Georgia Institute of Technology College of Sciences

A neural code C on n neurons is a collection of subsets of {1,2,...,n} which is used to encode the intersections of subsets U_1, U_2,...,U_n of some topological space. The study of neural codes reveals the ways in which geometric or topological properties can be encoded combinatorially. A prominent example is the property of max-intersection completeness: if a code C contains every possible intersection of its maximal codewords, then one can always find a collection of open convex U_1, U_2,..., U_n for which C is the code. In this talk I will answer a question posed by Curto et al. (2018), which asks if there is a way of determining max-intersection completeness from examination of the neural ideal, an algebraic counterpart to the neural code.

In this presentation, I will discuss my research in the field of data science, specifically in two areas: improving autoencoder interpolations and accelerating federated learning algorithms. My work combines advanced mathematical concepts with practical machine learning applications, contributing to both the theoretical and applied aspects of data science. The first part of my talk focuses on image sequence interpolation using autoencoders, which are essential tools in generative modeling. The focus is when there is only limited training data. By introducing a novel regularization term based on dynamic optimal transport to the loss function of autoencoder, my method can generate more robust and semantically coherent interpolation results. Additionally, the trained autoencoder can be used to generate barycenters. However, computation efficiency is a bottleneck of our method, and we are working on improving it. The second part of my presentation focuses on accelerating federated learning (FL) through the application of Anderson Acceleration. Our method achieves the same level of convergence performance as state-of-the-art second-order methods like GIANT by reweighting the local points and their gradients. However, our method only requires first-order information, making it a more practical and efficient choice for large-scale and complex training problems. Furthermore, our method is theoretically guaranteed to converge to the global minimizer with a linear rate.

It has been known since the pioneering work of J.L. Lions in 1960s that 3D hyper-dissipative (HD) Navier-Stokes (NS) system exhibits global-in-time regularity as long as the hyper-diffusion exponent is greater or equal to 5/4.  One should note that at 5/4, the system is critical, i.e., the energy level and the scaling -invariant level coincide. What happens in the super-critical regime, the hyper-diffusion exponent being strictly between 1 and 5/4 remained a mystery. 

The goal of this talk is to demonstrate that as soon as the hyper-diffusion exponent is greater than 1, a class of monotone blow-up scenarios consistent with the analytic structure of the flow (prior to the possible singular time) can be ruled out (a sort of 'runaway train' scenario). The argument is in the spirit of the regularity theory of the 3D HD NS system in 'turbulent scenario' (in the super-critical regime) developed by Grujic and Xu, relying on 'dynamic interpolation' – however, it is much shorter, tailored to the class of blow-up profiles in view. This is a joint work with Aseel Farhat.

Title: Topology, geometry and adaptivity in soft and living matter

Abstract:

Topology and adaptivity play fundamental roles in controlling the dynamics of biological and physical systems, from chromosomal DNA and biofilms to cilia carpets and worm collectives. How topological rules govern the self-adaptive dynamics of living matter remains poorly understood. Here we investigate the interplay between topology, geometry and reconfigurability in knotted and tangled matter. We first identify topological counting rules which predict the relative mechanical stability of human-designed knots, by developing a mapping between elastic knots and long-range ferromagnetic spin systems. Building upon this framework, we then examine the adaptive topological dynamics exhibited by California blackworms, which form living tangled structures in minutes but can rapidly untangle in milliseconds. Using blackworm locomotion datasets, we construct stochastic trajectory equations that explain how the dynamics of individual active filaments controls their emergent topological state. To further understand how tangled matter, along with more general biological networks, adapt to their surroundings, we introduce a theory of adaptive elastic networks which can learn mechanical information. By identifying how topology and adaptivity produce stable yet responsive structures, these results have applications in understanding broad classes of adaptive, self-optimizing biological systems.

Zoom: https://gatech.zoom.us/j/93619173236?pwd=ZGNRZUZ2emNJbG5pRzgzMnlFL1dzQT09

Braid groups are relatively simple to describe, but they have deep and intricate connections to many different areas of math. We will discuss three specific instances where the braid group on 3 strands arises in geometry and knot theory. In exploring connections between these perspectives, we will take a detour into the world of elliptic curves and their moduli space. As a result, we will see that these three perspectives are actually the same. Time permitting, we will explore generalizations of this to the braid group on n strands for n > 3.

Abstract: Classifying the invariant measures for a given dynamical system represents a fundamental challenge.

 Every graph is associated to a symmetric function constructed from proper colorings of the graph.  The Stanley-Stembridge conjecture posits that the expansion of the chromatic symmetric function into the elementary symmetric functions has positive coefficients for a certain class of graphs.  We explore a potential new approach to the Stanley-Stembridge Conjecture using combinatorial objects called "special rim hooks" and connect this to the "chromatic quasisymmetric functions" introduced by Shareshian and Wachs as a generalization of chromatic symmetric functions.  This is joint work with Meagan Hodge.