 - 10:40 (40min)

Title: Advances in the Score-based Generative Models: Theory and Application

Abstract: Diffusion models have recently acquired significant attention in the field of generative modeling of machine learning research due to their various theoretical advantages and remarkable applications in artificial intelligence, such as Stable Diffusion and DALL-E. In this presentation, we first introduce the theoretical background of the score-based diffusion models and present the latest results of their applications to machine learning. We also present advanced score-based generative models based on the time reversal theory of Lévy processes and diffusion theory in Hilbert space.


Time: 11:00 - 11:40 (40min)

Title: Target Space Entanglement

Abstract: We define a notion of Target Space Entanglement in Matrix Theory and study its properties.


Time: 19:00 - 19:40 (40min)

Title: On the dark side of the Universe

Abstract: In this talk, I discuss various topics on the dark side of the Universe from particle / astroparticle physics and cosmology, focusing on dark matter, dark radiation and force mediators such as dark photon and dark Higgs boson.


Time: 20:00 - 20:40 (40min)

Title: Primary Fields in Free CFT

Abstract: In this talk we will explain how to count and then construct the primary fields that appear in a product of free fields. We do this for scalar fields, for vector models and for matrix models.