| { |
| "paper_id": "2022", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T05:58:14.471195Z" |
| }, |
| "title": "Keynote Talk: Invited Talk 1", |
| "authors": [ |
| { |
| "first": "Mirella", |
| "middle": [], |
| "last": "Lapata", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| }, |
| { |
| "first": "Nina", |
| "middle": [], |
| "last": "Balcan", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| }, |
| { |
| "first": "Rada", |
| "middle": [], |
| "last": "Mihalcea", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| }, |
| { |
| "first": "Carlos", |
| "middle": [], |
| "last": "Guestrin", |
| "suffix": "", |
| "affiliation": {}, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "Invited Talk at the 2nd WIT: Workshop On Deriving Insights From User-Generated Text at ACL2022 Bio: Mirella Lapata is a professor in the School of Informatics at the University of Edinburgh. I'm affiliated with the Institute for Communicating and Collaborative Systems and the Edinburgh Natural Language Processing Group. Her research focuses on computational models for the representation, extraction, and generation of semantic information from structured and unstructured data, involving text and other modalities such as images, video, and large scale knowledge bases. I have worked on a variety of applied NLP tasks such as semantic parsing and semantic role labeling, discourse coherence, summarization, text simplification, concept-to-text generation, and question answering. I have also used computational models (drawing mainly on probabilistic generative models) to explore aspects of human cognition such as learning concepts, judging similarity, forming perceptual representations, and learning word meanings. The overarching goal of my research is to enable computers to understand requests and act on them, process and aggregate large amounts of data, and convey information based on them. Critical for all these tasks are models for extracting and representing meaning from natural language text, storing meanings internally, and working with stored meanings to derive further consequences. vii", |
| "pdf_parse": { |
| "paper_id": "2022", |
| "_pdf_hash": "", |
| "abstract": [ |
| { |
| "text": "Invited Talk at the 2nd WIT: Workshop On Deriving Insights From User-Generated Text at ACL2022 Bio: Mirella Lapata is a professor in the School of Informatics at the University of Edinburgh. I'm affiliated with the Institute for Communicating and Collaborative Systems and the Edinburgh Natural Language Processing Group. Her research focuses on computational models for the representation, extraction, and generation of semantic information from structured and unstructured data, involving text and other modalities such as images, video, and large scale knowledge bases. I have worked on a variety of applied NLP tasks such as semantic parsing and semantic role labeling, discourse coherence, summarization, text simplification, concept-to-text generation, and question answering. I have also used computational models (drawing mainly on probabilistic generative models) to explore aspects of human cognition such as learning concepts, judging similarity, forming perceptual representations, and learning word meanings. The overarching goal of my research is to enable computers to understand requests and act on them, process and aggregate large amounts of data, and convey information based on them. Critical for all these tasks are models for extracting and representing meaning from natural language text, storing meanings internally, and working with stored meanings to derive further consequences. vii", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Abstract", |
| "sec_num": null |
| } |
| ], |
| "body_text": [ |
| { |
| "text": "Welcome to the 2nd WIT (Workshop On Deriving Insights From User-Generated Text)! Recent advances in Conversational AI, Natural Language Processing, Natural Language Understanding, Language Generation, Machine Learning, Deep Learning, Knowledge Bases, and others, have demonstrated promising results and far-reaching uses of text. Such results can be seen in many different tasks including, but not limited to better extractions from user-generated content, better language models, new approaches related to (commonsense) knowledge-bases, knowledge graphs, better information seeking QA (or Dialogue) systems, etc. Classical data management problems such as data cleaning/integration and search may also benefit from these new approaches. The WIT workshop series was started to provide a venue to exploit and explore the use of advanced AI/ML/NLP techniques on user-generated text, which is rich in user insights and experiences. Therefore, the goal of this workshop series is to bring together researchers interested in the development and the application of novel approaches/models/systems to address challenges around harnessing text-heavy user-generated data that is available to organizations and over the Web. For this 2nd edition, the workshop will have a great line-up of invited speakers (Mirella Lapata -University of Edinburgh, Rada Mihalcea -University of Michigan, Ann Arbor, Nina Balcan -Carnegie Mellon University, Carlos Guestrin -Stanford University) as well