Conferences, Lectures, & Seminars
Events for October
-
AI Seminar-The Future is on Your Wrist: Challenges and Opportunities of Wearable Technologies
Fri, Oct 07, 2016 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Luca Foschini, Evidation Health
Talk Title: The Future is on Your Wrist: Challenges and Opportunities of Wearable Technologies
Series: Artificial Intelligence Seminar
Abstract: Wearable technologies have seen a tremendous development in recent years: step and calorie counters have long made their way to our phones and watches, and new consumer-grade sensors can now measure a breadth of physiological functions that until recently could only be found in the monitoring equipment of intensive care units. However, despite the undisputed short-term benefits due to the user increased awareness, quantifying the potential value of wearable technologies in improving longer-term health outcomes remains an open question. In this talk we will present evidence that activity tracking data contains a wealth of information that is predictive of metrics directly related to health outcomes, ranging from medication adherence to lifestyle. To this end, we will show how machine learning tools need to be adapted to take full advantage of densely sampled, multi-variate time series of tracker data. Finally, we will reflect on how the predictive power of wearable data can be harnessed to inform behavior change interventions, and how expertise in computer science, clinical medicine, and behavioral psychology will have to join forces to overcome obstacles in adoption, user engagement, and regulations.
Biography: As Co founder and Head of Data Science at Evidation Health, Luca Foschini PhD is responsible for data analytics, computing, research and development. Dr. Foschini has driven research collaborations with machine learning experts at NYU, behavioral economics departments at Harvard Business School and the Wharton School. Prior to this role, Dr. Foschini worked as R&D at Ask.com and was a visiting scholar at Google Research and ETH Zurich where he developed efficient algorithms for mining spatial data, partitioning large graphs, and detecting traffic anomalies in computer networks. He earned a PhD in Computer Science from UC Santa Barbara focusing on traffic analysis in computer and road networks. He has published numerous papers in the broader area of computer science and he co-authored several patents in information clustering and behavior phenotyping. Dr. Foschini is an alumnus of the Sant'Anna School of Pisa, Italy
Host: Emilio Ferrara
Location: Information Science Institute (ISI) - 11th Flr Conf Rm # 1135, Marina Del Rey
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. -
Tensor Decomposition Techniques for analysing time-varying networks
Tue, Oct 11, 2016 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Anna Sapienza, PhD in Applied Mathematics at the Polytechnic Univ. of Turin, working in the Data Science Lab at ISI Foundation, Turin, Italy
Series: Recruitng Seminar
Abstract: The increasing availability of high-dimensional data calls for new methods to extract meaningful information, such as groups of data correlations (i.e. communities, clusters) or unusual and unexpected data records (i.e. anomalies, outliers). Time-varying networks are particularly suitable objects to summarize a large amount of data into interpretable representations and are used to describe a great variety of complex systems. A fundamental challenge is to define models and tools that are able to capture and disentangle the structural and temporal properties from the time-varying networks and reproduce the observed features on dynamical processes occurring over the network, such as information diffusion, event cascades or disease spreading. Thus, the purpose of my Ph.D work is twofold: to extract the structural and temporal properties of time-varying networks to face problems as pattern detection and missing data recovery, and to analyze the interplay between these characteristics and dynamical processes.
Biography: Anna Sapienza is currently a Ph.D candidate at the Polytechnic University of Turin, she is completing the third year of her Ph.D studies. Her work was developed in the Data Science group at the I.S.I. Foundation of Turin under the supervision of Dr. Ciro Cattuto and Dr. Laetitia Gauvin. Her research interests stay at the intersection between computational social science, machine learning, and network analysis. Recently her work focused on the development of mathematical frameworks and tools for tensor factorization techniques and their applications for studying high-dimensional data.
Host: Emilio Ferrara and Kristina Lerman
Webcast: http://webcastermshd.isi.edu/Mediasite/Play/e7f614b9cffc415db4015dd86999db5f1dLocation: Information Science Institute (ISI) - 1135 - 11th fl Large CR
WebCast Link: http://webcastermshd.isi.edu/Mediasite/Play/e7f614b9cffc415db4015dd86999db5f1d
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-Matrix Completion, Saddlepoints, and Gradient Descent
Fri, Oct 14, 2016 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Jason Lee , USC
Talk Title: Matrix Completion, Saddlepoints, and Gradient Descent
Series: Artificial Intelligence Seminar
Abstract: Matrix completion is a fundamental machine learning problem with wide applications in collaborative filtering and recommender systems. Typically, matrix completion are solved by non-convex optimization procedures, which are empirically extremely successful. We prove that the symmetric matrix completion problem has no spurious local minima, meaning all local minima are also global. Thus the matrix completion objective has only saddlepoints an global minima.
