Events
May 2012
Maryland Cybersecurity Center (MC2) Symposium
The Maryland Cybersecurity Center (MC2) Symposium will be held on May 15-16 in the Riggs Alumni Center. Features cutting edge research on cybersecurity being done at the University of Maryland, tutorials by Maryland faculty, and panels and keynotes from outside experts.
More details at http://umiacs.umd.edu/mc2symposium/
April 2012
March 2012
February 2012
CS Open House Career Fair
The CS Open House Career Fair will be held on February 16, 2012 at 5:30PM. Individuals interested in attending can RSVP here.
November 2011
April 2011
CS Alumni of the Year 2011
CS Alumni of the year Dr. Aya Soffer, "It's time to get social: using social analytics for business advantage".
- Date: April 29, 2011
- Time: 10am
- Location: CSIC 2118
Dr. Soffer is Director of Information Management for Analytics Research in IBM.
CVL Distinguished Lecture: Guillermo Sapiro
Distinguished Seminar Series on Vision in honor of Prof. Azriel Rosenfeld presents Guillermo Sapiro, University of Minnesota, "Is there life after standard sparse models?".
- Date: April 15, 2011
- Time: 11:00 am
- Location: CSIC Building, 4th Floor
With all the recent activities and success in sparse modeling, we can ask ourself a number of question:
- Are sparse models still alive?
- Are they enough?
- Do we understand them better?
CVL Distinguished Lecture: Michael A. Arbib
Distinguished Seminar Series on Vision in honor of Prof. Azriel Rosenfeld presents Michael A. Arbib, University of Southern California, "Template construction grammar and the generation of descriptions of visual scenes".
- Date: April 1, 2011
- Time: 11:00 am
- Location: A.V. Williams Building, Room 2460
To explore how linguistic processes relate to brain mechanisms which integrate action and perception, we have developed a new kind of semantic representation, SemRep, and a system called Template Construction Grammar (TCG) in which constructions compete and cooperate to cover the SemRep to produce a description of a visual scene. The modeling is complemented by studying time-locked data on eye movements and verbal output as subjects describe visual scenes. The results support (1) the presence of a critical threshold for readout of assembled constructions which distinguishes a fragmentary description style from the style that favors grammatically correct and complete utterances; and (2) a dynamic process which integrates seemingly incompatible data supporting a macro-to-micro versus a micro-to-macro strategy under a single unified mechanism.
March 2011
CVL Seminar: Namrata Vaswani
CVL Seminar: Namrata Vaswani, "Recursive Sparse Recovery and Applications in Dynamic Imaging".
- Date: March 30, 2011
- Time: 11:00 am
- Location: AV Williams 1146
In this talk, I will discuss our recent work on Recursive Sparse Recovery (RecSparsRec) and show how it provides novel solutions to two very different problems in dynamic imaging. RecSparsRec refers to recursive approaches to causally recover a time sequence of signals/images from a greatly reduced number of measurements (compared to existing approaches), by utilizing their sparsity.
I will also briefly talk about our ongoing work on the difficult video surveillance problem of tracking multiple small-sized moving objects when the background scene itself is changing. This is posed as a problem of recursive robust principal components analysis (PCA) in the presence of correlated outliers. I will show how RecSparsRec is a key part of a completely new approach to solve this well-studied problem.
CVL Seminar: Daniel Glasner
CVL Seminar: Daniel Glasner, "Contour Based Joint Clustering of Multiple Segmentations".
- Date: March 29, 2011
- Time: 11:00 am
- Location: AV Williams 3258
We present an unsupervised shape based method for joint clustering of multiple image segmentations. Given two or more tightly related images, such as close frames in a video sequence, or images of the same scene taken under different lighting conditions, our method generates a joint segmentation of the images. We introduce a novel contour based representation that allows us to cast the shape-based joint clustering problem as a quadratic semi-assignment problem. Our score function is additive. We use complex valued affinities to assess the quality of matching the edge elements at the exterior bounding contour of clusters, while ignoring the contributions of elements that fall in the interior of the clusters. We further combine this contour-based score with region information and use a linear programming relaxation to solve for the joint clusters. We evaluate our approach on the occlusion boundary data-set of Stein et al.
CVL Distinguished Lecture: Bill Freeman
Distinguished Seminar Series on Vision in honor of Prof. Azriel Rosenfeld presents Bill Freeman, MIT, "Removing blur due to camera shake from images".
- Date: March 11, 2011
- Time: 11:00 am
- Location: CSIC Building, Room 1115
"Blind deconvolution" is a beautiful, ill-posed problem: given an image that has been blurred by some unknown convolution kernel, estimate the image before it was blurred. Lurking within this problem are nice, deep questions: What is an image, and how can you tell when one has been blurred? How should we solve very underdetermined inference problems?
CVL Seminar: Shawn D. Newsam
CVL Seminar: Shawn D. Newsam, University of California, Merced talks about "Proximate Sensing: Inferring What-Is-Where From Georeferenced Photo Collections".
