CIS Seminars & Events

Spring 2015 Colloquium Series

Unless otherwise noted, our lectures are held weekly on Tuesday and/or Thursday from 3:00 p.m. to 4:15 p.m. in Wu and Chen Auditorium, Levine Hall.

February 3rd

Xuehai Qian
EECS at Berkeley, Computer Science Division
University of California at Berkeley
"Taming Relaxed Memory Consistency and Non-determinism in Parallel System
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall

Read the Abstract and Bio

 

Abstract

With computer architectures moving towards an era dominated by
many-core machines and the ever-increasing demands of big data
processing, parallel programming has become the norm. Unfortunately,
most current programmers find parallelism challenging. It is urgent to
provide architectural and software supports to make parallel
applications easy to build, reason and debug. Among others, relaxed
memory consistency and non-determinism in particular make
shared-memory based parallel programming difficult.

In this talk, I will give an overview of our strategy to tame the two
factors. Specifically, I will present OmniOrder, a cache coherence
protocol for atomic blocks (transactions). It eliminates the effects
of relaxed consistency by supporting strict sequential consistency
with high performance. OmniOrder supports conflict serialization based
on the conventional directory-based protocol. I will also present
Pacifier, a deterministic record and replay scheme for relaxed
consistency models beyond Total-Store-Order (TSO). It helps to track
and understand the behaviors of relaxed consistency. Finally, I will
briefly discuss a software partial record and replay approach for
one-sided communication used in Partitioned Global Address Space
(PGAS).

Bio

Xuehai Qian is a postdoctoral researcher at University of California
Berkeley. He got the Ph.D from the Department of Computer Science at
the University of Illinois, Urbana-Champaign in 2013. His research
interests include parallel computer architecture, architectural support for
programming productivity and debugging support for
large-scale HPC applications. He received an MS in Computer Science
from the Institute of Computing Technology (ICT), Chinese Academy
of Sciences (CAS), and a BS in Computer Engineering from Beihang
University, Beijing. 

 

March 3rd, 2015 Warren Center Lecutre

Dr. Matthew Jackson
Professor of Economics at Stanford University
Title Gossip: Identifying Central Individuals in Networks and Diffusion Processes
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall

 

Read the Abstract and Bio
Abstract:

How can we identify the most influential nodes in a network for initiating diffusion? Are people able to easily identify those people in their communities who are best at spreading information, and if so How? Using theory and recent data, we examine these questions and see how the structure of social networks affects information transmission ranging from gossip to the diffusion of new products.   In particular, a model of diffusion is used to define centrality and shown to nest other measures of centrality as extreme special cases.  Then it will be shown that by tracking gossip within a network, nodes can easily learn to rank the centrality of other nodes without knowing anything about the network itself.  The theoretical predictions are consistent with data from rural India.  

 

Bio:

Matt Jackson is the William D. Eberle Professor of Economics at Stanford University where he returned as a faculty member in 2006 after receiving his PhD  in 1988. Professor Jackson is a fellow of the American Academy of Arts and Sciences, a Guggenheim fellow, a fellow of the Econometric Society, and an Economic Theory fellow of the Society for Advancement of Economic Theory. He was a Co-Editor at Econometrica and Games and Economic Behavior, and formerly served as an associate editor at several other journals including Journal of Economic Theory, Theoretical Economics and Social Choice and Welfare. 


Professor Jackson's research interests span a wide variety of areas within economics and economic theory, including political economy and voting, mechanism design and implementation, auction theory and market design, and learning and cognition. Recently, he is most known for his influential research on social networks and their relationship to economic outcomes. 

He has also recently written an influential textbook, "Social and Economic Networks", teaches a popular MOOC on Networks, and co-teaches a popular MOOC on game theory.
Reception immediately following

 

March 19th, 2015 Singh Program on Networked and Social Systems Engineering

Sue Gardner
Special Advisor to Wikipedia
"Yes! The Future of the Internet"
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall
Read the Abstract and Bio

Abstract: The internet used to be a Wild West, in which corporations, well-funded start-ups, and amateurs competed for user attention, with the amateurs often winning. The net gave ordinary people access to the means of production—a billion blogs, self-hosted sites, and proto-social networking sites were born! That was great, Sue Gardner contends, because the internet should be like a city, with shoe stores and banks and restaurants, but also with parks and libraries and schools. Today, though, the internet has matured. It is increasingly corporatized and commercialized, and ordinary people's open participation has declined. We participate in narrower, simpler ways than we used to, such as "liking" something on Facebook, or republishing other people's posts on Tumblr. In this state-of-the-union talk, delivered at a crucial moment, Sue Gardner looks to the future of the net, and examines the implications for democracy, journalism, education, free speech, creativity, openness—and much more. If we want an Internet that allows for healthy public spaces, she says, we are going to need to course-correct, including figuring out ways to pay for what we want. Otherwise, we risk ordinary people being flipped from creators into consumers, exactly like what happened with television 50 years ago. Bringing extraordinary insight to the net's growth so far, Gardner paints a realistic (and hopeful) picture of the ways we can still positively shape the greatest communications phenomenon we've ever known.

