- talks - lists.cs.princeton.edu

Xiaojuan Ma preFPO
by Melissa Lawson 06 Jan '10

06 Jan '10

Xiaojuan Ma will present her preFPO on Wednesday January 13 at 1:30PM in room 402. The members of her committee are: Perry Cook, advisor; Marilyn Tremaine (Rutgers) and Christiane Fellbaum, readers; Szymon Rusinkiewicz and Ken Steiglitz, nonreaders. Everyone is invited to attend her talk. Her abstract follows below. ------------------------------------------------- Title: A Multimedia Augmented Online Language Assistant for People with Aphasia and Other Language Barriers Abstract: Communication among language users take place usually in the spoken or written mode, as in face-to-face conversations, reading of books and newspapers, television watching and Internet browsing. However, for people with language disabilities (e.g. aphasia), people with low literacy, and people with poor command of a language, receiving and expressing information in this way is difficult. In particular, because of the inability to comprehend words and or to find words that express intended concepts, people with language barriers may be faced with great challenges performing everyday tasks such as ordering food in a restaurant and visiting a doctor. As an alternative to words, pictures and sounds have been designed, tested, verified and used to evoke concepts in computer interfaces, education, industry, and advertisement. However, icons created by artists, user-uploaded photos, auditory icons and earcons cannot always satisfy the need for communicating everyday concepts for people with language problems. We are building a Multimedia Augmented Online Language Assistant, which employs various multimedia forms, including web images, icons, animations, videos, and environmental sounds with the goal of assisting comprehension of common concepts. The Language Assistant is implemented as a popup dictionary in the form of a web browser extension. Users can select an unfamiliar word on a webpage to view its associated visual/auditory representation in a popup box. The Language Assistant can enhance concept understanding as people browse information on the Internet, and support face-to-face communication when people want to illustrate a term via a picture, a video or a sound when their conversation partner does not understand the spoken word. The Multimedia Augmented Online Language Assistant consists of a backend lexical network associated with large scale of multimedia data, and web interfaces to access the multimedia-concept associations. To develop the backend multimedia enhanced lexical database, we went through a cycle of design, construction, evaluation, and modification for each stimulus-to-concept associations. As an example we describe the creation of SoundNet, a nonspeech audio-to-concept semantic network which is part of the database. Soundnails, five-second environmental auditory representations, were extracted from sound effects libraries, and assigned to a vocabulary of common concepts. Two large scale studies were conducted via Amazon Mechanical Turk to evaluate the efficacy of soundnail representations. The first was a tagging study which collected human interpretations of the sources, locations, and interactions involved in the sound scene. The second study investigated how well people can understand words encoded in soundnails which are embedded in sentences compared to conventionally used icons and animations. Modification guidelines were proposed based on the study results. Through a series of studies, we verified that web images are as effective as stylized icons in conveying nouns, videos outperformed other stimuli in illustrating verbs, and nonspeech audio clips are better in distinguishing concepts like thunder, alarm, and sneezing. The Multimedia Augmented Online Language Assistant was shown to enhance information comprehension, and we are exploring its application in assisting communication involving medical care issues for people with language barriers and low levels of and medical literacy.

1 0

11am Wed Dec 16 talk by Michael Schapira on incentives issues in Internet routing
by Jennifer Rexford 16 Dec '09

16 Dec '09

Michael Schapira (http://www.cs.yale.edu/homes/schapira/) is visiting from Tuesday through Thursday this week. If you would like to meet with him, please drop me a note (jrex(a)cs.princeton.edu). Speaker: Michael Schapira, Yale University and UC Berkeley Title: Interdomain Routing: Stability, Security and Selfishness Date/time: 11am-noon on Wednesday December 16 Location: Room 402 in the Computer Science building Abstract: The Border Gateway Protocol (BGP) establishes routes between the many independently-administered networks that make up the Internet. Over the past two decades there has been exponential growth in the scale and complexity of the Internet. However, BGP has not changed significantly in comparison and, consequently, does not always cope well with modern-day challenges (bounded computational resources, economically driven manipulations, security attacks, and more). Understanding, ``fixing'' and redesigning interdomain routing necessitates taking a principled approach that bridges theory and systems research, and breaks traditional disciplinary barriers. I shall present conceptual frameworks for addressing today's challenges and novel routing schemes that improve on the existing ones. Specifically, I shall present (1) a necessary condition for BGP safety, i.e., guaranteed BGP convergence to a ``stable'' routing outcome; (2) an economic approach to BGP security; and (3) Neighbor-Specific BGP -- a modest extension to BGP that allows network administrators more expressive routing policies while improving global network stability. I shall also discuss the surprising implications of these results for other areas of research: distributed computing, game theory and mechanism design. Finally, I shall outline interesting directions for future work.