Next, we show that saddlepoints are easy to avoid for even Gradient Descent -- arguably the simplest optimization procedure. We prove that with probability 1, randomly initialized Gradient Descent converges to a local minimizer.
Biography: Jason Lee is an assistant professor in Data Sciences and Operations at the University of Southern California. Prior to that, he was a postdoctoral researcher at UC Berkeley working with Michael Jordan. Jason received his PhD at Stanford University advised by Trevor Hastie and Jonathan Taylor. His research interests are in statistics, machine learning, and optimization. Lately, he has worked on high dimensional statistical inference, analysis of non-convex optimization algorithms, and theory for deep learning.
Host: Emilio Ferrara
Location: Information Science Institute (ISI) - 6th Floor -CR # 689; ISI-Marina del Rey
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. -
EMNLP PRACTICE TALK: UNDERSTANDING NEURAL MACHINE TRANSLATION: LENGTH CONTROL AND SYNTACTIC STRUCTURE
Fri, Oct 14, 2016 @ 03:00 PM - 04:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Xing Shi, USC/ISI
Talk Title: EMNLP PRACTICE TALK: UNDERSTANDING NEURAL MACHINE TRANSLATION: LENGTH CONTROL AND SYNTACTIC STRUCTURE
Series: Natural Language Seminar
Abstract: Neural Machine Translation is powerful but we know little about the black box. We conduct the following two investigations to gain a better understanding: First, we investigate how neural, encoder-decoder translation systems output target strings of appropriate lengths, finding that a collection of hidden units learns to explicitly implement this functionality. Second, we investigate whether a neural, encoderdecoder translation system learns syntactic information on the source side as a by-product of training. We propose two methods to detect whether the encoder has learned local and global source syntax. A fine-grained analysis of the syntactic structure learned by the encoder reveals which kinds of syntax are learned and which are missing.
Biography: Xing Shi is a PhD student at ISI working with Prof. Kevin Knight.
Host: Kevin Knight
More Info: http://nlg.isi.edu/nl-seminar/
Location: Information Science Institute (ISI) - 6th Floor -CR # 689; ISI-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
Fri, Oct 21, 2016 @ 03:00 PM - 04:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Marjan Ghazvininejad, USC/ISI
Talk Title: EMNLP PRACTICE TALK. 1. GENERATING TOPICAL POETRY And 2. UNSUPERVISED NEURAL HIDDEN MARKOV MODELS
Series: Natural Language Seminar
Abstract: 1. In this talk I describe Hafez, a program that generates any number of distinct poems on a user-supplied topic. Poems obey rhythmic and rhyme constraints. I describe the poetry-generation algorithm, give experimental data concerning its parameters, and show its generality with respect to language and poetic form. 2. In this work, we present the first results for neuralizing an Unsupervised Hidden Markov Model. We evaluate our approach on tag induction. Our approach outperforms existing generative models and is competitive with the state-of-the-art though with a simpler model easily extended to include additional context.
Biography: Marjan Ghazvininejad is a PhD student at ISI working with Prof. Kevin Knight. Yonatan Bisk is a Postdoc at ISI working with Prof. Daniel Marcu.
Host: Xing Shi and Kevin Knight
More Info: http://nlg.isi.edu/nl-seminar/
Location: Information Science Institute (ISI) - 6th Floor -CR # 689; ISI-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. -
Learning from Zero: Recent Advances in Bootstrapping Semantic Parsers using Crowdsourcing
Fri, Oct 28, 2016 @ 03:00 PM - 04:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Yu Su, UCSB
Talk Title: Learning from Zero: Recent Advances in Bootstrapping Semantic Parsers using Crowdsourcing
Series: Natural Language Seminar
Abstract: Semantic parsing, which parses natural language into formal languages, has been applied to a wide range of structured data like relation databases, knowledge bases, and web tables. To learn a semantic parser for a new domain, the first challenge is always how to collect training data. While data collection using crowdsourcing has become a common practice in NLP, it's a particularly challenging and interesting problem when it comes to semantic parsing, and is still in its early stages. Given a domain and a formal language, how can we generate meaningful logical forms in a configurable way? How to design the annotation task so that crowdsourcing workers, who do not understand formal languages, can handle with ease? How can we exploit the compositional nature of formal languages to optimize the crowdsourcing process? In this talk I will introduce some recent advances in this direction, and present some preliminary answers to the above questions. The covered works mainly concern knowledge bases, but we will also cover some ongoing work concerning web APIs.
Biography: Yu Su is a fifth year PhD candidate in the Computer Science Department at UCSB, advised by Professor Xifeng Yan. Before that, He received a bachelor degree from Tsinghua University in 2012, with a major in Computer Science. He is interested in the interplay between language and formal meaning representations, including problems like semantic parsing, continuous knowledge representation, and natural language generation. He also enjoys applying deep learning on these problems.
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.
