- talks - lists.cs.princeton.edu

Y Zhu FPO
by Melissa Lawson 18 Aug '11

18 Aug '11

Yaping Zhu will present her FPO on Thursday August 25 at 1PM in Room 402. The members of her committee are: Jennifer Rexford, advisor; Vivek Pai and Aman Shaikh (AT&T), readers; Michael Freedman and Andrea LaPaugh, nonreaders. A copy of her thesis, "Minimizing Wide-Area Performance Disruptions in Inter-Domain Routing", is on display in Room 310. Everyone is invited to attend her talk.

1 0

S Ihm FPO
by Melissa Lawson 18 Aug '11

18 Aug '11

Sunghwan Ihm will present his FPO on Wednesday August 24 at 2PM in Room 402. The members of his committee are: Vivek Pai, advisor; Jennifer Rexford and Larry Peterson, readers; Michael Freedman and Andrea LaPaugh, nonreaders. A copy of his thesis, "Understanding and Improving Modern Web-Traffic Caching", is on display in Room 310. Everyone is invited to attend his talk.

1 0

J Jiang preFPO
by Melissa M. Lawson 02 Aug '11

02 Aug '11

Joe Wenjie Jiang will present his preFPO on Friday August 5 at 10:30AM in Room 318B (networking lab). Note location. His committee is: Jennifer Rexford, advisor; Andrea LaPaugh and Augustin Chaintreau (Columbia), readers; Mike Freedman and Mung Chiang (ELE), nonreaders. Everyone is invited to attend his talk. His abstract follows below. ---------------------- Title: Wide-Area Traffic Management for Cloud Services Abstract: Cloud services need an effective way to distribute contents across wide area networks. Providing large-scale, geographically-replicated online services presents new opportunities for coordination between server selection (to match subscribers with servers), traffic engineering (to select efficient paths for the traffic) and adaptive content caching (to place contents sources close to users). Traditional designs isolate these problems, which degrades performance, scalability, reliability and responsiveness. We leverage the theory of distributed optimization, collaborative game theory and randomized algorithms to provide solutions that jointly optimizes over different degrees of design freedom. This dissertation proposes a sets of solutions to wide-area and data center traffic management, which consists of the following three thrusts: (i) Cooperative content distribution and traffic engineering in an ISP network. We develop cooperation models with an increasing amount of information exchange between ISPs and content providers. We show that the lack of communication leads to sub-optimal performance, and even when extra information is used inappropriately. We propose a joint design solution to reduce such efficiency loss. (ii) Traffic management in Nano data centers, which represents the trend of decentralizing content delivery by placing contents close to the last mile of end user connectivity and utilizing their upload capability. We solve a joint content caching and request routing problem for minimizing cross-traffic and propose a distributed implementation. (iii) Cost-efficient data center traffic engineering with joint server placement and route selection, by exploiting multipath capability and virtual machine migrations. Our systems can be easily implemented with incremental modifications to today's architecture such as content-selection mechanisms, and infrastructure such as switches and end hosts. The effectiveness of our schemes is demonstrated by our evaluation with real-life traffic traces.

1 0

M Yu FPO
by Nicole E. Wagenblast 29 Jul '11

29 Jul '11

Minlan Yu will present her FPO on Friday August 5 at 3PM in Room 402. The members of her committee are: Jennifer Rexford, advisor; Michael Freedman and Albert Greenberg (MSR), readers; Vivek Pai and Ed Felten, nonreaders. A copy of her thesis, "Scalable Management of Enterprise and Data Center Networks", is on display in Room 310. Everyone is invited to attend her talk.

1 0

Z Khan FPO
by Melissa Lawson 26 Jul '11

26 Jul '11

Zia Khan will present his FPO on Tuesday August 2 at 2PM in Carl Icahn Labs Room 101 (note location!). The members of his committee are: Mona Singh, advisor; Leonid Kruglyak and Benjamin Garcia, readers; Hilary Coller and Saeed Tavazoie, nonreaders. A copy of his thesis, "Efficient Algorithms for Liquid Chromatography Coupled Mass Spectrometry Based Protein Quantification", is on display in Room 310. Everyone is invited to attend his talk.

1 0

11am Mon Jul 11 talk by Robin Kravets (UIUC) on mobility and disconnected networks in CS 402
by Jennifer Rexford 11 Jul '11

11 Jul '11

Speaker: Professor Robin Kravets, UIUC Title: Connecting the Disconnected Date/time: 11am-12pm on Monday July 11 Location: room 402 in the Computer Science building Abstract: Mobility, in all of its forms, results in networks with varying degrees of connectivity. Providing communication services in such networks relies on solutions that treat these disconnections as an expected part of the system, instead of as an exception to be handled out of band. However, handling disconnectivity breaks most architectures for more stable and connected networks. Instead of a standard layered approach to communication services, communication in such intermittently connected networks depends on three key components: contacts, or knowing who we meet, context, knowing what is in our neighborhood, and content, knowing how to find and deliver data. In this talk, I will present some of our current research on support for communication in intermittently connected networks, vehicular networks and mobile social networks, as well as the target applications and deployment testbeds. Bio: Robin Kravets is an Associate Professor at the University of Illinois at Urbana-Champaign and the head of the Mobius group at UIUC, which researches communication issues in all types of networks that are challenged by mobility, including wireless LANs, ad hoc networks, sensor network, delay and disruption tolerant networks, vehicular networks, mobile social networks and personal area networks. Her research focuses on solutions that enable effective power management, connectivity management, data transport, congestion management, location management, routing and security. For more information, visit http://mobius.cs.illinois.edu.

