Georgia Institute of Technology College of Sciences

The problem of partial permutation synchronization (PPS) provides a global mathematical formulation for the multiple image matching problem. In this matching problem, one is provided with possibly corrupted matches (i.e., partial permutations) between keypoints in pairs of images and the underlying task is to match keypoints in each image to universal 3D scene points (resulting in other partial permutations). For structure-from-motion (SfM) common datasets, previous PPS algorithms for image matching often become computationally intractable and demand an exceedingly large amount of memory. We address this issue by extending the recent framework of Cycle-Edge Message Passing (CEMP) to the setting of PPS despite the fact that partial permutations do not have a full group structure.  We emphasize mathematical difficulties that arise when extending CEMP to PPS and also explain the mathematical guarantees for the performance of the modified CEMP algorithm in the setting of adversarial corruption and sufficiently small noise. This is a joint work with Shaohan Li and Yunpeng Shi.

Generative networks have made it possible to generate meaningful signals such as images and texts. They were also extended to graphs and applied, for example, to generate molecules. However, the mathematical properties of generative methods are unclear, and training good generative models is difficult. Moreover, some basic and intuitive ideas of generative networks for signals and images do not apply to graphs and we thus focus on this talk on graph generation. An earlier joint work of the speaker generalized Mallat's scattering transform to graphs and later used it as an encoder within an autoencoder for graph generation (while applying a simple Gaussianization procedure to the output of the encoder) . For the graph scattering component, this work proved asymptotic invariance to permutations and stability to graph manipulations. The issue is that the decoder of this graph generation component used two fully connected networks and was not adapted to the graph structure. In fact, many other graph generation methods do not sufficiently utilize the graph structure. In order to address this issue, I will present a new recent joint work that develops a novel and trainable graph unpooling layer for effective graph generation. Given a graph with features, the unpooling layer enlarges this graph and learns its desired new structure and features. Since this unpooling layer is trainable, it can be applied to graph generation either in the decoder of a variational autoencoder or in the generator of a generative adversarial network (GAN). We establish connectivity and expressivity. That is, we prove that the unpooled graph remains connected and any connected graph can be sequentially unpooled from a 3-nodes graph. We apply the unpooling layer within the GAN generator and address the specific task of molecular generation. This is a joint work with Yinglong Guo and Dongmian Zou.

In the second talk of this seminar series, we continue to follow the manuscript of Carl Miller and building up concepts from algebraic topology. In particular, we will introduce chain complexes to define homology groups and provide some of the standard theory for them. 

Recent research on solving partial differential equations with deep neural networks (DNNs) has demonstrated that spatiotemporal-function approximators defined by auto-differentiation are effective    for approximating nonlinear problems. However, it remains a challenge to resolve discontinuities in nonlinear conservation laws using forward methods with DNNs without beginning with part of the solution. In this study, we incorporate first-order numerical schemes into DNNs to set up the loss function approximator instead of auto-differentiation from traditional deep learning framework such as the TensorFlow package, thereby improving the effectiveness of capturing discontinuities in Riemann problems. We introduce a novel neural network method.  A local low-cost solution is first used as the input of a neural network to predict the high-fidelity solution at a space-time location. The challenge lies in the fact that there is no way to distinguish a smeared discontinuity from a steep smooth solution in the input, thus resulting in “multiple predictions” of the neural network. To overcome the difficulty, two solutions of the conservation laws from a converging sequence, computed from low-cost numerical schemes, and in a local domain of dependence of the space-time location, serve as the input. Despite smeared input solutions, the output provides sharp approximations to solutions containing shocks and contact surfaces, and the method is efficient to use, once trained. It works not only for discontinuities, but also for smooth areas of the solution, implying broader applications for other differential equations.