25 accepted papers
Peer-to-peer learning is an increasingly popular framework that enables beyond-5G distributed edge devices to collaboratively train deep neural networks in a privacy-preserving manner without the aid of a central server. Neural network training algorithms for emerging environments, e.g., smart citie…
Uncertainty quantification of neural networks is critical to measuring the reliability and robustness of deep learning systems. However, this often involves costly or inaccurate sampling methods and approximations. This paper presents a sample-free moment propagation technique that propagates mean v…
Graph Convolutional Neural Networks (graph CNNs) adapt the traditional CNN architecture for use on graphs, replacing convolution layers with graph convolution layers. Although similar in architecture, graph CNNs are used for geometric deep learning whereas conventional CNNs are used for deep learnin…
Many Graph Signal Processing (GSP) applications consider product graphs, the product of smaller graphs. For example, with time-varying graph data, the graph shift can be the (Cartesian) product of a space graph and the cyclic time shift. Instead of treating the product graph as a single entity and a…
In a Networked Dynamical System (NDS), each node is a system whose dynamics are coupled with the dynamics of neighboring nodes. The global dynamics naturally builds on this network of couplings and it is often excited by a noise input with nontrivial structure. The underlying network is unknown in m…
This paper considers learning the hidden causal network of a linear networked dynamical system (NDS) from the time series data at some of its nodes -- partial observability. The dynamics of the NDS are driven by colored noise that generates spurious associations across pairs of nodes, rendering the…
While deep learning has been very successful in computer vision, real world operating conditions such as lighting variation, background clutter, or occlusion hinder its accuracy across several tasks. Prior work has shown that hybrid models—combining neural networks and heuristics/algorithms—can outp…
While much effort has been devoted to deriving and analyzing effective convex formulations of signal processing problems, the gradients of convex functions also have critical applications ranging from gradient-based optimization to optimal transport. Recent works have explored data-driven methods fo…
We study the problem of graph structure identification, i.e., of recovering the graph of dependencies among time series. We model these time series data as components of the state of linear stochastic networked dynamical systems. We assume partial observability, where the state evolution of only a s…
Few-shot classification aims at classifying categories of a novel task by learning from just a few (typically, 1 to 5) labeled examples. An effective approach to few-shot classification involves a prior model trained on a large-sample base domain, which is then finetuned over the novel few-shot task…
Unsupervised time series clustering is a challenging problem with diverse industrial applications such as anomaly detection, bio-wearables, etc. These applications typically involve small, low-power devices on the edge that collect and process real-time sensory signals. State-of-the-art time-series…
A feature-based model explanation denotes how much each input feature contributes to a model's output for a given data point. As the number of proposed explanation functions grows, we lack quantitative evaluation criteria to help practitioners know when to use which explanation function. This paper…
In this paper, we present a new approach to interpret deep learning models. By coupling mutual information with network science, we explore how information flows through feedforward networks. We show that efficiently approximating mutual information allows us to create an information measure that qu…
This paper studies the resilient distributed recovery of large fields under measurement attacks, by a team of agents, where each measures a small subset of the components of a large spatially distributed field. An adversary corrupts some of the measurements. The agents collaborate to process their m…
Internet of Things (IoT) applications for Smart Cities, such as systems for traffic control and pollution monitoring, increasingly rely on trustworthy and secure data analytics. Proper countermeasures are needed to ensure that IoT applications function reliably under security threats. This paper stu…
While domain adaptation has been actively researched, most algorithms focus on the single-source-single-target adaptation setting. In this paper we propose new generalization bounds and algorithms under both classification and regression settings for unsupervised multiple source domain adaptation. O…
While many approaches have been proposed to estimate and remove blur in a photo, few efforts were made to have an algorithm automatically understand the blur desirability: whether the blur is desired or not, and how it affects the quality of the photo. Such a task not only relies on low-level visual…
Teaching a robot to predict and mimic how a human moves or acts in the near future by observing a series of historical human movements is a crucial first step in human-robot interaction and collaboration. In this paper, we instrument a robot with such a prediction ability by leveraging recent deep l…
Network-based epidemics models try to characterize the impact of network topology, which represents contagion pathways, on the spread of infection. Although these models explicitly consider the dynamics of individuals in the given network (i.e., the state of the system is x(t)=[x <sub xmlns:mml="htt…
Gaussian belief propagation (BP) has been widely used for distributed estimation in large-scale networks such as the smart grid, communication networks, and social networks, where local meansurements/observations are scattered over a wide geographical area. However, the convergence of Gaussian BP is…
Design of filters for graph signal processing benefits from knowledge of the spectral decomposition of matrices that encode graphs, such as the adjacency matrix and the Laplacian matrix, used to define the shift operator. For shift matrices with real eigenvalues, which arise for symmetric graphs, th…
Finding densely connected subgraphs, also called communities, in networks are of interest for many applications. In previous work, we showed an optimization method for efficiently finding subgraphs denser than the overall network [1]. This result is derived from our studies of network processes, dyn…
A class of models for describing sets of time series generated by interacting agents using directed, weighted graphs is introduced. A computationally tractable algorithm for estimating the graph adjacency matrix of this model from observed time series data is presented. The performance guarantees of…
We study cyber-physical systems subject to dynamic sensor attacks, relating them to the system's strong observability. First, we find necessary and sufficient conditions for an attacker to create a dynamically undetectable sensor attack and relate these conditions to properties of the system dynamic…
This paper presents a computationally tractable algorithm for estimating the graph structure of graph signals is presented. The algorithm is demonstrated on simulated and real network time series datasets, and the performance of the new method is compared to that of related methods for estimating gr…