NewSum-EMNLP 2021: Program (9am - 6pm AST, Nov. 10th)

NewSum workshop main website
09:00 - 10:30    Morning Session I ( Zoom Link)
Chair: Fei Liu
Co-Chair: Yue Dong
09:00 - 09:10    Open remarks
NewSum Organizers
09:10 - 10:00    Keynote I - Sashi Narayan (Google)
Learning from Past: Bringing Planning Back to Neural Generators
Traditional NLG systems in Reiter and Dale’s vision were inherently grounded and controllable, thanks to a planning stage which played a crucial role in ordering and structuring the information, and in grounding the generation of text to the plan. Modern neural generation systems have advanced NLG beyond our imagination, yet some of the most desired properties such as grounding and controllability have been lost and are still to be mastered. In this talk, I will discuss why we need to bring back planning to neural generation and to make generation systems more grounded, controllable, inspectable and trustworthy. I will present several pieces of evidence supporting this direction exploring existing work in data-to-text and story generation, and in summarization.
10:00 - 10:10    Sentence-level Planning for Especially Abstractive Summarization
Andreas Marfurt1 and James Henderson2
1Idiap Research Institute and EPFL, 2Idiap Research Institute
File
10:10 - 10:20    Template-aware Attention Model for Earnings Call Report Generation
Yangchen Huang, Seyed Danial Mohseni Taheri, Prashant Dhingra
JP Morgan Chase & Company
File
10:20 - 10:25    Knowledge and Keywords Augmented Abstractive Sentence Summarization
Shuo Guan
NYU Courant
File
10:25 - 10:30    Rewards with Negative Examples for Reinforced Topic-Focused Abstractive Summarization
Khalil Mrini1, Can Liu2, Markus Dreyer2
1University of California, San Diego, 2Amazon.com
File
10:30 - 11:00    Coffee break GatherTown Link
11:00 - 12:00    Morning session II ( Zoom Link)
Chair: Yue Dong
Co-Chair: Jackie Cheung
11:00 - 11:50    Keynote II - Sebastian Gehrmann (Google)
Breaking News: It’s time to fix the evaluation of generated text
Language generation has undergone multiple paradigm shifts from constructed grammars and modular systems toward end-to-end supervised (neural) approaches, and now, almost every system is built on pretrained models. As a result, how generated text looks has changed a lot; it is now much more fluent and most of its issues relate to its content. Yet, we still use the same metrics, some of the same corpora, and how to conduct human evaluations remains a mystery. Throughout this talk, we will explore many examples of broken evaluations in summarization and other generation applications. I will discuss the implications that broken evaluation pipelines have on model development and the overall progress in the field. And I will show some promising results on developing evaluation suites, learned metrics, and meta-evaluations that have the potential to improve how generated text is evaluated.
11:50 - 12:00    A Novel Wikipedia based Dataset for Monolingual and Cross-Lingual Summarization
Mehwish Fatima1 and Michael Strube2
1Heidelberg Institute for Theoretical Studies (HITS gGmbH), 2Heidelberg Institute for Theoretical Studies
File
12:00 - 13:00    Lunch break GatherTown Link
13:00 - 14:30    Afternoon session I ( Zoom Link)
Chair: Lu Wang
Co-Chair: Yue Dong
13:00 - 13:50    Keynote III - Asli Celikyilmaz (Facebook AI Research)
Tune in To Your Language Model for Better Text Generation
With today’s neural language models, we can teach computers to summarize online meetings, write creative stories or articles about an event, hold longer conversations in customer-service applications, chit-chat about daily activities with individuals, describe pictures to visually impaired, to name a few. In this talk, I will discuss challenges and shortcomings of building such systems with the current neural text generation models focusing on issues relating to collecting and annotating training datasets and building new architectures to model the intrinsic structure of conversations. I will present our recent approaches that imbue transformer based neural generators with structural representations by way of implicit memory architectures and latent structural embeddings. I will conclude my talk pointing to avenues for future research.