- Date: March 9, 2011
- Time: 3:00 pm
- Location: A.V. Williams Building, Room 2120
In this talk, I will describe an interesting new research direction which I term Proximate Sensing that leverages ground-level georeferenced images to map what-is-where on the surface of the Earth much like the field of Remote Sensing has done for decades using overhead imagery. Enabled by the growing collections of community contributed photo collections, Proximate Sensing represents a rich framework in which to apply and evaluate current image understanding tasks such as scene classification and object recognition as well as motivate the development of novel problems. I will describe how Proximate Sensing can be considered part of the larger phenomena of Volunteered Geographic Information (VGI), a term coined by geographer Michael Goodchild in 2007 to refer to the growing collections of geographically relevant information provided voluntarily by individuals.
CVL Seminar: Ravi Ramamoorthi
Ravi Ramamoorthi, University of California, Berkeley talks about "Sampling and Reconstruction of High-Dimensional Visual Appearance".
- Date: March 9, 2011
- Time: 11:00 am
- Location: A.V. Williams Building, Room 2120
Producing realistic images in computer graphics, or acquiring the visual properties of realistic scenes in computer vision, often involves sampling high-dimensional datasets. One example comes from real-time rendering techniques, that often use a precomputed light transport matrix that can range from 4D - 6D, measuring variation across the image, as well as the space of lighting and viewing directions. Similar problems arise in appearance acquisition for computer vision, where one seeks to acquire data-driven models of shape and spatially-varying reflectance of real materials. Even in offline rendering in graphics, one is sampling a high-dimensional space, involving the pixel area, time for motion blur and depth of field for lens effects. Imaging applications are in many cases analogous to rendering, involving a sampling of the light field, or even just time for videos.
February 2011
January 2011
November 2010
Speaker: Robert Kleinberg
Robert Kleinberg, Cornell, talks about "Converting any algorithm into an incentive-compatible mechanism".
- Date: November 19, 2010
- Time: 1:00 pm
- Location: CSIC 2107
Does the complexity of algorithms increase dramatically when redesigning them to account for the incentives of selfish users? The theory of algorithmic mechanism design is largely founded on the presumption that the answer is affirmative, a presumption that has been rigorously confirmed under various interpretations of the question. This is unfortunate, since it would be very convenient if there existed generic procedures to convert any algorithm into an incentive-compatible mechanism with little or no computational overhead. In this talk, I will present a broad setting in which such generic procedures exist. I will describe a polynomial-time reduction transforming an arbitrary algorithm into a Bayesian incentive-compatible mechanism, given a suitable amount of information about the type distributions of agents. The talk will be self-contained; no prior knowledge of mechanism design theory will be assumed.
Speaker: Vikash Mansinghka
Vikash Mansinghka, CTO of Navia Systems talks about "Natively Probabilistic Computation: Principles, Artifacts and Applications".
- Date: November 15, 2010
- Time: 1:00 pm
- Location: AV Williams 3258
Complex probabilistic models and Bayesian inference are becoming increasingly critical across science and industry, especially in large-scale data analysis. They are also central to our best computational accounts of human cognition, perception and action. However, all these efforts struggle with the infamous curse of dimensionality. Rich probabilistic models can seem hard to write and even harder to solve, as specifying and calculating probabilities often appears to require the manipulation of exponentially (and sometimes infinitely) large tables of numbers.
We argue that these difficulties reflect a basic mismatch between the needs of probabilistic reasoning and the deterministic, functional orientation of our current hardware, programming languages and CS theory. To mitigate these issues, we have been developing a stack of abstractions for natively probabilistic computation, based around stochastic simulators (or samplers) for distributions, rather than evaluators for deterministic functions. Ultimately, our aim is to produce a model of computation and the associated hardware and programming tools that are as suited for uncertain inference and decision-making as our current computers are for precise arithmetic.
In this talk, I will give an overview of the entire stack of abstractions supporting natively probabilistic computation, with technical detail on several hardware and software artifacts we have implemented so far. I will also touch on some new theoretical results regarding the computational complexity of probabilistic programs. Throughout, I will motivate and connect this work to some current applications in biomedical data analysis and computer vision.
Speaker: Mark D. Hill
Professor Mark D. Hill, University of Wisconsin-Madison talks about "Amdahl's Law in the Multicore Era".
- Date: November 15, 2010
- Time: 2:00 pm
- Location: AV Williams 2460
Over the last several decades computer architects have been phenomenally successful turning the transistor bounty provided by Moore's Law into chips with ever increasing single-threaded performance. During many of these successful years, however, many researchers paid scant attention to multiprocessor work. Now as vendors turn to multicore chips, researchers are reacting with more papers on multi-threaded systems. While this is good, we are concerned that further work on single-thread performance will be squashed.
To help understand future high-level trade-offs, we develop a corollary to Amdahl's Law for multicore chips [Hill& Marty, IEEE Computer 2008]. It models fixed chip resources for alternative designs that use symmetric cores, asymmetric cores, or dynamic techniques that allow cores to work together on sequential execution. Our results encourage multicore designers to view performance of the entire chip rather than focus on core efficiencies. Moreover, we observe that obtaining optimal multicore performance requires further research BOTH in extracting more parallelism and making sequential cores faster.
