RDMA networks enable low latency and low CPU utilization, and their widespread adoption in datacenters enables improved application performance. However, there are performance isolation concerns for RDMA deployed in a shared cloud environment. In particular, we find that congestion control enforcement and congestion control algorithms in RDMA make the network susceptible to performance hacking attacks, which give the attacker extra bandwidth and cause severe congestion in the network.
In my talk, I will describe the work that I have been doing since March 2020, leading a multi-disciplinary team of 20+ volunteer scientists working very closely with the Presidency of the Valencian Government in Spain on 4 large areas: (1) human mobility modeling; (2) computational epidemiological models (both metapopulation, individual and LSTM-based models); (3) predictive models; and (4) citizen surveys via the COVID19impactsurvey with over 600,000 answers worldwide.
Minimum cut problems are among the most well-studied questions in combinatorial optimization. In this talk, I will introduce a simple but powerful new tool for solving minimum cut problems called the isolating cuts lemma. I will show how this tool can be employed to obtain faster algorithms for several fundamental min-cut problems, namely global min-cut, Steiner min-cut, and all-pairs min-cut. For these problems, the new results represent the first improvement in their runtimes in several decades.
Computational modeling of proteins is more biophysically accurate when modeling protein thermodynamic ensembles. Principled design algorithms should exploit statistical thermodynamics of non-covalent binding, and therefore require approximation of partition functions over protein conformation space. The MARK* algorithm in the OSPREY protein design software suite allows not only provable approximation of the partition function and estimation of binding affinity, but also approximation and visualization of protein energy landscapes.
Every representative democracy must specify a mechanism under which voters choose their representatives. The most common mechanism in the United States -- winner-take-all single-member districts -- both enables substantial partisan gerrymandering and constrains `fair' redistricting, preventing proportional representation in legislatures. We study the design of multi-member districts (MMDs), in which each district elects multiple representatives, potentially through a non-winner-takes-all voting rule. We carry out large-scale analyses for the U.S.
Accessibility is not new, and neither is ableism: Dispelling disability myths and putting accessibility into action
This second lecture in the Identity & Computing Lecture Series: Understanding Racism and Bias in Computing welcomes Dr. Michele Williams, who explains how people in computing must ensure inclusion of people with disabilities. This means teaching and creating technology inclusive of people with disabilities, and going deeper in understanding how ableism causes non-disabled people to not consider disabled people in the first place.
In the edge connectivity augmentation problem, given a weighted undirected graph and a target connectivity tau, we want to increase the graph's connectivity to tau by adding edges with minimum total weight. Our new algorithm uses poly-logarithmic calls to any max-flow algorithm, which yields a running time of O((m+n^(3/2)) polylog(n)) and improves on the previous best time of O(n^2 polylog(n)). We also obtain an identical improvement in the running time of the closely related edge splitting off problem in undirected graphs.
Duke TechConnect brings students and employers together for networking, education and connections. At TechConnect on September 22, 2021, employers will connect with Engineering and Computer Science students in a virtual networking environment powered by Handshake.
The inaugural lecture in the Identity & Computing Lecture Series: Understanding Racism and Bias in Computing. The landscape of information is rapidly shifting as new imperatives and demands push to the fore increasing investment in digital technologies. Yet, critical information scholars continue to demonstrate how digital technology and its narratives are shaped by and infused with values that are not impartial.
Duke Computer Science Professor Nicki Washington will be featured in a webinar today on Reimagining Education: Conversations on Character, Commitment & Community. Join Suzanne Shanahan, Nannerl O. Keohane Director of the Kenan Institute for Ethics and Associate Research Professor in Sociology and Nicki Washington for their conversation: “Why should computer science care about identity?”
Registration is required to attend this webinar:
Duke CS 2021 Graduates: Join our virtual celebration at 2 PM EDT on Sunday, May 2! The event will be live streamed over YouTube for friends and family, and CS graduates will be notified by email with the Zoom link to participate. CS faculty will welcome you to this Zoom meeting where you can interact with your peers, enjoy messages from CS alumni, hear graduates' names and degrees, awards, have some fun, and so much more. Don't miss it! https://graduation2021.cs.duke.edu
Deep Neural Networks (DNNs) enable computers to excel across many different applications such as image classification, speech recognition and robotics control. To accelerate DNN training and serving, parallel computing is widely adopted. System efficiency is a big issue when scaling out. In this talk, I will make three arguments towards better system efficiency in distributed DNN training and serving.
The mission of the Software and Systems Division (SSD) at the National Institute of Standards and Technology (NIST) is to work with industry, academia and other government agencies to accelerate the development and adoption of correct, reliable, testable software, leading to increased trust and confidence in deployed software; promulgate methods to develop better standards and testing tools for today's software infrastructures and tomorrow's next-generation software systems; advance the state of the art of software testing by developing scientifically rigorous, breakthrough techniques
Zoom’s platform provides video conferencing services for hundreds of millions of daily meeting participants. They use Zoom to conduct business, learn among classmates scattered by recent events, connect with friends and family, collaborate with colleagues, and in some cases, discuss critical matters of state. Zoom is working hard to improve meeting security for its users.
The early designers of the Internet fostered tremendous innovation by leaving much of the network’s functionality to the programmable computers at its periphery. Unfortunately, the *inside* of the network has been much harder to change. Yet, changing the network is important to make the Internet more reliable, secure, performant, and cost-effective. The networking research community has struggled for many years to make networks more programmable. What has worked, and what hasn't, and what lessons have we learned along the way?