Select a calendar:
Filter February Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Conferences, Lectures, & Seminars
Events for February
-
CS Colloquium: Vasilis Verroios (Stanford) - Combining Algorithms and Humans for Large-Scale Data Integration
Wed, Feb 01, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Vasilis Verroios , Stanford University
Talk Title: Combining Algorithms and Humans for Large-Scale Data Integration
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Modern enterprises collect data from their operations and the web, and strongly depend on the collected data to make important decisions. To analyze the collected data, enterprises need to first perform data integration, i.e., combine the data from the multiple sources to create a unified set.
Data integration involves some tasks that are still very hard for computer algorithms, like tasks involving images, video, natural language, or data semantics understanding. Since humans may be more accurate with such tasks, the approach of crowdsourcing has been proposed and applied by large companies and research organizations, over the last years. In crowdsourcing, humans are also involved, in order to enhance computer algorithms by completing small tasks, like classifying a forum comment as offensive or ironic. Crowdsourcing drastically improves the accuracy of the outcome compared to using only computer algorithms, however, it does not scale due to the large amount of time (and monetary compensation) required by humans. In this talk, I will discuss how to make crowdsourcing scalable for data integration.
Biography: Vasilis Verroios is a PhD candidate in the Computer Science Department, at Stanford University. His advisor is Hector Garcia-Molina. He received a B.S. and M.S. in Computer Science from the University of Athens, in 2006 and 2008, respectively. In the past, he has been a member of the "Management of Data, Information, & Knowledge Group" at the University of Athens, and he has worked for oDesk and Microsoft Research. His primary interests include data integration, data analytics, and data mining.
Host: Cyrus Shahabi
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Finale Doshi-Velez (Harvard) - Characterizing and Conquering Non-Identifiability in Non-negative Matrix Factorization
Thu, Feb 02, 2017 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Finale Doshi-Velez, Harvard
Talk Title: Characterizing and Conquering Non-Identifiability in Non-negative Matrix Factorization
Series: Yahoo! Labs Machine Learning Seminar Series
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium. Part of Yahoo! Labs Machine Learning Seminar Series.
Nonnegative matrix factorization (NMF) is a popular dimension reduction technique that produces interpretable decomposition of the data into parts. However, this decomposition is often not identifiable, even beyond simple cases of permutation and scaling. Non-identifiability is an important concern in practical data exploration settings, in which the basis of the NMF factorization may be interpreted as having some kind of meaning: it may be important to know that other non-negative characterizations of the data were also possible. While other studies have provide criteria under which NMF is unique, in this talk I'll discuss when and how an NMF might *not* be unique. Then I'll discuss some novel algorithms for characterizing the posterior in Bayesian NMF.
Biography: Finale Doshi-Velez is an Assistant Professor in Computer Science at Harvard University. Prior to that, she was a NSF CiTraCS postdoctoral fellow at Harvard Medical School and a Marshall Scholar at the University of Cambridge. She completed her PhD at MIT. Her interests lie in the intersection of healthcare and machine learning.
Host: Yan Liu
Location: Henry Salvatori Computer Science Center (SAL) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Leilani Battile (CSAIL MIT) - Behavior-Driven Optimizations for Big Data Exploration
Mon, Feb 06, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Leilani Battile , CSAIL MIT
Talk Title: Behavior-Driven Optimizations for Big Data Exploration
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
The physical and biological sciences are becoming more data driven, often due overwhelming quantities of data collected from satellites, telescopes, sequencers, and other sensors. One of the key issues for scientists who work with large datasets is efficient visualization of their data to extract patterns, observe anomalies, and debug their workflows. Though a variety of visualization tools exist to help people make sense of their data, these tools often rely on database management systems (or DBMSs) for data processing and storage; and unfortunately, DBMSs fail to process the data fast enough to support a fluid, interactive visualization experience.
My work blends optimization techniques from databases and methodology from HCI and visualzation in order to support interactive and iterative exploration of large datasets. In this talk, I will discuss Sculpin, a visual exploration system that learns user exploration patterns automatically, and exploits these patterns to pre-fetch data ahead of users as they explore. I will show that Sculpin's pre-fetching techniques provide significant performance benefits compared to existing systems. I will then discuss our ongoing work with Sculpin, which aims to avoid wasting computational resources, while still providing a fluid, interactive exploration experience for users. To do this, we combine data-prefetching with incremental data processing and visualization-focused caching optimizations, and incorporate these techniques in Sculpin to further boost performance.