1 1

Colloquium Speaker, Dec. 16, 2009- David Naumann
by Nicole E. Wagenblast 14 Dec '09

14 Dec '09

Smart Assertions for Dumb Provers David Naumann, Stevens Institute of Technology 4:30 PM Small Auditorium CS 105 Assertions are widely used for runtime checking of software correctness, and increasingly used for compile time checking of simple properties like array index bounds. Structural integrity of pointer based data structures and design patterns often involves recursively defined properties which are costly to check at runtime and which appear to be unsuited to static checking using fully automated theorem provers. Remarkably, judicious use of ghost state lets programmers express recursive properties in pure first order logic and successfully use fast provers based on SAT solving. In this talk I will introduce the technique in elementary terms, for programmers, who may find it immediately useful in testing. I will also outline research questions about how to use the technique in frame conditions, in refactoring, in foundational proofs, in refinement types, in terms of aspects, and in checking information flow properties.

1 1

Colloquium Speaker, Dec. 9, 2009- Marc Fiuczynski
by Nicole E. Wagenblast 07 Dec '09

07 Dec '09

PlanetLab: From Testbed to Service Deployment Platform Marc Fiuczynski, Princeton Computer Science 4:30 PM Small Auditorium CS 105 Abstract: PlanetLab was originally designed to be a realistic testbed for networking and distributed systems research. It has served that role quite well. More recently, however, it has also become a practical platform for rapidly deploying (and then effectively managing) network services that have value to a wide-range of user communities (and a diverse set of network environments). This talk describes three such deployments: (1) providing web connectivity to remote villages and schools; (2) capturing gold-standard measurements of broadband connectivity in the US; and (3) supporting high-speed content distribution in a Telco access network.

1 0

Colloquium Speaker, Dec. 9, 2009- Marc Fiuczynski
by Michele J. Brown 03 Dec '09

03 Dec '09

PlanetLab: From Testbed to Service Deployment Platform Marc Fiuczynski, Princeton Computer Science 4:30 PM Small Auditorium CS 105 Abstract: PlanetLab was originally designed to be a realistic testbed for networking and distributed systems research. It has served that role quite well. More recently, however, it has also become a practical platform for rapidly deploying (and then effectively managing) network services that have value to a wide-range of user communities (and a diverse set of network environments). This talk describes three such deployments: (1) providing web connectivity to remote villages and schools; (2) capturing gold-standard measurements of broadband connectivity in the US; and (3) supporting high-speed content distribution in a Telco access network.

1 0

S Nikolova preFPO
by Melissa Lawson 24 Nov '09

24 Nov '09

Sonya Nikolova will present her preFPO on Wednesday December 2 at 10AM in Room 402. The members of her committee are: Perry Cook, advisor; Marilyn Tremaine (Rutgers) and Christiane Fellbaum, readers; Adam Finkelstein and Szymon Rusinkiewicz, nonreaders. Everyone is invited to attend her talk. Her abstract follows below. ***** Title and Abstract: Improving Vocabulary Organization in Assistive Communication Tools: A Mixed-Initiative Approach Navigating a vocabulary consisting of thousands of entries in order to select appropriate words for building communication is challenging for individuals with lexical access disorders like those caused by aphasia, a cognitive disorder. Most existing assistive communication vocabularies have a lexical organization scheme based on a simple list of words. Some word collections are organized in hierarchies which often leads to deep and confusing searches; others are simply a list of arbitrary categories which causes excessive scrolling and a sense of disorganization. Ineffective vocabulary organization and navigation hurt the usability and adoption of assistive communication tools and ultimately fail to help users engage in practical communication. We have developed a multi-modal visual vocabulary that enables improved navigation and effective word finding by modeling a speaker's "mental lexicon", where words are stored and organized in ways that allow efficient access and retrieval. Due to impaired links in their mental lexicon, people with aphasia have persistent difficulties accessing and retrieving words that express intended concepts. The Visual Vocabulary for Aphasia (ViVA) attempts to compensate for some of these missing or impaired semantic connections by organizing words in a dynamic semantic network where links between words reflect semantic association measures based on WordNet, human judgments of semantic similarity, and past vocabulary usage. ViVA is both adaptable, able to be customized by the user, and adaptive, able to dynamically change to better suit users' past actions and future needs. This mixed-initiative approach enables the user to feel in control by making changes and anticipating ones that have been initiated by the system while still allowing adaptive methods to help determine where and when changes are required. ViVA's adaptable component allows the user to add and remove vocabulary items, and group them in personalized categories. The adaptive component updates the semantic network according to vocabulary usage, user preferences, and a number of semantic association measures. Thus, ViVA tailors the vocabulary organization according to both user-specific information and general knowledge of human semantic memory. We report results from two studies evaluating how ViVA compares to a vocabulary access system in which words are organized hierarchically into common categories. Since our target user population is people with cognitive impairments, further work involves studying the effect of adaptation on users' cognitive load and investigating the tradeoff between improved navigation and changing vocabulary structure due to adaptation.

