Conferences, Lectures, & Seminars
Events for April
-
AI SEMINAR
Fri, Apr 01, 2016 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Mahdi Soltanolkotabi, Assistant Professor at USC
Talk Title: Finding low-complexity models without the shackles of convexity
Series: AI Seminar
Abstract: In many applications, one wishes to estimate a large number of parameters from highly incomplete data samples. Low-dimensional models such as sparsity, low-rank, etc provide a principled approach for addressing the challenges posed by such high-dimensional data. The last decade has witnessed a flurry of activity in understanding when and how it is possible to find low complexity models via convex relaxations. However, the computational cost of such convex schemes can be prohibitive. In fact, in this talk I will argue that over insistence on convex methods has stymied progress in many application domains. I will discuss my ongoing research efforts to unshackle such problems from the confines of convexity opening the door for new applications.
I will discuss three concrete problems characterized by incomplete information about a low-complexity object of interest. The first is the century-old phase retrieval problem where one wishes to recover a signal from magnitude only measurements--phase information is completely missing. The second is a problem in data analysis, where we observe only a few incomplete linear measurements from a data matrix (e.g. a few entries) and wish to reliably infer all of the entries of the matrix. The third problem involves the recovery of a structured image from highly compressed information--most measurements are missing. To retrieve seemingly lost information I will present novel non-convex algorithms for these problems. Surprisingly, despite the lack of convexity these algorithms can provably converge to the global optimum and hence impute the missing information precisely.
Biography: Mahdi Soltanolkotabi completed his Ph.D. in electrical engineering at Stanford University in 2014. He was a postdoctoral researcher in the Algorithms, Machines, and People AMP lab and the EECS and Statistics departments at UC Berkeley during the 2014-2015 academic year. His research focuses on design and mathematical understanding of computationally efficient algorithms for optimization, high dimensional statistics, machine learning, signal processing and computational imaging. Recently, a main focus of his research has been on developing and analyzing algorithms for non-convex optimization with provable guarantees of convergence to the global optimum.
WILL NOT BE WEBCASTED
Host: Emilio Ferrara
Location: Information Science Institute (ISI) - 1135 - 11th fl Large CR
Audiences: Everyone Is Invited
Contact: Alma Nava / Information Sciences Institute
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
NL Seminar-Harnessing reviews to build richer models of opinions
Fri, Apr 01, 2016 @ 03:00 PM - 04:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Julian McAuley , UCSD
Talk Title: Harnessing reviews to build richer models of opinions
Series: Natural Language Seminar
Abstract: Online reviews are often our first port of call when considering products and purchases online. Yet navigating huge volumes of reviews (many of which we might disagree with) is laborious, especially when we are interested in some niche aspect of a product. This suggests a need to build models that are capable of capturing the complex and idiosyncratic semantics of reviews, in order to build richer and more personalized recommender systems. In this talk I'll discuss three such directions: First, how can reviews be harnessed to better understand the dimensions (or facets) of people's opinions? Second, how can reviews be used to answer targeted questions, that may be subjective or require personalized responses? And third, how can reviews themselves be synthesized, so as to predict what a reviewer would say, even for products they haven't seen yet?
Biography: Dr. McAuley has been an Assistant Professer in the Computer Science Department at the University of California, San Diego since 2014. Previously he was a postdoctoral scholar at Stanford University after receiving his PhD from the Australian National University in 2011. His research is concerned with developing predictive models of human behavior using large volumes of online activity data.
Host: Xing Shi and Kevin Knight
More Info: http://nlg.isi.edu/nl-seminar/
Location: Information Science Institute (ISI) - 11th Flr Conf Rm # 1135, Marina Del Rey
Audiences: Everyone Is Invited
Contact: Peter Zamar
Event Link: http://nlg.isi.edu/nl-seminar/
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
AI Seminar-Prominent features of rumors in social networks
Fri, Apr 08, 2016 @ 11:00 AM - 11:45 AM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Meeyoung Cha , KAIST
Talk Title: Prominent features of rumors in social networks
Series: Artificial Intelligence Seminar
Abstract: *This is the First of 2 AI Seminar Talks on FRI. 4/8
Social psychology literature defines a rumor as a story in general circulation without confirmation or certainty to facts. Rumors arise in the context of ambiguity, when the meaning of a situation is not readily apparent or when people feel an acute need for
Security. Rumors hence are a powerful, pervasive, and persistent force affecting people and groups. This talk will introduce efforts on identifying rumors using massive data in social media. I will discuss the distinct patterns we observed from rumor diffusions in terms of the following aspects: temporal, structural, and linguistic.