Biography: Leilani Battile is a Computer Science Ph.D. candidate in the CSAIL Database Group at MIT, advised by Prof. Michael Stonebraker. Her research interests lie at the intersection of data management, user interface design, and visual analytics, with the aim of building intuitive and scalable database exploration tools. She was a National Science Foundation Graduate Fellow from 2011 to 2013. She obtained a M.S. from MIT in 2013, and a B.S. in Computer Engineering from the University of Washington in 2011.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Heather Culbertson (Stanford University) - Realistic and Intuitive Haptic Feedback for Communication in Virtual and Real-World Environments
Tue, Feb 07, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Heather Culbertson, Stanford University
Talk Title: Realistic and Intuitive Haptic Feedback for Communication in Virtual and Real-World Environments
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
The haptic (touch) sensations felt when interacting with the physical world create a rich and varied impression of objects and their environment. Humans are capable of gathering a significant amount of information through touch with their environment, allowing them to assess object properties and qualities, dexterously handle objects, and communicate social cues and emotions. Humans are spending significantly more time in the digital world, however, and are increasingly interacting with people and objects through a digital medium. Unfortunately, digital interactions remain unsatisfying and limited, representing the human as having only two sensory inputs: visual and auditory.
This talk will focus on the investigation of haptic devices and rendering algorithms to provide humans with touch information when communicating through a computer. I will present a background on the sense of touch, and illustrate how we can leverage this knowledge in order to design haptic devices and rendering systems that allow the human to communicate through the digital world in a natural and intuitive way. I will highlight contributions I have made in furthering haptic realism in virtual reality through the creation of highly realistic virtual objects. These objects are created by modeling high-frequency acceleration, force, and speed data recorded during physical interactions and displaying the appropriate haptic signals during rendering. I will then describe advances I have made in novel wearable haptic devices for communicating information to a human using intuitive and natural cues.
Biography: Heather Culbertson is a Postdoctoral Research Fellow in the Department of Mechanical Engineering at Stanford University where she works in the Collaborative Haptics and Robotics in Medicine (CHARM) Lab. She received her PhD in the Department of Mechanical Engineering and Applied Mechanics (MEAM) at the University of Pennsylvania in 2015 working in the Haptics Group, part of the General Robotics, Automation, Sensing and Perception (GRASP) Laboratory. She completed a Masters in MEAM at the University of Pennsylvania in May of 2013, and earned a BS degree in mechanical engineering at the University of Nevada, Reno in 2010.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Dorsa Sadigh (UC Berkeley) -Towards a Theory of Safe and Interactive Autonomy
Thu, Feb 09, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Dorsa Sadigh, UC Berkeley
Talk Title: Towards a Theory of Safe and Interactive Autonomy
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Today's society is rapidly advancing towards cyber-physical systems (CPS) that interact and collaborate with humans, e.g., semi-autonomous vehicles interacting with drivers and pedestrians, medical robots used in collaboration with doctors, or service robots interacting with their users in smart homes. The safety-critical nature of these systems requires us to provide provably correct guarantees about their performance in interaction with humans. The goal of my research is to enable such human-cyber-physical systems (h-CPS) to be safe and interactive. I aim to develop a formalism for design of algorithms and mathematical models that facilitate correct-by-construction control for safe and interactive autonomy.
In this talk, I will first discuss interactive autonomy, where we use algorithmic human-robot interaction to be mindful of the effects of autonomous systems on humans, and further leverage these effects for better safety, efficiency, coordination, and estimation. I will then talk about safe autonomy, where we provide correctness guarantees, while taking into account the uncertainty arising from the environment. Further, I will discuss a diagnosis and repair algorithm for systematic transfer of control to the human in unrealizable settings. While the algorithms and techniques introduced can be applied to many h-CPS applications, in this talk, I will focus on the implications of my work for semi-autonomous driving.