 


Bio: Sue Gardner is the former Director of and now special advisor to Wikipedia. She's also the only Canadian to make it onto Forbes' The World's 100 Most Powerful Women list. At Wikipedia, Gardner introduced major initiatives focused on organizational maturity, long-term sustainability, and increased participation, reach, and quality of the Foundation's free-knowledge projects. She was formerly head of CBC.CA, the internet platform for the Canadian Broadcasting Corporation: Canada's radio, TV, and new media public broadcaster and the nation's largest journalistic organization. Under her leadership, CBC.CA experienced a historic audience surge and launched many new major multimedia technologies, including podcasting, breaking news alerts, live event blogging, and many forms of user interaction.
In March 2013, Gardner announced she would be stepping down from Wikimedia. She wrote: "The movement and the Wikimedia Foundation are in a strong place now . . . I feel that although we’re in good shape, with a promising future, the same is not true for the Internet itself . . . Increasingly, I’m finding myself uncomfortable about how the Internet’s developing, who’s influencing its development, and who is not . . . There are many organizations and individuals advocating for the public interest online - what’s good for ordinary people - but other interests are more numerous and powerful than they are. I want that to change. And that’s what I want to do next."

Since beginning her career in 1990 at the now-classic radio program As It Happens, Gardner has worked in all forms of media to create high-quality, award-winning programming. As a journalist, she specialized in pop culture, social issues and media analysis, covering stories such as manipulation of the news media during the first Gulf War, the rise of gated communities in California, the racial implications of the return of the death penalty to New York, changing feminist attitudes towards pornography, the dawn of interactive media, and the rise and fall of rave culture in the UK.

Gardner is a member of the Online News Association, the Society for News Design, Women in Film and Television, the Canadian Association of Journalists and Canadian Women In Communications.

 

March 24th, 2015

Abhishek Bhattacharjee
Computer Science Department
Rutgers University

Title: "Virtual Memory in Next-Generation Heterogeneous Manycore Systems"
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall
Read the Abstract and Bio

Abstract:

Since its inception, virtual memory has become a powerful and ubiquitous abstraction for allocating and managing memory with a flexible and clean programming model. Typically, the systems community has been comfortable paying a performance tax for these programmability benefits. Unfortunately, emerging software with large data requirements and deeper stacks (e.g., large graphs, key value stores, virtualization), and emerging hardware accelerators requiring manual data orchestration by the CPU are increasing this performance tax drastically, while also conceding various programmability benefits of virtual memory.

In this talk, I discuss techniques to reclaim this lost performance and programmability by enriching existing address translation hardware to more elasticity adapt to memory allocation aspects of the operating system. Specifically, I show how hardware support that detects patterns in page table allocation can be used to design low-overhead, high performance address translation hardware. In addition, I discuss how to design memory management units for accelerators in support of unified address spaces. Overall, these techniques are broadly applicable across both server and client systems.

Bio:

Abhishek Bhattacharjee is an assistant professor in the department of computer science at Rutgers University. His interests span the the interactions between architecture and operating systems. Abhishek received his PhD from Princeton University in 2010 and the NSF Career award in 2013.

 

March 31st, 2015

Byron Wallace
Department of Computer Science and School of Information
University of Texas at Austin
Title: " Automating evidence synthesis via machine learning and natural language processing
"
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall
Read the Abstract and Bio

 

Abstract: Evidence-based medicine (EBM) looks to inform patient care with the totality of available relevant evidence. Systematic reviews are the cornerstone of EBM and are critical to modern healthcare, informing everything from national health policy to bedside decision-making. But conducting systematic reviews is extremely laborious (and hence expensive): producing a single review requires thousands of person-hours. Moreover, the exponential expansion of the biomedical literature base has imposed an unprecedented burden on reviewers, thus multiplying costs. Researchers can no longer keep up with the primary literature, and this hinders the practice of evidence-based care.

To mitigate this issue, I will discuss past and recent advances in machine learning and natural language processing methods that look to optimize the practice of EBM. These include methods for semi-automating evidence identification (i.e., citation screening) and more recent work on automating the extraction of structured data from full-text published articles describing clinical trials. As I will discuss, these problems pose challenging problems from a machine learning vantage point, and hence motivate the development of novel approaches. I will present evaluations of these methods in the context of EBM and propose new directions moving forward toward automating evidence synthesis.