Speaker: Graham Cormode
Graham Cormode, AT&T Labs-Research talks about "Streaming Graph Computations with a Helpful Advisor".
- Date: November 12, 2010
- Time: 1:00 pm
- Location: CSIC 3117
When handling large quantities of data, it is desirable to outsource the computational effort to a third party: this captures current efforts in cloud computing, but also scenarios within trusted computing systems and inter-organizational data sharing. When the third party is not fully trusted, it is necessary to give assurance that the computation has been perfomed correctly. This talk presents some recent results in designing new protocols for verifying computations which are streaming in nature: the verifier (data owner) needs only a single pass over the input, storing a sublinear amount of information, and follows a simple protocol with a prover (serice provider) that takes a small number of rounds. A dishonest prover fools the verifier with only polynomially small probabiliy, while an honest prover's answer is always accepted. Within this model, we show protocols for a number of graph stream problems. Without annotations, streaming algorithms for graph problems generally require significant memory; we show that for many standard problems, including all graph problems that can be expressed with totally unimodular integer programming formulations, only a constant number of hash values are needed for single-pass algorithms given linear-sized proofs. Further, we show how to trade off slightly increased memory requirements for much reduced proof sizes.
Speaker: Anthony Hunter
Anthony Hunter, Professor of Artificial Intelligence, and Head of the Intelligent Systems Group, in the UCL Department of Computer Science talks about "Introduction to Computational Models of Argument".
- Date: November 10, 2010
- Time: 11:00 am
- Location: AV Williams 3258
Computational models of argument are being developed with the aim of reflecting how human argumentation uses conflicting information to construct and analyse arguments. Argumentation involves identifying arguments and counterarguments relevant to an issue (e.g. What are the pros and cons for the safety of mobile phones for children?). Argumentation may also involve weighing, comparing, or evaluating arguments (e.g. What sense can we make of the arguments concerning mobile phones for children?) and it may involve drawing conclusions (e.g. A parent answering the question "Are mobile phones safe for my children?"). In addition, argumentation may involve convincing an audience (e.g. A politician making the case that mobile phones should be banned for children because the risk of radiation damage is too great or a parent trying to convince a young child that they are too young to have a mobile phone).
In this talk, we will look at both graph-based and logic-based formalizations of argumentation. We will consider some of the basic concepts, review proposals from various research groups, and briefly consider some potential application areas.
Distinguished Lecture Series presents Andrew Odlyzko
Distinguished Lecture Series presents Andrew Odlyzko on November 8, 2010.
- Title: Providing Security with Insecure Systems
- Date: Monday November 8, 2010
- Time: 4pm
- Location: CSIC Building, Room 1115
Network security is terrible, and we are constantly threatened with the prospect of imminent doom. Yet such warnings have been common for the last two decades. In spite of that, the situation has not gotten any better. On the other hand, there have not been any great disasters either. To understand this paradox, we need to consider not just the technology, but also the economics, sociology, and psychology of security. Any technology that requires care from millions of people, most very unsophisticated in technical issues, will be limited in its effectiveness by what those people are willing and able to do. This suggests that one can provide adequate security using contrarian approaches that violate traditional security and system engineering precepts (such as encouraging "spaghetti code").
Distinguished Lecture Series presents Andrew Myers
Distinguished Lecture Series presents Andrew Myers on November 1, 2010.
- Title: Programming secure, composable distributed systems with a higher-level language abstraction
- Date: Monday November 1, 2010
- Time: 4pm
- Location: CSIC Building, Room 1115
Computation and persistent storage are rapidly moving into the distributed domain. Yet we are offered very weak security and privacy assurance, especially as complex information systems share information across trust boundaries. Fabric aims to improve this situation by introducing a higher-level abstraction for building complex distributed information systems securely and composably.
Fabric is a decentralized system that allows heterogeneous network nodes to securely share both information and computation resources despite mutual distrust. Its high-level programming language makes distribution and persistence largely transparent to programmers. Its Java-like object model is extended with data resources labeled with confidentiality and integrity policies, enforced by a combination of compile-time and run-time mechanisms. Optimistic, nested transactions ensure consistency across all objects and nodes. A peer-to-peer dissemination layer helps to increase availability and to balance load. Results from applications built using Fabric suggest that Fabric has a clean, concise programming model, offers good performance, and enforces security.
October 2010
Distinguished Lecture Series presents Dan Suciu
Distinguished Lecture Series presents Dan Suciu.
- Title: Querying Probabilistic Data
- Date: Monday October 18, 2010
- Time: 4pm
- Location: CSIC Building, Room 1115
A major challenge in data management is how to manage uncertain data. Many reasons for the uncertainty exists: the data may be extracted automatically from text, it may be derived from the physical world such as RFID data, it may be integrated using fuzzy matches, or may be the result of complex stochastic models. Whatever the reason for the uncertainty, a data management system needs to offer predictable performance to queries over such data.
