Event Archive

The pandemic ushered in a time of unprecedented crisis in education, with students experiencing months of interrupted or unfinished learning. Student performance data shows that all students’ learning has been affected, but the impact is far greater for some student populations, particularly students of color and students from low-income backgrounds.

Hear executive insights from Srinivasan “Chini” Krishnan, MS '91, PhD '92 (Computer Science), the founder and CEO of GetInsured.com, a technology company partnering with states to provide state-based healthcare exchanges to their citizens. Today, the company is the leading provider of health insurance enrollment technology for the state governments of California, Idaho, Minnesota, Pennsylvania, Nevada, New Jersey and Washington. GetInsured technology has enabled more than 15 million enrollments to date. The event will feature Dr.

Thanks to neural networks (NNs), faster computation, and massive datasets, machine learning (ML) is under increasing pressure to provide automated solutions to even harder real-world tasks beyond human performance with ever faster response times due to potentially huge technological and societal benefits.

The All-Pairs Max-Flow problem (APMF) asks to compute the maximum flow (or equivalently, the minimum cut) between all pairs of nodes in a graph. The naive solution of making n^2 calls to a (single-pair) max-flow algorithm was beaten in 1961 by a remarkable algorithm of Gomory and Hu that only makes n-1 calls. Within the same time bound, their algorithm also produces a cut-equivalent tree (a.k.a. GH-Tree) that preserves all pairwise minimum cuts exactly.

CS Department members, please join us for this special opportunity to come together in fellowship and welcome our newest faculty, graduate students, undergraduate students, postdocs, and staff. The meeting will begin at 12:00 pm, with remarks from the department chair and presentation of awards. Boxed lunches will be provided for you to take with you and enjoy.

We hope you can attend this special event in-person, but if not, please join us via Zoom! The Zoom link will be emailed to CS department members.

Fueled by massive amounts of data, models produced by machine-learning (ML) algorithms, especially deep neural networks (DNNs), are being used in diverse domains where trustworthiness is a concern, including automotive systems, finance, healthcare, natural language processing, and malware detection. Of particular concern is the use of ML algorithms in cyber-physical systems (CPS), such as self-driving cars and aviation, where an adversary can cause serious consequences. Interest in this area of research has simply

In the first part we cover five current specific challenges through examples: (1) discrimination (e.g., facial recognition, justice, sharing economy, language models); (2) phrenology (e.g., biometric based predictions); (3) unfair digital commerce (e.g., exposure and popularity bias); (4) stupid models (e.g., Signal, minimal adversarial AI) and (5) indiscriminated use of computing resources (e.g., large language models). These examples do have a personal bias but set the context for the second part where we address four generic challenges: (1) too many principles (e.g., principles vs.

In this paper we study the training dynamics for gradient flow on over-parametrized tensor decomposition problems. Empirically, such training process often first fits larger components and then discovers smaller components, which is similar to a tensor deflation process that is commonly used in tensor decomposition algorithms. We prove that for orthogonally decomposable tensor, a slightly modified version of gradient flow would follow a tensor deflation process and recover all the tensor components.

Voronoi and Delaunay refinement are standard algorithmic techniques used in surface reconstruction and mesh generation in Euclidean spaces.  For point sets, the resulting output is a triangulation of a superset of the points in which every triangle has angles larger than a user-defined constant.  This guarantees that the total complexity is linear in the number of points (even in dimensions greater than two).  These simple refinement heuristics are local updates that naturally produce asymptotically optimal size outputs.  In this talk, I will discuss how one can translate

Quantum many-body systems are very hard to simulate, since computational resources (time and memory) typically grow exponentially with the system size. However, quantum computers or analog quantum simulators may perform that task in a much more efficient way. In this talk, Professor Ignacio Cirac will review some of the quantum algorithms that have been proposed for this task and then explain the advantages and disadvantages of analog quantum simulators.

The Demikernel is a new datapath OS and architecture for microsecond-scale datacenter systems and kernel-bypass devices. Demikernel accommodates heterogenous kernel-bypass devices with a flexible library OS architecture and new high-level datapath API with an asynchronous I/O interface and zero-copy memory semantics for microsecond I/O processing. Demikernel implements this API for RDMA, DPDK and SPDK devices with new nanosecond-scale I/O stacks in Rust.

Relational queries, and in particular joins, often generate large output results when executed over a huge dataset. In such cases, it is prohibitively expensive to store the whole materialized output if we plan to reuse it further down a data processing pipeline. In this talk, I will present several techniques that allow us to avoid this materialization step and discuss their applications.

Register Now
Interested in Computer Science graduate school? Have questions? Faculty, PhD students, and representatives from the computer science programs at Duke, NC State, and UNC will come together to help demystify the grad school process. Hear first-hand accounts about applying to graduate schools, path to graduation, and career opportunities.  

In this talk, I will revisit some fundamental problems in computational geometry related to range searching.  I will describe new algorithms for "Hopcroft's problem" -- finding incidences between n points and n lines in 2D -- with O(n^{4/3}) run time, improving Matousek's previous result from 30 years ago by a 2^{O(log^*n)} factor.  The new techniques have numerous other applications (in 2D and higher), allowing us to remove extraneous factors from many known theoretical time bounds.

Joint work with Da Wei Zheng (to appear in SODA'22).

Many have predicted the future of the Web to be the integration of Web content with the real-world through technologies such as augmented reality. This overlay of virtual content on top of the physical world, called the Spatial Web (in different contexts AR Cloud, MetaVerse, Digital Twin), holds promise for dramatically changing the Internet as we see it today, and has broad applications.

SECURITY READING GROUP

I will discuss the Steiner forest problem in the Euclidean plane, where the input is a multiset of points, partitioned into k color classes,  and the goal is to find a minimum-cost Euclidean graph G such that every color class is connected. We study this problem in dynamic streams,  where the input is provided by a stream of insertions and deletions of colored points from the discrete grid [\Delta]^2. 

Given a pair of graphs, the problem of graph matching or network alignment refers to finding a bijection between the vertex sets so that the edge sets are maximally aligned. This is a ubiquitous problem arising in a variety of applications across diverse fields such as network privacy, computational biology, computer vision, and natural language processing. Graph matching is an instance of the notoriously difficult quadratic assignment problem (QAP), which is NP-hard to solve or to approximate.