Bio: Byron Wallace is an assistant professor in the School of Information at the University of Texas at Austin. He holds a PhD in Computer Science from Tufts University, where he was advised by Carla Brodley. Prior to joining UT, he was research faculty at Brown University, where he was part of the Center for Evidence-Based Medicine and also affiliated with the Brown Laboratory for Linguistic Information Processing. His primary research is in machine learning and natural language processing methods, with an emphasis on their application in health informations (and especially evidence-based medicine).

Wallace's work is supported by grants from the NSF and the ARO. He was recognized as The Runner Up for the 2013 ACM Special Interest Group on Knowledge Discovery and Data Mining (SIG KDD) for his thesis work.

 

April 6th, 2015

Zachary Palmer
Ph.D. candidate, Computer Science Dept.
Johns Hopkins
Title: "Refactoring: The Same is Better"
4:30 pm, Heilmeier Hall (Towne 100)


Read the Abstract and Bio

 

Abstract: To refactor code is to modify it so that it accomplishes the same task in a different way.  Although we make our best efforts to design code to be appropriate for our requirements, those requirements change and, over time, the design becomes less suited.  We address this problem through refactoring: we change the design of the code without changing its behavior.  Although refactoring by itself produces no outward improvement, it acts as an investment: with a design more appropriate to our updated requirements, we are better able to add features and correct mistakes.

In this lecture, we discuss the refactoring process and some basic refactorings by example.  We begin with a code base and some new requirements.  We apply standard software engineering practices such as revision control and unit testing during the process and discuss the abstract definition of each refactoring as it is applied.  By the end of this lecture, we arrive at a codebase which is structurally improved and which includes the new requirements in a pleasing and maintainable way.


Bio: Zachary Palmer is a Ph.D. candidate in Computer Science at Johns Hopkins University, advised by Dr. Scott F. Smith.  His research focuses on the development of typed scripting languages and compile-time metaprogramming.  He previously worked as a software engineer for Amentra, Inc. (now Red Hat Consulting Services) with a specific focus on feature addition and maintenance on enterprise projects.  He has taught numerous courses during his graduate program at JHU, for which he has won the 2015 Professor Joel Dean Excellence in Teaching Award and the 2010 Whiting School of Engineering Outstanding Teaching Award.

 

April 7th, 2015

Sumit Gulwani
Principal Researcher
Microsoft Research
Title: "Data Manipulation using Programming By Examples and Natural Language"
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall
Read the Abstract and Bio

 

Abstract: Data is locked up in semi-structured formats (such as spreadsheets, text/log files, webpages, pdf documents) in both consumer and enterprise space. Getting data out of these documents into structured formats that allow the data to be explored and analyzed is time consuming and error prone. While data scientists typically spend 80% of their time cleaning data, programmatic solutions to data manipulation are beyond the expertise of 99% of those end users who do not know programming.

The paradigms of programming by examples (PBE) and programming by natural language (PBNL) have the potential to make data wrangling a delightful experience for the masses. In order to bring PBE and PBNL technologies to market, two key technical challenges need to be addressed: (a) developing efficient search algorithms that can explore the huge state space of programs to find those that match the user specification, and (b) developing effective ambiguity resolution techniques to deal with the inherent ambiguity in the user specification. Our state-of-the-art search algorithms employ deductive reasoning and domain-specific languages that restrict search space to achieve real-time efficiency. Our ambiguity resolution techniques include machine learning based ranking of synthesized programs, support for navigation between synthesized programs paraphrased into structured English, and active learning based interaction models. In this talk, I will demo few technologies that have been developed using these principles. Some of these technologies have also been shipped as part of major Microsoft products.

Bio: Sumit Gulwani is a principal researcher at Microsoft Research and an adjunct faculty in the Computer Science Department at IIT Kanpur, India. His research interests lie in the cross-disciplinary areas of automating end-user programming and building intelligent tutoring systems. He is a recipient of the ACM SIGPLAN Robin Milner Young Researcher Award. He obtained his PhD from UC-Berkeley in 2005, and was awarded the ACM SIGPLAN Outstanding Doctoral Dissertation Award. He obtained his BTech from IIT Kanpur, and was awarded the President's Gold Medal.

 

 

April 9th, 2015

Zachary Kincaid
University of Toronto
Title: " Parallel Proofs for Parallel Programs"
12:00pm - 1:15pm, 307 Levine Hall
Read the Abstract and Bio

Abstract: Today's software systems are large and complex - beyond the scope of comprehension of any one person.  My research is motivated by the question of how machines can help humans better understand their code and aid the construction of reliable, secure, and efficient systems.  Multi-threading is a long-standing obstacle to reasoning by both humans and machines.  Conventionally, this obstacle is met by developing clever ways to reason about the program as if it were executing sequentially.  In this talk, I will argue that we should embrace parallelism, not hide from it.  I will discuss new *fundamentally parallel* foundations for automated program analysis, which allow the parallelism present in a program to be explicitly maintained and enable tractable automated reasoning and succinct proofs.  In other words: my talk will be on parallel proofs for parallel programs.  