Biography: Dorsa Sadigh is a Ph.D. candidate in the Electrical Engineering and Computer Sciences department at UC Berkeley. Her research interests lie in the intersection of control theory, formal methods, and human-robot interactions. Specifically, she works on developing a unified framework for safe and interactive autonomy. Dorsa received her B.S. from Berkeley EECS in 2012. She was awarded the NDSEG and NSF graduate research fellowships in 2013. She was the recipient of the 2016 Leon O. Chua department award and the 2011 Arthur M. Hopkin department award for achievement in the field of nonlinear science, and she received the Google Anita Borg Scholarship in 2016.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Anirudh Sivaraman (CSAIL MIT) - Making the fastest routers programmable
Mon, Feb 13, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Anirudh Sivaraman, CSAIL MIT
Talk Title: Making the fastest routers programmable
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Historically, the evolution of network routers was driven primarily by performance. Recently, owing to the need for better control over network operations and the constant demand for new features, programmability of routers has become as important as performance.
However, today's fastest routers, which run at line rate, use fixed-function hardware, which cannot be modified after deployment. I will describe two router primitives we have developed to build programmable routers at line rate. The first is a programmable pocket scheduler. The second is a way to execute stateful packet-processing algorithms to manage network resources. Together, these primitives allow us to program several packet-processing functions at line rate, such as in-network congestion control, active queue management, data-plane load balancing, network measurement, and packet scheduling.
This talk is based on joint work with collaborators at MIT, Barefoot Networks, Cisco Systems, Microsoft Research, Stanford University, and the University of Washington.
Biography: Anirudh Sivaraman is a Ph.D. student at MIT, advised by Hari Balakrishnan and Mohammad Alizadeh. His recent research work has focused on hardware and software for programmable high-speed routers. He has also been actively involved in the design and evolution of the P4 language for programmable network devices. His past research includes work on congestion control, network emulation, improving Web performance, and network measurement. He received the MIT EECS department's Frederick C. Hennie III Teaching Award in 2012 and shared the Internet Research Task Force's Applied Networking Research Prize in 2014.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Aurojit Panda (UC Berkeley) - A New Approach to Network Functions
Tue, Feb 14, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Aurojit Panda, UC Berkeley
Talk Title: A New Approach to Network Functions
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Modern networks do far more than just deliver packets, and provide network functions -- including firewalls, caches, and WAN optimizers -” that are crucial for scaling networks, ensuring security and enabling new applications. Network functions were traditionally implemented using dedicated hardware middleboxes, but in recent years they are increasingly being deployed as VMs on commodity servers. While many herald this move towards network function virtualization (NFV) as a great step forward, I argue that accepted virtualization techniques are ill-suited to network functions. In this talk I describe NetBricks -” a new approach to building and running virtualized network functions that speeds development and increases performance. I end the talk by discussing the implications of being able to easily create and insert new network functions.
Biography: Aurojit Panda is a PhD candidate in Computer Science at the University of California Berkeley, where he is advised by Scott Shenker . His work spans programming languages, networking and systems, and his recent work has investigated network verification, consensus algorithms in software defined networks and frameworks for building network functions.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Theodoros Rekatsinas (Stanford University) - Data Integration with Unreliable Sources
Wed, Feb 15, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Theodoros Rekatsinas, Stanford University
Talk Title: Data Integration with Unreliable Sources
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Data integration is an essential element of data-intensive science and modern analytics. Users often need to combine data from different sources to gain new scientific knowledge, obtain accurate insights, and create new services. However, today's upsurge in the number and heterogeneity-”in terms of format and reliability-”of data sources limits the ability of users to reason about the value of data. This raises the fundamental questions: what makes a data source useful to end users, how can we integrate unreliable data, and which sources we need to combine to maximize the user's utility?
In this talk, I discuss how to assess and leverage the quality and reliability of data to make data integration more efficient. Specifically, I demonstrate how statistical learning is the key to managing large volumes of heterogeneous sources effectively. Building upon this observation, I introduce new solutions to classical data integration problems, such as data conflict resolution and data cleaning, and show that these solutions outperform their traditional counterparts by large margins. I finish with an outlook on how recent advancements in machine learning have the potential to streamline the construction of end-to-end data curation systems and bring data closer to users.