(Published at IEEE 13th International Conference on Data Mining Conference 2013, Joint work with Sejeong Kwon, Kyomin Jung, Wei Chen, Yajun Wang)
Biography: Meeyoung Cha is an associate professor at Graduate School of Culture Technology in KAIST and currently a Visiting Professor at Facebook. Her research interests are in the analysis of large-scale online social networks with emphasis the spread of information, moods, and user influence. She received the best paper awards at ACM IMC 2007 for analyzing long-tail videos in YouTube and at ICWSM 2012 for studying social conventions in Twitter. Her research has been published in leading journals and conferences including PLoS One, Information Sciences, WWW, and ICWSM, and has been featured at the popular media outlets including the New York Times websites, Harvard Business Review's research blog, the Washington Post, the New Scientist.
Host: Emilio Ferrara
Webcast: http://webcasterms1.isi.edu/mediasite/Viewer/?peid=270f829804634fd8b615e50d00f243e41dLocation: Information Science Institute (ISI) - 11th Flr Conf Rm # 1135, Marina Del Rey
WebCast Link: http://webcasterms1.isi.edu/mediasite/Viewer/?peid=270f829804634fd8b615e50d00f243e41d
Audiences: Everyone Is Invited
Contact: Peter Zamar
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
NL Seminar-Learning Distributed Representations from Network Data and Human Navigation
Fri, Apr 08, 2016 @ 03:00 PM - 04:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Hao Wu, USC/ISI
Talk Title: Learning Distributed Representations from Network Data and Human Navigation
Series: Natural Language Seminar
Abstract: The increasing growth of network data such as linked documents on the Web and social networks, has imposed great challenges on automatic data analysis. We study the problem of learning representations of network data, which is of critical for applications including data classification, ranking and link prediction. We present neural network embedding algorithms to learn distributed representations of network data that capture the deep context of each data point, and human cognition in navigation data. To improve the scalability of our algorithms, we use efficient optimization and sampling methods.
Biography: Hao Wu is a PhD student at USC/ISI, advised by Kristina Lerman.
Host: Xing Shi and Kevin Knight
More Info: http://nlg.isi.edu/nl-seminar/
Location: Information Science Institute (ISI) - 11th Flr Conf Rm # 1135, Marina Del Rey
Audiences: Everyone Is Invited
Contact: Peter Zamar
Event Link: http://nlg.isi.edu/nl-seminar/
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
NL Seminar- Decoding Neuro-Semantic Representation of Stories across Languages
Fri, Apr 15, 2016 @ 03:00 PM - 04:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Morteza Dehghani (USC), USC/ISI
Talk Title: Decoding Neuro-Semantic Representation of Stories across Languages
Series: Natural Language Seminar
Abstract: Understanding how conceptual knowledge is represented and organized in the human brain is one of the core problems of cognitive science, and many studies have aimed at exploring and understanding the similarities of neuro-semantic representations of concepts. A general approach that has been particularly fruitful in this domain is the investigation of the relationship between various corpus statistics of words and neural activity during exposure to those words. In this work, we examine the neuro-semantic representations of stories across three different languages. We demonstrate that using new advances in vector-based representation of text and paragraphs, fMRI signals can be reliably mapped to story representations. We also show that such representations can capture common neuro-semantic representation of stories across different languages. Finally, performing search-light analysis using over a billion regressions, we show that activation patterns in the default mode network of the brain are the most reliable features for decoding stories.
Biography: Morteza is an Assistant Professor of psychology, computer science and the Brain and Creativity Institute at University of Southern California. His research spans the boundary between psychology and artificial intelligence, as does his education. His work investigates properties of cognition by using documents of the social discourse, such as narratives, social media, transcriptions of speeches and news articles, in conjunction to behavioral studies.
Host: Xing Shi and Kevin Knight
More Info: http://nlg.isi.edu/nl-seminar/
Location: Information Science Institute (ISI) - 11th Flr Conf Rm # 1135, Marina Del Rey
Audiences: Everyone Is Invited
Contact: Peter Zamar
Event Link: http://nlg.isi.edu/nl-seminar/
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
AI SEMINAR
Mon, Apr 18, 2016 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Peter Fennell, Postdoc at Univ. of Limerick, Ireland
Talk Title: Information diffusion on Twitter: a hazard rates approach
Series: AI Seminar
Abstract: Online social networks provide a digital footprint of the manner in which people interact in their everyday lives. Information diffuses through online social networks as a result of users creating and sharing information, and such diffusions are highly complex because of the nature of individuals and because of the structure of the network that connects them.