 

April 9th, 2015

Warren Center Distinguished Lecture
Alessandro Acquisti
Carnegie Mellon University
Title: "On the Roots of Privacy Concerns"
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall
Read the Abstract and Bio

 

Abstract: Human beings have evolved to detect and react to threats in their physical environment, and have developed perceptual systems to assess physical, sensorial stimuli for current, material risks. In cyberspace, those stimuli can be absent, subdued, or deliberately manipulated by antagonistic third parties. Security and privacy concerns that would normally be activated in the offline world, therefore, can remain muted, and defense behaviors can be hampered, online. In order to start understanding the interrelationships between online and offline threat detection and online decision making, we investigate the extent to which "visceral" stimuli in the physical world can impact security and privacy behavior in cyberspace. In this talk, I will present the design and results of a stream of controlled human subject experiments that explore the influence of sensorial stimuli (indicating the presence of other human beings in the proximal space of a subject) on subjects' online disclosure of personal, and highly sensitive, behaviors.

 

 

Joint work with Laura Brandimarte (CMU) and Jeff Hancock (Cornell)

 

April 14th, 2015

Jimmy Yang
Senior Director of Research at Yahoo Labs
Title: "Optimization in Online Advertising"
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall
Read the Abstract and Bio

 

Abstract: As a fast growing business, online advertising has become the largest media channel since 2013. With the benefits of flexible and dynamic format, ability to track and measure user response, global coverage and real-time delivery, it creates increasingly higher value for advertisers and publishers. Online advertising is an ecosystem composed of users, publishers and advertisers, who closely interact with each other. Each party has their own goals participating in the ecosystem, where optimization plays a critical role to reach the goals. In this talk, we will first briefly review the characteristics and trend of online advertising. Then we will present some of the important optimization problems, from the perspectives of the publishers and advertisers respectively. In general, the publishers are interested in maximizing their revenue given available inventory of supply, while the advertisers are interested in maximizing their ROI (Return On Investment) given available budget. These problems are modeled in different ways and solved to provide both online and offline solutions that guide the business practice. The talk will be concluded with a summary and discussion of potential research opportunities.

Jimmy Yang is a Senior Director of Research at Yahoo Labs. He leads the group of Advertising Insights, Decisions, and Optimization, which performs applied research in data analysis, model building, algorithm design, offline simulation and online experimentation to create value for advertisers and publishers through areas such as online traffic forecasting, supply and demand optimization, pricing and revenue management, bid and budget recommendation, and ad effectiveness measurement across multiple platforms and devices. Jimmy holds a Ph.D. degree in Electrical and Computer Engineering from University of California at Davis.

 


April 16th, 2015

Dawn Song
Computer Science Division, University of California, Berkeley
Title: "No More Cat and Mouse: Towards Building Systems Secure by Construction"
3:00pm - 4:15pm, Wu and Chen Auditorium, Levine Hall
Read the Abstract and Bio

Abstract: Malicious cyber attacks are wreaking havoc on the Internet and continue to increase in scale, sophistication, and severity. How can we fundamentally break the cat-and-mouse game and change the attack-defense arms race dynamic? In this talk, I will explore different approaches to security and present new techniques to make it much easier to build secure systems. I will show examples from our BitBlaze and WebBlaze project that demonstrate the principle of Secure by Construction, and describe how this approach provides better security than traditional approaches. Our solution has helped secure many widely deployed software systems including high-profile Google applications and FreeBSD. I will discuss future directions in next-generation security solutions.

Bio: Dawn Song is Professor of Computer Science at UC Berkeley. Prior to joining UC Berkeley, she was an Assistant Professor at Carnegie Mellon University from 2002 to 2007. Her research mainly focuses on improving security in the real world, combining techniques from different areas including programming languages, systems, machine learning, and theory. She also has broad interest in other areas of computer science including new computing and programming paradigms. She is the recipient of various awards including the MacArthur Fellowship, the Guggenheim Fellowship, the NSF CAREER Award, the Alfred P. Sloan Research Fellowship, the MIT Technology Review TR-35 Award, the IBM and Google Faculty Award, the George Tallman Ladd Research Award, the Okawa Foundation Research Award, the Li Ka Shing Foundation Women in Science Distinguished Lecture Series Award, and Best Paper Awards from top conferences. She is the founder of Ensighta Security, Inc., acquired by FireEye Inc., and Safely Inc., acquired by Menlo Security.

 


 

TBA

TBA
Read the Abstract and Bio

 

TBA