Biography: Theodoros (Theo) Rekatsinas is a Moore Data Postdoctoral Fellow at Stanford working with Christopher Ré; he earned his Ph.D. in Computer Science from the University of Maryland, where he was advised by Amol Deshpande and Lise Getoor. His research interests are in data management, with a focus on data integration, data cleaning, and uncertain data. Theo's work on using quality-aware data integration techniques to forecast the emergence and progression of disease outbreaks received the Best Paper Award at SDM 2015. Theo was awarded the Larry S. Davis Doctoral Dissertation award in 2015.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Rashmi K. Vinayak (UC Berkeley) - Smart redundancy for big-data systems: Theory and Practice
Thu, Feb 16, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Rashmi K. Vinayak, UC Berkeley
Talk Title: Smart redundancy for big-data systems: Theory and Practice
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Large-scale distributed storage and caching systems form the foundation of big-data systems. A key scalability challenge in distributed storage systems is achieving fault tolerance in a resource-efficient manner. Towards addressing this challenge, erasure codes provide a storage-efficient alternative to the traditional approach of data replication. However, classical erasure codes come with critical drawbacks: while optimal in utilizing storage space, they significantly increase the usage of other important cluster resources such as network and I/O. In the first part of the talk, I present new erasure codes and theoretical optimality guarantees. The proposed codes reduce the network and I/O usage by 35-70% for typical parameters while retaining the storage efficiency of classical codes. I then present an erasure-coded storage system that employs the proposed codes, and demonstrate significant benefits over the state-of-the-art in evaluations under production setting at Facebook. Our codes have been incorporated into Apache Hadoop 3.0. The second part of the talk focuses on achieving high performance in distributed caching systems. These systems routinely face the challenges of skew in data popularity, background traffic imbalance, and server failures, which result in load imbalance across servers and degradation in read latencies. I present EC-Cache, a cluster cache that employs erasure coding to achieve a 3-5x improvement as compared to the state-of-the-art.
Biography: Rashmi K. Vinayak is a postdoctoral researcher in the EECS department at UC Berkeley, where she received her PhD in 2016. Her dissertation received the Eli Jury Award 2016 from the EECS department at UC Berkeley for outstanding achievement in the area of systems, communications, control, or signal processing. Rashmi is also a recipient of the Facebook Fellowship 2012-13, the Microsoft Research PhD Fellowship 2013-15, and the Google Anita Borg Memorial Scholarship 2015-16. She is also the recipient of the IEEE Data Storage Best Paper and Best Student Paper Awards for the years 2011/2012. Her research interests lie in the theoretical and system challenges that arise in storage and analysis of big data, with a current focus on erasure coding for big-data systems.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Shivaram Venkataraman (UC Berkeley) - Scalable Systems for Fast and Easy Machine Learning
Tue, Feb 21, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Shivaram Venkataraman, UC Berkeley
Talk Title: Scalable Systems for Fast and Easy Machine Learning
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Machine learning models trained on massive datasets power a number of applications; from machine translation to detecting supernovae in astrophysics. However the end of Moore's law and the shift towards distributed computing architectures presents many new challenges for building and executing such applications in a scalable fashion.
In this talk I will present my research on systems that make it easier to develop new machine learning applications and scale them while achieving high performance. I will first present programming models that let users easily build distributed machine learning applications. Next, I will show how we can exploit the structure of machine learning workloads to build low-overhead performance models that can help users understand scalability and simplify large scale deployments. Finally, I will describe scheduling techniques that can improve scalability and achieve high performance when using distributed data processing frameworks.
Biography: Shivaram Venkataraman is a PhD Candidate at the University of California, Berkeley and is advised by Mike Franklin and Ion Stoica. His research interests are in designing systems and algorithms for large scale data processing and machine-learning. He is a recipient of the Siebel Scholarship and best-of-conference citations at VLDB and KDD. Before coming to Berkeley, he completed his M.S at the University of Illinois, Urbana-Champaign.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Xiang Ren (UIUC) - Effort-Light StructMine: Turning Massive Corpora into Structures
Thu, Feb 23, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Xiang Ren, UIUC
Talk Title: Effort-Light StructMine: Turning Massive Corpora into Structures
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
The real-world data, though massive, are hard for machines to resolve as they are largely unstructured and in the form of natural-language text. One of the grand challenges is to turn such massive corpora into machine-actionable structures. Yet, most existing systems have heavy reliance on human effort in the process of structuring various corpora, slowing down the development of downstream applications.
