Advances in miniature electronics, distributed algorithms and manufacturing technology have enabled a new generation of smart composites that tightly integrate sensing, actuation, computation and communication. Such “robotic materials” are inspired by multi-functional natural structures such as the skin of the cuttlefish that can change its color and patterning, bird wings that can change their shape, or the human skin that provides tactile sensing at high dynamic range. Correll will describe a series of recent results that best illustrate the benefits of material integrated computation: high-bandwidth sensing for texture recognition and localization in artificial skins, distributed optimization for controlling shape change, distributed classification for recognizing gestures drawn onto a modular facade and feedback control of soft robotic actuators. He will then describe current challenges in robotic grasping and manipulation, and demonstrate how robotic materials can provide critical sensing and control during a series of manipulation tasks.
Nikolaus Correll is an assistant professor in Computer Science at the University of Colorado Boulder with courtesy appointments in Aerospace Engineering Sciences; Electrical, Computer and Energy Engineering; and Materials Engineering. Nikolaus obtained a degree in electrical engineering from ETH Zurich in 2003 with visits at Lund Tekniska Hogskola, Sweden, and Caltech, and earned a PhD in computer science from EPFL in Lausanne, Switzerland in 2007 working with Alcherio Martinoli. He did a post-doc at MIT CSAIL from 2007-2009. Nikolaus is the recipient of a 2012 NSF CAREER award, a 2012 NASA Early Career Faculty Fellowship and the 2016 Provost Faculty Achievement award at CU Boulder.# vimeo.com/186315667 Uploaded 29 Plays 0 Likes 0 Comments
Machine learning is ubiquitous, but most users treat it as a black box: a handy tool that suggests purchases, flags spam or autocompletes text. Boyd-Graber presents qualities that ubiquitous machine learning should have to allow for a future filled with fruitful, natural interactions with humans: interpretability, interactivity and an understanding of human qualities. After introducing these properties, he presents machine learning applications that begin to fulfill these desirable properties. He begins with a traditional information processing task--making sense and categorizing large document collections--and shows that machine learning methods can provide interpretable, efficient techniques to categorize large document collections with a human in the loop. From there, Boyd-Graber turns to techniques to help computers understand and detect when texts reveal their writer's ideology or duplicity. Finally, he ends with a setting combining all of these properties: language-based games and simultaneous machine translation.
Jordan Boyd-Graber is an assistant professor in the University of Colorado Boulder's Computer Science Department, formerly serving as an assistant professor at the University of Maryland. Before joining Maryland in 2010, he did his PhD thesis on "Linguistic Extensions of Topic Models" with David Blei at Princeton. Boyd-Graber's research focus is in applying machine learning and Bayesian probabilistic models to problems that help us better understand social interaction or the human cognitive process. He and his students have won "best of" awards at NIPS (2009, 2015), NAACL (2016), and CoNLL (2015), and Boyd-Graber won the British Computing Society's 2015 Karen Spärk Jones Award. His research has been funded by DARPA, IARPA, NSF, NCSES, ARL, OMO, NIH and Lockheed Martin and has been featured by CNN, Huffington Post, New York Magazine, Talking Machines and the Wall Street Journal.# vimeo.com/186315649 Uploaded 40 Plays 0 Likes 0 Comments
While we have seen an explosion in the reasoning capabilities of automated program analyzers in the last decade, it is still quite rare to see advanced program analyzers assisting software developers to find and fix bugs. This situation is not due to a lack of effort in deployment but rather a relative lack of techniques that enable the human user to take advantage of the wealth of information derivable by an analyzer.
In this talk, Chang presents some of their approaches that move beyond the singular focus on automated reasoning engines to develop effective techniques that encompass the entire process of applying tools during software development through novel ways to mix specification of properties, alarm triage, and mining source code repositories with program analysis. His team's work is motivated by real-world application domains, such as developing Android applications, which exhibits heavy use of the heap and event-driven concurrency that is challenging for tools and humans alike.