In this talk I will address a fundamental computational problem in probabilistic databases: given a query, what is the complexity of evaluating it over probabilistic databases? Probabilistic inference is known to be hard in general, but once we fix a query, it becomes a specialized problem. I will show that Unions of Conjunctive Queries (also known as non-recursive datalog rules) admit a dichotomy: every query is either provably #P hard, or can be evaluated in PTIME. For practicaly purposes, the most interesting part of this dichotomy is the PTIME algorithm. It uses in a fundamental way the Mobius' inversion formula on finite lattices (which is the inclusion-exclusion formula plus term cancellation), and, because of that, it can perform probabilistic inference in PTIME on classes of Boolean expressions where other established methods fail, including OBDDs, FBDDs, inference based on bounded tree widths, or d-DNNF's.
2010 Fall Colloquium: Ben Bederson
Fall Colloquium presents Professor Ben Bederson on October 11, 2010.
- Title: The Role of Humans in Computing
- Date: Monday October 11, 2010
- Time: 4pm
- Location: CSIC Building, Room 1115
The fields of natural language processing, computer vision (and artificial intelligence in general) have an important characteristic in common: All seek to automate tasks that humans do naturally so that they can be done more quickly and in greater quantity. Whether it be recognizing the faces of missing children in airports, translating documents between languages, or summarizing the opinions of Iranian blogs, it would be immensely beneficial to society if these problems could be solved quickly, accurately, and cheaply at the push of a button. Unfortunately, neither automated nor manual solutions are simultaneously fast, accurate, and cheap.
Human Computation (HComp) - sometimes called collective intelligence, crowdsourcing, or wisdom of the crowds - is the strategy of combiningthe strengths of computers and humans by assigning small, independent tasks to a large number of human contributors connected by the internet. Then, the next step is to integrate that work with automated systems.
In this talk, I will discuss my work in this area applied to natural language translation. Human translation is expensive and slow, and often unavailable between uncommon language pairs. Machine translation is inexpensive and fast, but quality remains unreliable. In an effort to find a balance between speed and quality, I will describe two iterative translation processes designed to leverage the massive number of online users who have limited or no bilingual skill. I will describe this work in the context of our larger vision of more generally combining human and machine participation in solving complex problems.
2010 Fall Colloquium: Amol Deshpande
Fall Colloquium presents Professor Amol Deshpande.
- Title: Managing and Querying Large Probabilistic Databases
- Date: Monday October 4, 2010
- Time: 4pm
- Location: CSIC Building, Room 1115
Increasing numbers of real-world application domains are generating data that is inherently noisy, incomplete, and probabilistic in nature. Statistical analysis and probabilistic inference, widely used in those domains, often introduce additional layers of uncertainty. Examples include sensor data analysis, data integration and information extraction on the Web, social network analysis, and scientific and biomedical data management. Managing and querying such data requires us to combine the tools and the techniques from a variety of disciplines including databases, first-order logic, and probabilistic reasoning. There has been much work at the intersection of these research areas in recent years. The work on probabilistic databases has made great advances in efficiently executing SQL and inference queries over large-scale uncertain datasets. The research in first-order probabilistic models and lifted inference has resulted in several techniques for efficiently integrating first-order logic and probabilistic reasoning.
In this talk, I will present some of the foundations of large-scale probabilistic data management, and the challenges in scaling the representational power and the reasoning capabilities of probabilistic databases. I will primarily focus on the PrDB probabilistic data management system that we are building here at the University of Maryland. Unlike the other work on this topic, PrDB is designed to represent uncertain data with rich correlation structures, and it uses probabilistic graphical models as the basic representation model. I will discuss how PrDB supports compact specification of uncertainties at different abstraction levels, from "schema-level" uncertainties that apply to entire relations to "tuple-specific" uncertainties that apply to a specific tuple or a specific set of tuples; I will also discuss how this relates to the work on first-order probabilistic models. Query evaluation in PrDB can be formulated as inference in appropriately constructed graphical models, and I will briefly present some of the key novel techniques that we have developed for efficient query evaluation. Time permitting, I will also discuss our work on ranking and top-k query processing over probabilistic datasets. I will conclude with a discussion of some of the open research challenges moving forward.
September 2010
2010 Fall Colloquium: Mihai Pop
Fall Colloquium Series presents Professor Mihai Pop.
- Title: Computational Challenges in Genomics Research
- Date: Monday September 27, 2010
- Time: 4pm
- Location: CSIC Building, Room 1115
During the past few years we have witnessed dramatic advances in DNA sequencing and mapping technologies. These technologies generate data orders of magnitude faster, and at just a fraction of the costs previously possible. As a result, DNA sequencing is rapidly becoming a critical tool in many areas of biology research. At the same time, the wealth of data being generated is rapidly challenging the capacity of the computational infrastructure available to researchers. Also, as sequencing is being applied in new contexts, the resulting data cannot be effectively analyzed by existing computational tools.
In my talk I will describe recent research from my lab that addresses emerging computational challenges in the analysis of genomic data.