In this talk, I will introduce a data-driven framework, Effort-Light StructMine, that extracts structured facts from massive corpora without explicit human labeling effort. In particular, I will discuss how to solve three StructMine tasks under Effort-Light StructMine framework: from identifying typed entities in text, to fine-grained entity typing, to extracting typed relationships between entities. Together, these three solutions form a clear roadmap for turning a massive corpus into a structured network to represent its factual knowledge. Finally, I will share some directions towards mining corpus-specific structured networks for knowledge discovery.
Biography: Xiang Ren is a Computer Science PhD candidate at University of Illinois at Urbana-Champaign, working with Jiawei Han and the Data and Information System ï¼DAISï¼Research Lab. Xiang's research develops data-driven methods for turning unstructured text data into machine-actionable structures. More broadly, his research interests span data mining, machine learning, and natural language processing, with a focus on making sense of massive text corpora. His research has been recognized with a Google PhD Fellowship, Yahoo!-DAIS Research Excellence Award, C. W. Gear Outstanding Graduate Student Award, and has been transferred to US Army Research Lab, NIH, Microsoft, Yelp and TripAdvisor.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Ellie Pavlick (University of Pennsylvania) - Natural Language Understanding with Paraphrases and Composition
Mon, Feb 27, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Ellie Pavlick, University of Pennsylvania
Talk Title: Natural Language Understanding with Paraphrases and Composition
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Natural language processing (NLP) aims to teach computers to understand human language. NLP has enabled some of the most visible applications of artificial intelligence, including Google search, IBM Watson, and Apple's Siri. As AI is applied to increasingly complex domains such as health care, education, and government, NLP will play a crucial role in allowing computational systems to access the vast amount of human knowledge documented in the form of unstructured speech and text.
In this talk, I will discuss my work on training computers to make inferences about what is true or false based on information expressed in natural language. My approach combines machine learning with insights from formal linguistics in order to build data-driven models of semantics which are more precise and interpretable than would be possible using linguistically naive approaches. I will begin with my work on automatically adding semantic annotations to the 100 million phrase pairs in the Paraphrase Database (PPDB). These annotations provide the type of information necessary for carrying out precise inferences in natural language, transforming the database into a largest available lexical semantics resource for natural language processing. I will then turn to the problem of compositional entailment, and present an algorithm for performing inferences about long phrases which are unlikely to have been observed in data. Finally, I will discuss my current work on pragmatic reasoning: when and how humans derive meaning from a sentence beyond what is literally contained in the words. I will describe the difficulties that such "common-sense" inference poses for automatic language understanding, and present my on-going work on models for overcoming these challenges.
Biography: Ellie Pavlick is a PhD student at the University of Pennsylvania, advised by Dr. Chris Callison-Burch. Her dissertation focuses on natural language inference and entailment. Outside of her dissertation research, Ellie has published work on stylistic variation in paraphrase--e.g. how paraphrases can effect the formality or the complexity of language--and on applications of crowdsourcing to natural language processing and social science problems. She has been involved in the design and instruction of Penn's first undergraduate course on Crowdsourcing and Human Computation (NETS 213). Ellie is a 2016 Facebook PhD Fellow, and has interned at Google Research, Yahoo Labs, and the Allen Institute for Artificial Intelligence.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CAIS Seminar Series: Dr. Pascal Van Hentenryck (University of Michigan) - The Case of Infrastructure Optimization
Mon, Feb 27, 2017 @ 04:00 PM - 04:50 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Dr. Pascal Van Hentenryck, University of Michigan
Talk Title: The Case of Infrastructure Optimization
Series: Center for AI in Society (CAIS) Seminar Series
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
In the last decade, massive amount of information has been collected about critical infrastructures, including the transportation network and the electrical power system. These data sets, together with progress in Artificial Intelligence and Operations Research, make it possible to analyze, predict, and optimize these infrastructures with unprecedented fidelity. This talk demonstrates the societal benefits of this transformation on a number of case studies in evacuation planning, public transportation, and power restoration.