In this talk, I will discuss an approach to understanding information diffusion on online social networks, with specific reference to a case study on Twitter. Here, we use hazard rates to quantify how individuals respond to information, and show the non-linear nature of the response of individuals to multiple signals from their peers. Such hazard rates can be implemented in mathematical frameworks for understanding and predictive purposes, allowing us to analyze the information diffusion and its dependence on the structure of the online social network. I will discuss the importance of empirical observations of large social network datasets in constructing mathematical models of information diffusion, and the necessity of such realistic models for both prediction and analysis.
Biography: Peter Fennell is a postdoctoral researcher in the Dept. of Mathematics and Statistics in the University of Limerick, Ireland. Peter's research focuses on diffusive processes on complex networks, and in using probabilistic frameworks to understand such processes and their interplay with the network through which they spread. Recently, Peter has been awarded a James S. McDonnell postdoctoral fellowship to further his research in the area of information diffusion on online social networks. This project will combine extensive empirical examinations of online social networks along with mathematical modeling to gain an understanding of the underlying mechanisms behind information diffusion
Host: Kristina Lerman
Webcast: http://webcasterms1.isi.edu/mediasite/Viewer/?peid=336e5bb472104d289de47f5e8ef7331c1dLocation: Information Science Institute (ISI) - 1135 - 11th fl Large CR
WebCast Link: http://webcasterms1.isi.edu/mediasite/Viewer/?peid=336e5bb472104d289de47f5e8ef7331c1d
Audiences: Everyone Is Invited
Contact: Alma Nava / Information Sciences Institute
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
AI Seminar
Fri, Apr 29, 2016 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Naira Hovakimyan, Professor, University Illinois Urbana Champagne
Talk Title: Aerial Co-robots of Future: How Far We Are?
Abstract: The presentation will give a historical overview of flight control technology from its inception till its maturation. Parallel developments in aerial robotics will be reviewed from the perspective of aerospace industry standards, prioritizing safety, resilience and reliability of operations. Special focus will be placed on cooperative control of UAVs for various mission scenarios in military operations. Flight tests of a subscale commercial jet at NASA and Learjet at Edwards Air Force base will be used to demonstrate the efficiency of the methods developed over the past ten years. Lessons learned will be summarized, and the opportunities in public safety, elderly care, package delivery, precision farming and digital agriculture will be discussed.
Biography: Bio of Naira Hovakimyan
Naira Hovakimyan graduated with MS degree in Theoretical Mechanics and Applied Mathematics in 1988 from Yerevan State University in Armenia. She got her Ph.D. in Physics and Mathematics in 1992, in Moscow, from the Institute of Applied Mathematics of Russian Academy of Sciences, majoring in optimal control and differential games. Before joining the faculty of UIUC in 2008, she has spent time as a research scientist at Stuttgart University in Germany, at INRIA in France, at Georgia Institute of Technology, and she was on faculty of Aerospace and Ocean engineering of Virginia Tech during 2003-2008. She is currently W. Grafton and Lillian B. Wilkins Professor of Mechanical Science and Engineering at UIUC. In 2015 she was named as inaugural director for Intelligent Robotics Lab of CSL at UIUC. She has co-authored a book and more than 300 refereed publications. She is the recipient of the SICE International scholarship for the best paper of a young investigator in the VII ISDG Symposium (Japan, 1996), the 2011 recipient of AIAA Mechanics and Control of Flight award and the 2015 recipient of SWE Achievement Award. In 2014 she was awarded the Humboldt prize for her lifetime achievements and was recognized as Hans Fischer senior fellow of Technical University of Munich. In 2015 she was recognized by UIUC Engineering Council award for Excellence in Advising. She is an associate fellow and life member of AIAA, a Senior Member of IEEE, and a member of SIAM, AMS, SWE, ASME and ISDG. Naira is co-founder of IntelinAir, Inc., a company that commercializes data-drones for delivering actionable information from aerial imagery for various industries. Her work in robotics for elderly care was featured in the New York Times. Her research interests are in the theory of robust adaptive control and estimation, control in the presence of limited information, networks of autonomous systems, game theory and applications of those in safety-critical systems of aerospace, mechanical, electrical, petroleum and biomedical engineering.
Host: Aram Galstyan
More Info: TBA
Location: Information Science Institute (ISI) - 11th floor large conference room
Audiences: Everyone Is Invited
Contact: Kary LAU
Event Link: TBA
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
AI Seminar
Fri, Apr 29, 2016 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Naira Hovakimyan, Professor, University Illinois Urbana Champagne
Talk Title: Aerial Co-robots of Future: How Far We Are?