13:50 - 13:55    Evaluation of Summarization Systems across Gender, Age, and Race
Anna Jørgensen1 and Anders Søgaard2
1University of Amsterdam, 2University of Copenhagen
File
13:55 - 14:00    Evaluation of Abstractive Summarisation Models with Machine Translation in Deliberative Processes
Miguel Arana-Catania1, Rob Procter1, Yulan He1, Maria Liakata2
1University of Warwick, 2Queen Mary University of London
File
14:00 - 14:10    Capturing Speaker Incorrectness: Speaker-Focused Post-Correction for Abstractive Dialogue Summarization
Dongyub Lee1, Jungwoo Lim2, Taesun Whang3, chanhee lee2, Seungwoo Cho4, Mingun Park5, Heuiseok Lim2
1Kakao Corp, 2Korea University, 3Wisenut Inc., 4Kakao Enterprise at South Korea, 5Microsoft
File
14:10 - 14:20    Measuring Similarity of Opinion-bearing Sentences
Wenyi Tay1, Xiuzhen Zhang1, Stephen Wan2, Sarvnaz Karimi2
1RMIT University, 2CSIRO
File
14:20 - 14:30    EASE: Extractive-Abstractive Summarization End-to-End using the Information Bottleneck Principle
Haoran Li1, Arash Einolghozati1, Srinivasan Iyer1, Bhargavi Paranjape2, Yashar Mehdad3, Sonal Gupta1, Marjan Ghazvininejad4
1Facebook, 2University of Washington, 3Facebook AI, 4Facebook AI Research
File
14:30 - 15:00    Coffee break GatherTown Link
15:00 - 15:35    Afternoon session II ( Zoom Link)
Chair: Jackie Cheung
Co-Chair: Giuseppe Carenini
15:00 - 15:10    Context or No Context? A preliminary exploration of human-in-the-loop approach for Incremental Temporal Summarization in meetings
Nicole Beckage, Shachi H Kumar, Saurav Sahay, Ramesh Manuvinakurike
Intel Labs
File
15:10 - 15:20    Are We Summarizing the Right Way? A Survey of Dialogue Summarization Data Sets
Don Tuggener1, Margot Mieskes2, Jan Deriu1, Mark Cieliebak1
1Zurich University of Applied Sciences, 2University of Applied Sciences, Darmstadt
File
15:20 - 15:30    Modeling Endorsement for Multi-Document Abstractive Summarization
Logan Lebanoff1, Bingqing Wang2, Zhe Feng3, Fei Liu4
1Soar Technology, Inc., 2Bosch Research & Technology Center North America, 3Bosch, 4University of Central Florida
File
15:30 - 15:35    SUBSUME: A Dataset for Subjective Summary Extraction from Wikipedia Documents
Nishant Yadav1, Matteo Brucato2, Anna Fariha2, Oscar Youngquist2, Julian Killingback3, Alexandra Meliou4, Peter Haas2
1UMass Amherst, 2University of Massachusetts Amherst, 3The University of Massachusetts Amherst, 4University of Massachusetts, Amherst
File
15:35 - 16:15    EMNLP Finding papers - Summarization ( Zoom Link)
Chair: Jackie Cheung
Co-Chair: Giuseppe Carenini
15:35 - 15:40    Exploring Multitask Learning for Low-Resource Abstractive Summarization
Ahmed Magooda, Diane Litman, Mohamed Elaraby
University of Pittsburgh
File
15:40 - 15:45    Mitigating Data Scarceness through Data Synthesis, Augmentation and Curriculum for Abstractive Summarization
Ahmed Magooda and Diane Litman
University of Pittsburgh
File
15:45 - 15:50    TWEETSUMM - A Dialog Summarization Dataset for Customer Service
Guy Feigenblat , Chulaka Gunasekara , Benjamin Sznajder , Ranit Aaronov, David Konopnicki, Sachindra Joshi
IBM Research
15:50 - 15:55    Convex Aggregation for Opinion Summarization
Hayate Iso1, Xiaolan Wang1, Yoshihiko Suhara1, Stefanos Angelidis2, Wang-Chiew Tan3
1Megagon Labs, 2University of Edinburgh, 3Facebook AI
15:55 - 16:00    "Let Your Characters Tell Their Story'': A Dataset for Character-Centric Narrative Understanding
Faeze Brahman1, Meng Huang2, Oyvind Tafjord3, Chao Zhao4, Mrinmaya Sachan5, Snigdha Chaturvedi6
1UC Santa Cruz, 2University of Chicago, 3AI2, 4University of North Carolina at Chapel Hill, 5ETH Zurich, 6University of North Carolina, Chapel Hill
File
16:00 - 16:05    Retrieval Augmented Code Generation and Summarization
Md Rizwan Parvez1, Wasi Ahmad2, Saikat Chakraborty3, Baishakhi Ray3, Kai-Wei Chang4
1University of California Los Angeles, 2University of California, Los Angeles, 3Columbia University, 4UCLA
File
16:05 - 16:10    MiRANews: Dataset and Benchmarks for Multi-Resource-Assisted News Summarization
Xinnuo Xu1, Ondřej Dušek2, Shashi Narayan3, Verena Rieser1, Ioannis Konstas1
1Heriot-Watt University, 2Charles University, 3Google
File
16:10 - 16:15    Leveraging Pretrained Models for Automatic Summarization of Doctor-Patient Conversations
Longxiang Zhang1, Renato Negrinho2, Arindam Ghosh3, Vasudevan Jagannathan3, Hamid Reza Hassanzadeh4, Thomas Schaaf1, Matthew R. Gormley2
13M | M*Modal, 2Carnegie Mellon University, 33M, 4NLP Researcher at 3M HIS
File
16:15 - 16:45    Coffee break GatherTown Link
16:45 - 18:00    Afternoon session III
Chair: Giuseppe Carenini
Co-Chair: Fei Liu, Jackie Cheung, Lu Wang, Yue Dong
16:45 - 16:55    TLDR9+: A Large Scale Resource for Extreme Summarization of Social Media Posts
Sajad Sotudeh1, Hanieh Deilamsalehy2, Franck Dernoncourt2, Nazli Goharian1
1Georgetown University, 2Adobe Research
File
16:55 - 17:00    A New Dataset and Efficient Baselines for Document-level Text Simplification in German
Annette Rios, Nicolas Spring, Tannon Kew, Marek Kostrzewa, Andreas Säuberli, Mathias Müller, Sarah Ebling
University of Zurich
File
17:00 - 18:00    Mentoring Program ( Zoom Link)