1 1

MohammadHossein Bateni FPO
by Nicole E. Wagenblast 03 Jun '11

03 Jun '11

MohammadHossein Bateni will present his FPO on Thursday June 9 at 1PM in Room 402. The members of his committee are: Moses Charikar, advisor; Sanjeev Arora and MohammadTaghi Hajiaghayi (Maryland), readers; Bernard Chazelle and Rob Schapire, nonreaders. A copy of his thesis, "A Primal-Dual Clustering Technique with Applications in Network Design" is on display in Room 310. Everyone is invited to attend his talk.

1 0

Martin Suchara FPO
by Nicole E. Wagenblast 03 Jun '11

03 Jun '11

Martin Suchara will present his FPO on Wednesday June 15 at 1PM in Room 402. The members of his committee are: Jennifer Rexford, advisor; Mung Chiang (ELE) and Gordon Wilfong (Bell Labs), readers; Michael Freedman and Ed Felten, nonreaders. A copy of his thesis, "Reliable Internet Routing" is on display in Room 310. Everyone is invited to attend his talk.

1 0

Z Wang preFPO
by Melissa Lawson 23 May '11

23 May '11

Zhe Wang will present his preFPO on Friday May 27 at 2PM in Room 402. The members of his committee are: Kai Li, advisor; Moses Charikar and Olga Troyanskaya, readers; Jennifer Rexford and Andrea LaPaugh; nonreaders. Everyone is invited to attend his talk. His abstract follows below. -------------------------- Title: Designing similarity search systems for multimodal data As images, videos and all kinds of non-text sensory data have taken the majority of the storage space, how to efficiently index and search them becomes an issue. The query by example kind of content based search method has become more popular and often can be used to compliment the existing search method. We have designed multiple search system to demonstrate the content based search capability and would discuss our experience about building such systems. First we describe VFerret system where we combine the image visual feature with audio features for personal video search. Second, we explore different aggregation methods to aggregate results from multiple features in a system to detect image spam emails. In the last system, we design a new method to combine image similarity search with product navigation to help user to find desired product in the product category tree. One of the common problem for the similarity search system is the space utilization for the feature vectors used to represent objects. Our group had developed sketch as a compact data structure to alleviate such problem. In order to use the sketch effectively, I have worked out a method to model the sketch and predict the best parameters to use given a sample dataset. Our experience in integrating the similarity search capability into different kinds of system show that the content based search can be an effective method to enhance the existing systems. By creatively exploit multiple modality from the data, we demonstrate much better performance than the traditional systems.

1 0

Monday, 11AM: Geoff Hinton on neural networks
by Nicole E. Wagenblast 16 May '11

16 May '11

hi cs faculty and graduate students, geoff hinton from U of T is speaking tomorrow at 11am in the small auditorium. geoff is one of the founders of modern machine learning and artificial intelligence. (and, he's an incredibly witty and engaging speaker.) see the full announcement below. best dave ---------- Forwarded message ---------- From: David Mimno < mimno(a)cs.princeton.edu > Date: Thu, May 12, 2011 at 2:57 PM Subject: [Ml-stat-talks] Monday, 11AM: Geoff Hinton on neural networks To: ml-stat-talks < ml-stat-talks(a)lists.cs.princeton.edu > For those not familiar with his work: "Geoffrey Hinton talks about neural networks" is roughly equivalent to "J.S. Bach talks about fugues". This is not to be missed. Drag your friends. Geoff's schedule is filling up, but if you would like to meet with him before the talk or in the afternoon, please let me know. Monday, May 16, 11AM (note the change), CS105 ============================= How to force unsupervised neural networks to discover the right representation of images Geoffrey Hinton University of Toronto One appealing way to design an object recognition system is to define objects recursively in terms of their parts and the required spatial relationships between the parts and the whole. These relationships can be represented by the coordinate transformation between an intrinsic frame of reference embedded in the part and an intrinsic frame embedded in the whole. This transformation is unaffected by the viewpoint so this form of knowledge about the shape of an object is viewpoint invariant. A natural way for a neural network to implement this knowledge is by using a matrix of weights to represent each part-whole relationship and a vector of neural activities to represent the pose of each part or whole relative to the viewer. The pose of the whole can then be predicted from the poses of the parts and, if the predictions agree, the whole is present. This leads to neural networks that can recognize objects over a wide range of viewpoints using neural activities that are ``equivariant'' rather than invariant: as the viewpoint varies the neural activities all vary even though the knowledge is viewpoint-invariant. The ``capsules'' that implement the lowest-level parts in the shape hierarchy need to extract explicit pose parameters from pixel intensities and these pose parameters need to have the right form to allow coordinate transformations to be implemented by matrix multiplies. These capsules are quite easy to learn from pairs of transformed images if the neural net has direct, non-visual access to the transformations, as it would if it controlled them. (Joint work with Sida Wang and Alex Krizhevsky)

1 0