oral (and poster) presentations of contributed research papers. Following the tradition started in the 1st WIT, the 2nd WIT will host a panel of experts from the academia and industry to discuss and share their experiences and challenges faced in deriving insights from user-generated text. The panel is tentatively titled \"User generated content and deep learning: Sorting out 'the good, the bad, and the ugly\"' and is intended to highlight and surface the effects of training data on downstream applications and whether or not organizations prepare efforts around removing biases in data that they use for training or other purposes. We would like to congratulate the authors of accepted papers, as well as to thank all the authors of submitted papers, members of the Program Committee and all the ACL main conference organization team. 2nd WIT Organizing Committee", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Introduction", |
| "sec_num": null |
| }, |
| { |
| "text": "Abstract: Invited Talk at the 2nd WIT: Workshop On Deriving Insights From User-Generated Text at ACL2022 Bio: Maria-Florina (Nina) Balcan is the Cadence Design Systems Professor of Computer Science at the School of Computer Science (MLD and CSD) at Carnegie Mellon University, she is also Sloan Fellow and Microsoft Faculty Fellow. Nina's main research interests are in machine learning, artificial intelligence, and theoretical computer science. Current research focus includes developing foundations and principled, practical algorithms for important modern learning paradigms. These include interactive learning, distributed learning, learning representations, life-long learning, and metalearning. Her research addresses important challenges of these settings, including statistical efficiency, computational efficiency, noise tolerance, limited supervision or interaction, privacy, low communication, and incentives. Other research topics are i) Foundations and applications of data driven algorithm design. Design and analysis of algorithms on realistic instances (a.k.a. beyond worst case).; ii) Computational and data-driven approaches in game theory and economics; iii) computational, learning theoretic, and game theoretic aspects of multi-agent systems, and iv) Analyzing the overall behavior of complex systems in which multiple agents with limited information are adapting their behavior based on past experience, both in social and engineered systems contexts.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Nina Balcan Carnegie Mellon University", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [], |
| "bib_entries": { |
| "BIBREF0": { |
| "ref_id": "b0", |
| "title": "Unsupervised Abstractive Dialogue Summarization with Word Graphs and POV Conversion Seongmin Park and", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Unsupervised Abstractive Dialogue Summarization with Word Graphs and POV Conversion Seongmin Park and Jihwa Lee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1", |
| "links": null |
| }, |
| "BIBREF1": { |
| "ref_id": "b1", |
| "title": "An Interactive Analysis of User-reported Long COVID Symptoms using Twitter Data Lin Miao, Mark Last and Marina Litvak", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "An Interactive Analysis of User-reported Long COVID Symptoms using Twitter Data Lin Miao, Mark Last and Marina Litvak . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10", |
| "links": null |
| }, |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "Bi-Directional Recurrent Neural Ordinary Differential Equations for Social Media Text Classification Maunika Tamire, Srinivas Anumasa and", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bi-Directional Recurrent Neural Ordinary Differential Equations for Social Media Text Classification Maunika Tamire, Srinivas Anumasa and P. K. Srijith . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20", |
| "links": null |
| }, |
| "BIBREF4": { |
| "ref_id": "b4", |
| "title": "An Interactive Analysis of User-reported Long COVID Symptoms using Twitter Data Lin Miao, Mark Last and Marina Litvak Bi-Directional Recurrent Neural Ordinary Differential Equations for Social Media Text Classification Maunika Tamire", |
| "authors": [], |
| "year": null, |
| "venue": "Srinivas Anumasa and P. K. Srijith", |
| "volume": "12", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "An Interactive Analysis of User-reported Long COVID Symptoms using Twitter Data Lin Miao, Mark Last and Marina Litvak Bi-Directional Recurrent Neural Ordinary Differential Equations for Social Media Text Classification Maunika Tamire, Srinivas Anumasa and P. K. Srijith 12:30 -11:30 Invited Talk II 12:30 -14:00 Lunch Break 15:00 -14:00 Invited Talk III 15:00 -15:30 Coffee Break 15:30 -15:45 Session 2", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "Unsupervised Abstractive Dialogue Summarization with Word Graphs and POV Conversion Seongmin Park and Jihwa Lee", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Unsupervised Abstractive Dialogue Summarization with Word Graphs and POV Conversion Seongmin Park and Jihwa Lee 15:45 -16:45 Invited Talk IV 16:45 -17:45 Panel xii", |
| "links": null |
| } |
| }, |
| "ref_entries": {} |
| } |
| } |