1 0

Thursday Lunch Talk: The economics of virtualization [Nov 19]
by jensg＠princeton.edu 18 Nov '09

18 Nov '09

Tomorrow, Paul Laskowski (UC Berkeley) is presenting his work on the economics of network innovation and virtualization. I have attached the announcement below. Please join us for the talk at CITP. If you are interested in meeting with him please let me know your availability. Paul is at CITP from Wednesday - Friday. Jens Grossklags (Postdoctoral Research Associate, CITP) ------------------- Paul Laskowski - The Economics of Virtualization Date: Thursday, November 19, 2009 Time: 12:30 -1:30 pm Location: 306 Sherrerd Hall Food at 12:30 pm. Discussion begins at 12:45 pm. Everyone invited. The difficulty of making changes to the internet architecture has spawned widespread interest in large-scale, “virtualized” testbeds as a place to deploy new services. Despite the excitement, uncertainty surrounds the question of how technologies can bridge the gap from testbed to global availability. It is recognized that no amount of validation will spur today’s ISPs to make architectural changes, so if new services are to reach a widespread audience, the testbed itself must provide that reach. This suggests two questions: First, would today’s network providers (or a new set of providers) ever support a virtualized architecture on a global scale? Second, even if they did, would such a network, spanning a great many domains, support the adoption of new services or upgrades to the infrastructure? In this talk, I will argue that the answers to these questions depend critically on how money flows to network and service providers. A novel economic theory, rooted in the classic model of Cournot competition, allows us to compare market types with regard to service innovation and network upgrade. According to this analysis, there is a danger that a virtualized testbed inherits the market structure prevalent in the internet architecture, causing investment levels to remain poor. On the other hand, alternate market designs can dramatically improve incentives to invest in services, and even in network upgrades, but may encounter resistance from network providers who are likely to see reduced profits. I will discuss how these alternate market types may be implemented.))

1 0

Today's Colloquium Cancelled
by Andrew W. Appel 18 Nov '09

18 Nov '09

Today's colloquium is cancelled. David Naumann cannot speak due to an illness.

1 0

David Naumann--Colloquium Speaker, Wednesday, October 18, 2009
by Michele J. Brown 16 Nov '09

16 Nov '09

Title: Smart Assertions for Dumb Provers David Naumann, Stevens Institute of Technology 4:30PM- Small Auditorium CS105 Host: Andrew Appel Assertions are widely used for runtime checking of software correctness, and increasingly used for compile time checking of simple properties like array index bounds. Structural integrity of pointer based data structures and design patterns often involves recursively defined properties which are costly to check at runtime and which appear to be unsuited to static checking using fully automated theorem provers. Remarkably, judicious use of ghost state lets programmers express recursive properties in pure first order logic and successfully use fast provers based on SAT solving. In this talk I will introduce the technique in elementary terms, for programmers, who may find it immediately useful in testing. I will also outline research questions about how to use the technique in frame conditions, in refactoring, in foundational proofs, in refinement types, in terms of aspects, and in checking information flow properties. Joint work with Anindya Banerjee, Mike Barnett, and Stan Rosenberg

1 0

Colloquium Speaker, Wednesday, November 18, 2009- David Naumann
by Michele J. Brown 13 Nov '09

13 Nov '09

Title: Smart Assertions for Dumb Provers David Naumann, Stevens Institute of Technology 4:30PM- Small Auditorium CS105 Host: Andrew Appel Assertions are widely used for runtime checking of software correctness, and increasingly used for compile time checking of simple properties like array index bounds. For pointer based data structures and design patterns, structural integrity conditions often involve recursively defined properties which are costly to check at runtime and which appear to be unsuited to static checking using fully automated theorem provers. Remarkably, judicious use of ghost state lets programmers express recursive properties in pure first order logic, thereby enabling the use of fast provers based on SAT solving. In this talk I will introduce the technique and pose research questions about how to use the technique in frame conditions, in foundational proofs, in refinement types for higher order programs, in terms of aspects, and in checking information flow properties. Joint work with Anindya Banerjee, Mike Barnett, and Stan Rosenberg

1 0