Abstract: The presentation will give a historical overview of flight control technology from its inception till its maturation. Parallel developments in aerial robotics will be reviewed from the perspective of aerospace industry standards, prioritizing safety, resilience and reliability of operations. Special focus will be placed on cooperative control of UAVs for various mission scenarios in military operations. Flight tests of a subscale commercial jet at NASA and Learjet at Edwards Air Force base will be used to demonstrate the efficiency of the methods developed over the past ten years. Lessons learned will be summarized, and the opportunities in public safety, elderly care, package delivery, precision farming and digital agriculture will be discussed.
Biography: Bio of Naira Hovakimyan
Naira Hovakimyan graduated with MS degree in Theoretical Mechanics and Applied Mathematics in 1988 from Yerevan State University in Armenia. She got her Ph.D. in Physics and Mathematics in 1992, in Moscow, from the Institute of Applied Mathematics of Russian Academy of Sciences, majoring in optimal control and differential games. Before joining the faculty of UIUC in 2008, she has spent time as a research scientist at Stuttgart University in Germany, at INRIA in France, at Georgia Institute of Technology, and she was on faculty of Aerospace and Ocean engineering of Virginia Tech during 2003-2008. She is currently W. Grafton and Lillian B. Wilkins Professor of Mechanical Science and Engineering at UIUC. In 2015 she was named as inaugural director for Intelligent Robotics Lab of CSL at UIUC. She has co-authored a book and more than 300 refereed publications. She is the recipient of the SICE International scholarship for the best paper of a young investigator in the VII ISDG Symposium (Japan, 1996), the 2011 recipient of AIAA Mechanics and Control of Flight award and the 2015 recipient of SWE Achievement Award. In 2014 she was awarded the Humboldt prize for her lifetime achievements and was recognized as Hans Fischer senior fellow of Technical University of Munich. In 2015 she was recognized by UIUC Engineering Council award for Excellence in Advising. She is an associate fellow and life member of AIAA, a Senior Member of IEEE, and a member of SIAM, AMS, SWE, ASME and ISDG. Naira is co-founder of IntelinAir, Inc., a company that commercializes data-drones for delivering actionable information from aerial imagery for various industries. Her work in robotics for elderly care was featured in the New York Times. Her research interests are in the theory of robust adaptive control and estimation, control in the presence of limited information, networks of autonomous systems, game theory and applications of those in safety-critical systems of aerospace, mechanical, electrical, petroleum and biomedical engineering.
Host: Aram Galstyan
More Info: TBA
Location: Information Science Institute (ISI) - 11th floor large conference room
Audiences: Everyone Is Invited
Contact: Kary LAU
Event Link: TBA
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
NL Seminar-Deep learning solutions to computational phenotyping in health care
Fri, Apr 29, 2016 @ 03:00 PM - 04:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Zhengping Che, USC
Talk Title: Deep learning solutions to computational phenotyping in health care
Series: Natural Language Seminar
Abstract: Exponential growth in electronic health care data has resulted in new opportunities and urgent needs to discover meaningful data-driven representations and patterns of diseases. Recent rise of this research field with more available data and new applications also has introduced several challenges. In this talk, we will present our deep learning solutions to address some of the challenges. First, health care data is inherently heterogeneous, with a variety of missing values and from multiple data sources. We propose variations of Gated Recurrent Unit (GRU) to explore and utilize the informative missingness in health care data, and hierarchical multimodal deep models to utilize the relations between different data sources. Second, model interpretability is not only important but necessary for care providers and clinical experts. We introduce a simple yet effective knowledge distillation approach called interpretable mimic learning to learn interpretable gradient boosting tree models while mimicking the performance of deep learning models.
Biography: Zhengping Che is a third year PhD candidate in the Computer Science Department at the University of Southern California, advised by Professor Yan Liu. Before that, he received his bachelor degree in Computer Science from Pilot CS Class (Yao Class) at Tsinghua University, China. His primary research interest lies in the area of deep learning and its applications in health care domain, especially on multivariate time series data.
Host: Xing Shi and Kevin Knight
More Info: http://nlg.isi.edu/nl-seminar/
Location: Information Science Institute (ISI) - 11th Flr Conf Rm # 1135, Marina Del Rey
Audiences: Everyone Is Invited
Contact: Peter Zamar
Event Link: http://nlg.isi.edu/nl-seminar/
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor.