During the first half of the talk I will discuss several results we obtained in the broad area of string algorithms. These results include theoretical analyses of the computational complexity of genome assembly, the development of algorithms for inexact matching capable of rapidly processing large numbers of short DNA sequences, and the use of cloud computing infrastructure to "commodify" the computational analysis of large genomic data-sets.
In the second part of my talk I will focus on problems arising from metagenomics research - the genomic analysis of communities of microbes. I will present new algorithms we developed for metagenomic assembly, as well as initial results on using time-series data to infer and model the dynamic interactions between microbes inhabiting an environment.
2010 Fall Colloquium: Neil Spring
Fall Colloquium Series presents Professor Neil Spring.
- Title: Underspecified Network Protocols Complicate Research
- Date: Monday September 20, 2010
- Time: 4pm
- Location: CSIC Building, Room 1115
Network protocols are well-specified where it matters: features necessary for interoperation. For other behaviors, such as heuristics or corner cases, standards are silent or ignored. I describe two research efforts that, at the start, seemed to be straightforward but became interesting because of varied implementation choices in routers and in wireless interfaces.
Maranello is our modification of the ARQ (acknowledgment) component of 802.11 to support partial retransmissions. The potential increase in throughput provided by Maranello depends on the aggressiveness of the underlying implementation: how it performs backoff and how it selects fallback rates. I will describe our exploration of these heuristics and show how well Maranello works when interfaces are well-behaved.
Discarte is our Internet mapping technique that uses the record route IP option to complement the traditional traceroute method. Having two methods to measure the same features exposes weaknesses of each, which in turn makes it possible to use clever techniques to correct errors. I will describe our application of disjunctive logic programming to scalably analyze measurements to derive maps with few errors.
July 2010
CS 4 HS @ UMD
CS4HS workshop sponsored by UMD and Google, July 12-14, in the CSIC Bldg. The CS4HS workshop, led by Dr. Bill Pugh, will provide High School teachers with learning modules, hands-on experience, and current research resources on how to teach computational thinking to High School students.
Website: http://www.cs.umd.edu/highschool
April 2010
Raymond Reiter Lecture
Professor Emeritus Jack Minker has been invited to present a lecture in honor of Raymond Reiter at the 2010 International Conference on Nonmonotonic Reasoning in Toronto, Canada, in mid-May 2010. The lecture is historical in nature, but is oriented towards those familiar with nonmonotonic reasoning, a subfield of artificial intelligence. Professor Minker will present his lecture at the University of Maryland. Preceding his lecture, he will briefly discuss nonmonotonic reasoning so those unfamiliar with the field will be able to follow the lecture. An abstract of the talk he will present at 2010 NMR follows.
- Title: Raymond Reiter and Nonmonotonic Reasoning
- Date: April 28, 2010
- Time: 1:30-3pm
- Location: AV Williams Bldg., ISR Conf.Room 1146
Ray Reiter, was one of the leading researchers in the field of artificial intelligence at the time of his untimely death, September 16, 2002, at the age of 63. He received many awards for his seminal contributions: Fellow of the ACM, the AAAI, the Royal Society of Canada and 1993 IJCAI Outstanding Contribution Award. Ray was a principal founder of nonmonotonic reasoning, generally considered to have started in 1980. Thirty years later, in 2010, in the city of his birth, Toronto, we appropriately celebrate Ray at the International Conference on Nonmonotonic Reasoning. In this talk, I discuss Ray, the person, his major contributions to artificial intelligence in data and knowledge bases, default reasoning, cognitive robotics and other topics related to nonmonotonic reasoning. I also discuss work spawned by his contributions. It is a tribute to Ray that his seminal research is still as vibrant and cited today as it was 30 years ago.
Honorary Seminar: Ben Shneiderman for his election to the NAE
University of Maryland Department of Computer Science is proud to present a seminar on the occasion of Ben Shneiderman's Election to the National Academy of Engineering. Department Chair Larry Davis will make introductory remarks.
- Title: Information Visualization for Knowledge Discovery
- Date: Tuesday April 27, 2010
- Time: 2pm
- Location: AV Williams Building, Room 2460
Interactive information visualization tools provide researchers with remarkable capabilities to support discovery. These telescopes for high-dimensional data combine powerful statistical methods with user-controlled interfaces. Users can begin with an overview, zoom in on areas of interest, filter out unwanted items, and then click for details-on-demand. With careful design and efficient algorithms, the dynamic queries approach to data exploration can provide 100msec updates even for million-record databases.
Maryland Day 2010
Maryland Day 2010 takes place April 24, 10-4pm. Computer Science has two demos located in the tent in front of the Math Building.
Distinguished Lecture Series presents Eric Horvitz
Distinguished Lecture Series presents Eric Horvitz, "Machine Intelligence and the Open World".
- Date: April 19, 2010
- Time: 4:00 pm
- Location: CSIC Building, Room 1115
Systems that sense, learn, and reason from streams of data promise to provide extraordinary value to people and society. Harnessing computational principles to build systems that operate in the open world can also teach us about the sufficiency of existing models, and frame new directions for research. I will discuss efforts on learning and inference in the open world, highlighting key ideas in the context of projects in transportation, energy, and healthcare. Finally, I will discuss opportunities for building systems with new kinds of open-world competencies by weaving together components that leverage advances from several research subdisciplines.