Biography: Dr. Pascal Van Hentenryck is the Seth Bonder Collegiate Professor of Engineering at the University of Michigan. He is Professor of Industrial and Operations Engineering, Professor of Electrical Engineering and Computer Science, and core faculty in the Michigan Institute of Data Science. He is the author of the pioneering CHIP and OPL optimization systems, which have been widely used in academia and industry. Dr. Van Hentenryck is the author of five MIT Press books and is a fellow of AAAI and INFORMS.
Host: Milind Tambe
Location: John Stauffer Science Lecture Hall (SLH) - 100
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Vinodkumar Prabhakaran (Stanford University) - NLP for Social Good: Inferring Social Context from Language
Tue, Feb 28, 2017 @ 11:00 AM - 12:20 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Vinodkumar Prabhakaran, Stanford University
Talk Title: NLP for Social Good: Inferring Social Context from Language
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
The vast quantities of language data online and offline offer tremendous opportunities to study society through language. In this talk, I show how natural language processing techniques can be expanded from understanding the meanings of words and sentences, to inferring the underlying social structures and processes they reflect and identifying crucial shortcomings in them. I apply these techniques to computationally detect two ways in which the social context affects the use of language: social relations affecting how people interact with one another, and social constructs shaping how institutions interact with communities. In the first part, I show how to computationally detect manifestations of social power in workplace interactions between individuals -” providing means for organizations to detect incivility at workplace. In the second part, I show how to computationally investigate the ways race shapes the interactions between the police and the communities they serve -” providing means for departments to address and monitor racial disparities in policing. My research looks beyond words and phrases, and introduce ways to infer richer rhetorical and dialog information like conversational structure and respect that reflect the social context, demonstrating the importance of deeper language processing for the computational social sciences.
Biography: Vinodkumar Prabhakaran is a postdoctoral fellow in the computer science department at Stanford University. His research falls in the inter-disciplinary field of computational sociolinguistics, in which he builds and uses computational tools to analyze linguistic patterns that reveal the underlying social contexts in which language is used. He received his PhD in Computer Science from Columbia University in 2015. In his doctoral thesis, he studied how machine learning and natural language processing techniques can help detect the underlying social power structures that guide social interactions. As part of his research, he has also made significant contributions to core NLP problems such as extracting information from text, as well as modeling structures of dialog and discourse.
Host: CS Department
Location: Ronald Tutor Hall of Engineering (RTH) - 217
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
-
CS Colloquium: Fan Long (MIT CSAIL) - Learning How to Patch Software Errors Automatically
Tue, Feb 28, 2017 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Fan Long, MIT CSAIL
Talk Title: Learning How to Patch Software Errors Automatically
Series: CS Colloquium
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium.
Software systems are increasingly integrated into every part of our society. As the number of systems and our dependence on these systems continue to grow, making these systems reliable and secure becomes an increasingly important challenge for our society and a daunting task for software developers.
Automatic patch generation holds out the promise of automatically correcting software defects without the need for developers to manually diagnose, understand, and correct these defects. In this talk, I will present two novel automatic patch generation systems, Prophet and Genesis, both of which learn from past successful human patches to automatically fix defects. By collectively leveraging development efforts worldwide, Prophet and Genesis automatically generate correct patches for real-world defects in large open-source C and Java applications with up to millions lines of code. This research also demonstrates that the growing volume of software programs is not just a challenge but also a great opportunity. Exploiting this opportunity can enable revolutionary new automated techniques that enhance software reliability and security.
Biography: Fan Long is a PhD candidate in Computer Science at Massachusetts Institute of Technology (MIT). His research to date has focused on developing automated programming systems to improve software reliability and security. He has developed systems that automatically identify and eliminate errors in large software programs and systems that enable software programs to operate successfully in spite of the presence of errors. He holds a BE from Tsinghua University and a MS from MIT.
Host: CS Department
Location: Henry Salvatori Computer Science Center (SAL) - 101
Audiences: Everyone Is Invited
Contact: Assistant to CS chair