Bor-Yuh Evan Chang (http://www.cs.colorado.edu/~bec) is an Assistant Professor of Computer Science at the University of Colorado Boulder. He is interested in tools and techniques for building, understanding, and ensuring reliable computational systems. His techniques target using novel ways of interacting with the programmer to design more precise and practical program analyses. He joined the CU Boulder faculty in 2009 after receiving his PhD in 2008 from the University of California Berkeley and his BS in 2002 from Carnegie Mellon University. He is a recipient of an NSF CAREER Award (2010).# vimeo.com/186309950 Uploaded 19 Plays 0 Likes 0 Comments
This talk will review cybersecurity challenges and opportunities from the perspective of the DHS S&T Cyber Security Division. DHS S&T CSD funds R&D at universities, businesses and labs. The approach is part of a broader Federal Cybersecurity Research and Development Strategic Plan, which included input from two “National Conversations” held in Colorado and was released by the White House in February.
Cybersecurity is no longer confined to desktop computers and data centers. Nearly every industry sector from energy and finance to food and agriculture have benefited from new cyber components. Along with these advances have come cyber-related vulnerabilities. Researchers have demonstrated techniques for remotely hacking vehicles on the Interstate and have shown how remote commands can cause medical devices to deliver fatal doses. Actual cyber incidents have disrupted systems from 911 centers in the U.S. to the Ukrainian power grid. The convergence of the cyber and physical worlds brings new research opportunities for faculty in a variety of fields, requires new skills from our students, and also brings risks if we fail to adapt. Today’s automotive engineer will continue to deal with traditional mechanical engineering properties but must also understand the embedded control units in the vehicle. Further, future automotive technical managers must also consider the implications connectivity brings for privacy and how these systems may change underlying business assumptions.
Dan Massey is a program manager in the Cyber Security Division (CSD) for the Homeland Security Advanced Research Projects Agency at DHS S&T. His portfolio includes security for cyber physical systems, distributed denial of service defense, routing security, and open source security. Prior to his position in CSD, Massey worked as a tenured associate professor at Colorado State University and was previously a project leader and research assistant professor at the University of Southern California Information Science Institute. He earned his bachelor’s degree in mathematics and computer science and his master’s degree in applied mathematics all from the University of California, San Diego and his doctorate in computer science from the University of California, Los Angeles.# vimeo.com/186309949 Uploaded 22 Plays 0 Likes 0 Comments
How is social media changing the way social networks solve complex tasks? And what are the physical and cognitive limits of these networks? I will describe a number of real-world experiments on mobilizing very large numbers of people via social media to achieve complex tasks, from finding balloons tethered at random locations all over a continent in under 9 hours, to locating individuals in remote cities in under 12 hours using only their mug shots, to reconstructing shredded documents with the help of hundreds of volunteers. I will also discuss how things can go wrong, from deliberate misinformation and sabotage, to cognitive biases that can impede the collective intelligence and learning in networks.
Iyad Rahwan is an associate professor of Media Arts and Sciences at the MIT Media Lab, where he heads the Scalable Cooperation group. His research explores mechanisms and technologies that promote collective intelligence and large-scale cooperation. Rahwan's work appeared in major academic journals, including Science and PNAS. He holds a PhD from the University of Melbourne, Australia# vimeo.com/168348143 Uploaded 107 Plays 1 Like 0 Comments
Computer Science Colloquia
The University of Colorado Boulder Department of Computer Science holds colloquia throughout the fall and spring semesters. These colloquia, open to the public, are typically held on Thursday afternoons, but sometimes occur at other times as well. Recordings…
The University of Colorado Boulder Department of Computer Science holds colloquia throughout the fall and spring semesters. These colloquia, open to the public, are typically held on Thursday afternoons, but sometimes occur at other times as well. Recordings are typically posted the following week.
If you would like to receive email notification of upcoming colloquia, please subscribe to our Colloquia Mailing List (colorado.edu/cs/colloquia/colloquia-mailing-list).
Browse This Channel
More stuff from “Computer Science Colloquia”
Heads up: the shoutbox will be retiring soon. It’s tired of working, and can’t wait to relax. You can still send a message to the channel owner, though!