2010 Spring Colloquium: V.S. Subrahmanian
Professor V.S. Subrahmanian will be giving a talk, "Selected Research in Probabilistic Logics".
- Date: April 5, 2010
- Time: 4:00 pm
- Location: CSIC Building, Room 1115
Probabilistic logics (due originally to Boole) are used to reason about situations where dependencies between events are either non-existent or not understood. In this talk, I will describe the basic ideas underlying probabilistic logics and describe some recent results on solving the "most probable world" problem as well as results on intelligibly representing the uncertainty in answers to queries to probabilistic logic programs.
In addition, I will also briefly describe some ongoing work related to social networks and geospatial reasoning.
March 2010
CMPS Open House
The College of Computer, Mathematical and Physical Sciences will hold an Open House for admitted students March 25, 2010.
The faculty, students and staff of the College of Computer, Mathematical and Physical Sciences are eager to answer your questions about our undergraduate programs and the many opportunities we provide. Let us welcome you and share our excitement for scientific discovery with you.
Opening remarks will begin at 5:30 PM and will be followed by informal discussions with faculty and a reception. This is a parent-friendly event!
2010 Spring Colloquium: Dianne O'Leary
Professor Dianne O'Leary will be giving a talk, "Confidence and Misplaced Confidence in Image Reconstruction"
- Date: March 8, 2010
- Time: 4:00 pm
- Location: CSIC 1115
Forming the image from a CAT scan and taking the blur out of vacation pictures are problems that are ill-posed. By definition, small changes in the data to an ill-posed problem make arbitrarily large changes in the solution. How can we hope to solve such problems using noisy data and inexact computer arithmetic?
In this talk we discuss the use of side conditions and bias constraints to improve the quality of solutions. We discuss their impact on solution algorithms and show their effect on our confidence in the results.
CRA Outstanding Undergraduate Researcher Awards
Richard Matthew McCutchen, a Junior majoring in Computer Science and Mathematics is one of the CRA Outstanding Undergraduate Researcher Award winners for 2010. Two additional CS students received awards, John Silberholz, who was a finalist and Allie Hoch received an honorable mention in the contest.
- Date: Monday, March 1, 2010
- Time: 4:00 pm
- Location: CSIC 1115
Richard Matthew McCutchen, "Streaming algorithms for clustering of point sets".
John Silberholz, "Integrating Post-Newtonian Equations on Graphics Processing Units".
Allie Hoch, "Finding Meaningful Patterns in Gene Annotation Graphs".
February 2010
Guest Lecture: Hamid Pirahesh, IBM Fellow
Hamid Pirahesh, IBM Fellow talks about "Impact of Cloud Computing on Emerging Analytic Software Systems and Solutions". Abstract & Bio
- Date: February 23, 2010
- Time: 4:00 pm
- Location: AV Williams 1146
November 2009
Distinguished Lecture Series presents Divesh Srivastava
Distinguished Lecture Series presents Divesh Srivastava, "Dependence and Truth".
- Date: December 7, 2009
- Time: 4:00 pm
- Location: CSIC Building, Room 1115
The Web has enabled the availability of a huge amount of useful information, but has also eased the ability to spread false information and rumors across multiple sources, making it hard to distinguish between what is true and what is not. Since it is important to permit the expression of dissenting and conflicting opinions, it would be a fallacy to try to ensure that the Web provides only consistent information. However, to help in separating the wheat from the chaff, it is essential to be able to determine dependence between sources. Given the huge number of data sources and the vast volume of conflicting data available on the Web, doing so in a scalable manner is very challenging. Read more...
October 2009
Joint Distinguished Lecturer Geri Gay
Geri Gay, Kenneth J. Bissett Professor and Chair of Communication at Cornell University presents "Mobile Health - Social Influence and Emotional Support in Context".
- Date: October 22, 2009
- Time: 4:00 pm
- Location: CSIC Building, Room 1115
This talk will focus on how mobile phones can be used to employ various forms of motivation, both social and individual, to encourage healthy behavior. Motivators such as competition, cooperation, control, and recognition that have been leveraged to bring about behavior change in many circumstances. Social influence has been shown to play an important role in persuasion and the motivation of behavior change; countless studies, both involving technology and not, have shown that individuals grouped with peers have better results in quitting drinking, quitting smoking, losing weight, exercising, and even surviving cancer. This research explores social influence and also emotional support through two context-aware mobile applications developed in our HCI lab at Cornell and designed to encourage healthy behavior.
This lecture is jointly sponsored by the College of Information Studies and the Department of Computer Science. Read more...
View slides and presentation: http://www.cs.umd.edu/talks/dls/2009/gerigay/
September 2009
Distinguished Lecture Series presents Jon Kleinberg
Jon Kleinberg, Tisch University Professor from Cornell University presents "Meme-tracking, Scheduling, and the Dynamics of the News Cycle".
