Conferences, Lectures, & Seminars
Events for September
-
NL Seminar-More than the sum of their parts: Translating idioms without destroying their meaning
Thu, Sep 05, 2019 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Denis Emelin and Prince Wang, USC/ISI
Talk Title: More than the sum of their parts: Translating idioms without destroying their meaning
Series: Natural Language Seminar
Abstract: Translating idioms is hard. As low-frequency linguistic events with a non compositional meaning, idiomatic expressions are at odds with contemporary neural machine translation methods. Accordingly, the literal translation of idiomatic phrases which fails to preserve their semantic content represents an often observed failure case in NMT models. To facilitate future work on idiom translation, the current project sets out to compile a large-coverage, multilingual corpus of parallel sentences containing idiomatic expressions, augmented with their respective monolingual definitions. With this resource in hand, we next aim to propose models which can effectively exploit idiom definitions to avoid literal translation errors. As part of the evaluation of the constructed corpus, we demonstrate that idioms continue to pose a veritable challenge for state of the art NMT models.
Biography: Denis is a second-year PhD candidate at the University of Edinburgh, advised by Dr. Rico Sennrich. His background is in machine translation, natural language understanding, and linguistics.
Host: Emily Sheng
More Info: https://nlg.isi.edu/nl-seminar
Webcast: https://bluejeans.com/s/8Lu7w/Location: Information Science Institute (ISI) - CR #689
WebCast Link: https://bluejeans.com/s/8Lu7w/
Audiences: Everyone Is Invited
Contact: Peter Zamar
Event Link: https://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- Allen NLP Tools Workshop
Thu, Sep 19, 2019 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Seraphina Goldfarb-Tarrant, USC/ISI
Talk Title: AllenNLP Tools Workshop
Series: Natural Language Seminar
Abstract: This is a practical talk that highlights some of the areas where AllenNLP the NLP research library excels, and gives a look at new features being released. It will focus on the ways that use of the library can enable reproducibility, interpretability, and visualizations.
Biography: Seraphina Goldfarb-Tarrant is a Research Programmer at ISI, doing work in NLG. She finished her Master's at the University of Washington, and is beginning her PhD at the University of Edinburgh.
Host: Emily Sheng
More Info: https://nlg.isi.edu/nl-seminar
Webcast: https://bluejeans.com/s/OUQy4/Location: Information Science Institute (ISI) - CR #689
WebCast Link: https://bluejeans.com/s/OUQy4/
Audiences: Everyone Is Invited
Contact: Peter Zamar
Event Link: https://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. -
Center for Knowledge-Driven Interdisciplinary Data Science (CKIDS)
Mon, Sep 23, 2019 @ 04:00 PM - 05:00 PM
Thomas Lord Department of Computer Science, Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Odd Erik Gundersen, Adjunct Associate Professor, Norwegian University of Science and Technology
Talk Title: Reproducibility in AI: Standing on the Feet of Giants
Series: Invited Lecture Series
Abstract: First, we need a common understanding of what reproducibility is. Then, I will talk about some of the challenges we face related to reproducing empirical AI research and give some examples of studies that have tried to reproduce results from AI and machine learning. Having this understanding we can identify what we need to do to improve the reproducibility of our own experiments.
Biography: Odd Erik Gundersen is an adjunct associate professor at the Norwegian University of Science and Technology (NTNU) in Trondheim, Norway, where he teaches courses and supervises master students in AI. He received his PhD from the Norwegian University of Science and Technology. Gundersen has applied AI in the industry, mostly for startups, since 2006. He has conducted several analysis of reproducibility in the artificial intelligence and machine learning literature, and has developed guidelines for reproducibility in data science. Currently, he investigates how AI can be applied in the renewable energy sector and for driver training.
For more information and future speakers, please visit:
https://sites.usc.edu/ckids/events/invited-lecture-series/
Host: Yolanda Gil, Director of Center for Knowledge-Driven Interdisciplinary Data Science (CKIDS)
More Info: https://sites.usc.edu/ckids/events/invited-lecture-series/
Location: Ronald Tutor Hall of Engineering (RTH) - 115
Audiences: Everyone Is Invited
Contact: Alma Nava / Information Sciences Institute
Event Link: https://sites.usc.edu/ckids/events/invited-lecture-series/
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 Question Answering by Reasoning Across Documents with Graph Convolutional Networks
Thu, Sep 26, 2019 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Nicola De Cao, University of Amsterdam
Talk Title: Question Answering by Reasoning Across Documents with Graph Convolutional Networks
Series: Natural Language Seminar
Abstract: Most research in reading comprehension has focused on answering questions based on individual documents or even single paragraphs. We introduce a neural model which integrates and reasons relying on information spread within documents and across multiple documents. We frame it as an inference problem on a graph. Mentions of entities are nodes of this graph while edges encode relations between different mentions e.g. within and cross document co reference. Graph convolutional networks GCNs are applied to these graphs and trained to perform multi-step reasoning. Our Entity GCN method is scalable and compact, and it achieves state of the art results on a multi-document question answering dataset, WikiHo.
Biography: Nicola is a first year Ph.D. candidate at the Institute for Logic, Language and Computation ILLC at the University of Amsterdam.
He is appointed at the School of Informatics at the University of Edinburgh supervised by Prof Ivan Titov, and he is part of the Edinburgh NLP group. His work focuses on unstructured Machine Reading Comprehension also know as Question Answering.
Host: Emily Sheng
More Info: https://nlg.isi.edu/nl-seminar
Webcast: https://bluejeans.com/s/sgwNF/Location: Information Science Institute (ISI) - CR #689
WebCast Link: https://bluejeans.com/s/sgwNF/
Audiences: Everyone Is Invited
Contact: Peter Zamar
Event Link: https://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.