- Date: October 15, 2009
- Time: 4:00 pm
- Location: CSIC Building, Room 1115
The flow of news through on-line networks has created a complex landscape of media sources and led to rich datasets that provide glimpses into how news is produced, shaped, and consumed. We begin by discussing methods for studying how news stories spread through such a system, using an approach that tracks short pieces of as they travel and mutate across news sources. This type of analysis can be effective at capturing temporal patterns in the news over a daily time-scale --- in particular, the succession of story lines that evolve, compete for attention, and collectively produce an effect that commentators refer to as the `news cycle.' We then show how a detailed analysis of temporal dynamics can suggest novel optimization problems in the scheduling of news stories and other on-line media. Specifically, given a supply of featured content and data on user attention over time, we consider how to sequence the content in a way that maximizes the size of the audience.
This is joint work with Lars Backstrom and Jure Leskovec. Read more...
View slides and presentation: http://www.cs.umd.edu/talks/dls/2009/kleinberg/
May 2009
Workshop on Many-Core Computing
The University of Maryland will host a workshop on parallel computing systems on May 29, 2009, titled "Theory and Many-Cores (T&MC): What Does Theory Have to Say About Many-Core Computing?." The workshop is being organized by Professor Uzi Vishkin (ECE/UMIACS).
The main objective of the workshop will be to explore opportunities for theoretical computer science research and education in the emerging era of many-core computing, and develop understanding of the role that theory should play in it.
Workshop website: http://www.umiacs.umd.edu/conferences/tmc2009/
HCIL 26th Annual Symposium
Mark your calendars! May 28th - 29th, 2009.
HCIL's 26th Annual Symposium will highlight the cutting-edge research being conducted in the Human-Computer Interaction Lab at the University of Maryland. The Symposium will take place Thursday, May 28th, followed by a day of tutorials and workshops on Friday, May 29th. Registration begins in March!
April 2009
Researchers to demonstrate V911 system on Maryland Day
COLLEGE PARK, Md. - University of Maryland researchers have created a new emergency alert technology for cell phones and PDA's called V911, which they say could help improve safety on campuses across the U.S.
Created by scientists and students from the UM Institute for Advanced Computer Studies (UMIACS), the V911 technology is one tool in MyeVyu, downloadable software package for cell phones and PDAs that provides a host of new networking and information access capabilities, including a direct link between the user and campus police dispatch. With the touch of a single button, a user in distress can alert campus police of their identity and location, and stream live, GPS-enhanced video and audio of the incident situation. This new technology will be demonstrated for the public at Maryland Day on Saturday, April 25, 2009.
CS Alum of the Year
Vipin Kumar, Department of Computer Science Alum of the year will be giving a talk entitled "Discovery of Patterns in Global Earth Science Data using Data Mining".
- Date: Friday, April 24, 2009
- Time: 1:30pm - 2:30pm
- Location: CSIC Building, Room 1122
Vipin Kumar is a William Norris Professor and Head of Computer Science & Engineering at the University of Minnesota.
November 2008
Distinguished Lecture Series presents Prabhakar Raghavan
Prabhakar Raghavan, head of Yahoo! Research, presents "New Sciences for a new Web".
- Date: December 1, 2008
- Time: 4:00 pm
- Location: CSIC 1115
The web experience has changed from a human interacting with a browser, to the emergence of a plethora of social media experiences. One consequence is that we need research advances that straddle the boundaries between computational and social sciences, the latter including microeconomics, cognitive psychology and sociology. It also raises difficult questions on the use of data - ranging from the algorithmic to the societal.
This lecture will attempt to chart this interdisciplinary research agenda, arguing that the most influential research will require heavy interaction between these "hard" and "soft" sciences. Read more...
Distinguished Lecture Series presents Thomas Malone
Thomas Malone from MIT's Center for Collective Intelligence presents "Classifying and Designing Collective Intelligence".
- Date: November 3, 2008
- Time: 4:00 pm
- Location: CSIC 1115
Some of the most interesting recent Internet applications (like Wikipedia, Digg, and Google) are examples of what is variously called "Web 2.0" or "collective intelligence." Designing such systems well is not just a matter of software design; it's also a matter of organizational design. This talk will describe some examples of these new types of systems and present early results of a project whose goals are to systematically classify them and identify the conditions in which they are useful. One goal of this work is to develop principles for designing both the software and the other organizational elements needed for successful collective intelligence. Read more...
October 2008
Colloquium Series presents David Jacobs
David Jacobs continues the 2008 Colloquium Series with his research "Matching Images with Deformations".
- Date: October 20, 2008
- Time: 4:00 pm
- Location: CSIC 1115
When we match images that come from the same object, we must often allow for 2D, non-linear deformations. These can model changes in shape that can occur when an object deforms or an articulated object moves its parts, differences in shape between different instances of the same type of object, or variations in apparent shape due to changes in viewpoint. This talk will provide an overview of several approaches to matching that stress finding problem formulations that yield computationally efficient algorithms. Read more...
20th Maryland Theoretical Computer Science Day
The University of Maryland Institute for Advanced Computer Studies (UMIACS) is organizing a Theoretical Computer Science Day to be held on Tuesday, October 14, 2008 at the University of Maryland, College Park. All talks will be held in room AVW 2460 A.V.Williams Building.
September 2008
Colloquium Series presents Bobby Bhattacharjee
Bobby Bhattacharjee continues the 2008 Colloquium Series with his research "Systems without Cooperation".
- Date: October 13, 2008
- Time: 4:00 pm
- Location: CSIC 1115
Early systems on the Internet were exercises in feasibility: institutions and users cooperated to build, deploy, and demonstrate applications and uses of packet switching. Modern Internet protocols must assume a user base that is not necessarily cooperative or altruistic. Fundamental Internet protocols were characterized by cooperative primitives such as voluntary backoff, exchange of accurate local state, and trusted authorities. It is relatively easy to exploit these primitives in non-cooperative settings; instead, researchers are developing new protocols that are characterized by the application of game theory and mechanism design.
I will discuss challenges that arise from mismatched assumptions and describe how to build practical systems that provide strong guarantees, outline fundamental problems with current approaches with regard to information leakage and abuse, and outline our current work addressing these problems. Read more...
Colloquium Series presents Jeff Foster
Jeff Foster starts off the 2008 Colloquium Series with his research "Static Analysis to Improve Software Reliability and Security".
- Date: September 15, 2008
- Time: 4:00 pm
- Location: CSIC 1115
The goal of my research is to discover fundamental new ways to improve software reliability and security. My focus is on developing language-based tools and techniques that find and prevent software defects. In this talk, I will give an overview of this research area and discuss three recent efforts by my research group in more detail: Read more...
May 2008
HCIL 25th Anniversary and Annual Symposium
Come celebrate the Human-Computer Interaction Lab's 25th anniversary by joining us for a very special Symposium on May 29, 2008. Not only will you hear talks about cutting-edge research being conducted at the HCIL, but this year we will begin the Symposium with a very special keynote panel, "25 Years of HCI, 25 Years of the HCIL." Esteemed colleagues from outside of the HCIL will offer their reflections. In addition, this year we will continue the tradition of demos and posters following the talks, but these will happen as a part of lab tours where you will be able to see our new facilities. The following day, May 30, 2008 there will be a wide variety of tutorials and workshops that can't be missed. Be sure to sign up early, since space is limited.
Department Awards Ceremony, 2008
The Department hosted its 2008 annual Awards Ceremony on April 18, recognizing the achievements of faculty, staff and students over the past year. Congratulations to all recipients on their accomplishments and hard work.
Botball Robotics Contest 2008
The annual Botball Robotics contest for the DC region will be held in Ritchie Coliseum on May 3. Botball engages students of all backgrounds in science, technology, engineering, and math (STEM) as they work together as a team to achieve a long-term goal. Botball gives students the opportunity to be on the creative side of technology as they design, build, program, and document a pair of autonomous mobile robots to play in a competition.
April 2008
March 2008
Graduate Visit Day 2008
The Computer Science Department's annual Graduate Visit Day was held on March 28, 2008. Visit Day began with a welcome/information session, continued with lunch followed by a poster session, included individual meetings with faculty and a campus tour, and concluded with dinner out in small groups, hosted by faculty and current students. The entire department pitched in to help make Visit Day a great success.
Meet the Family
Newly-admitted Terps from the class of 2012 were welcomed at a "Meet the Family!" reception held on the evening March 13th. These students and their families are deep in the springtime ritual of comparing schools to which they have been accepted, and were greeted by faculty, students and alum all eager to have them confirm with College Park. The event included presentations by faculty on the wide variety of research opportunities open to undergrads in our program, plus 'student only' panels conducted by present CS majors.
HCIL UX (User Experience) Laboratory
Increasingly, rigorous evaluations of software with human subjects are expected in computer science research. Computer science researchers are also studying how humans perform tasks so that they can evaluate how best to augment these tasks with technology.
To support these research needs, we have built a usability lab as a service to the Computer Science department and UMIACS in room 3452 of A.V. Williams. This lab is equipped with many things needed to conduct a user study, including video recording, screen capture and analysis software.
The HCIL UX lab was dedicated March 12th, 2008. More pictures and information will be posted.
High School Programming Contest
The UMD High School Programming Contest brings talented students from high schools throughout the DC metropolitan area to the campus to participate in a three hour competition. Students competing in teams of four demonstrate their programming skills and problem solving abilities by attempting to solve eight programming problems in Java, using the Eclipse programming environment, on Apple MacBooks.
The 2008 Contest took place Saturday, March 8, at the CSIC Building at the University of Maryland, College Park. 39 teams from Maryland, Virginia, and the District of Columbia competed in the contest.
Graduate Visit Day 2008
The Department of Computer Science will hold a visit day for prospective graduate students on Friday, March 28th 2008. The visit day is organized to provide prospective graduate students with a chance to get an up-close view of the department. It is an opportunity for the students to meet faculty and students of the department and learn about the graduate program.


