• The workshop will take place in person at 502 - Cowlitz and virtually via Zoom and Discord.

  • Zoom link: on the workshop page via Underline (available after registering).

  • Discord link (if you get an error trying to accept the invite, try logging into Discord first).

• Keynotes will be held in person and streamed on Zoom; Q&A's will take place via Discord text channels.

• Papers will be presented via hybrid panels (in person and on Zoom); discussion will take place via Discord text channels.

  • • Each paper will have 5 minutes to present (in the order listed below), and then there will be a panel discussion at the end.

• Breakout discussions will be held in person and via Discord voice channel.

• For posting or following on Twitter, please use the hashtag #hcinlp2022.

08:30am - 10:00am:

• Welcome

• 08:30-9:30am: Keynote #1: Diyi Yang (in-person and Zoom, Q&A via Discord text channels)

Human-Centered Design and Evaluation for Natural Language Processing

Abstract: Recent advances in natural language processing especially around big models have enabled extensive successful applications. However, there are a growing amount of evidences and concerns towards the negative aspects of NLP systems such as biases and the lack of input from users. How can we build NLP systems that are more user centric and be more aware of human factors?  In this talk, we will present two case studies on how human-centered design and evaluation can be leveraged to build responsible NLP applications. The first one utilizes participatory design to construct a corpus for African American Vernacular English to study dialect disparity. The second part presents an interactive system with various user studies for visualizing models’ toxic predictions, while providing alternative suggestions for flagged toxic language. 

Bio: Diyi Yang is an assistant professor in the School of Interactive Computing at Georgia Tech. She received her PhD from Language Technologies Institute at Carnegie Mellon University in 2019.  Her research interests are computational social science and natural language processing. Her research goal is to understand the social aspects of language and to  build socially aware NLP systems to better support human-human and human-computer interaction. Her work has received multiple best paper nominations or awards at ACL, ICWSM, EMNLP, SIGCHI, and CSCW.  She is a recipient of Forbes 30 under 30 in Science (2020),  IEEE “AI 10 to Watch” (2020), the Intel Rising Star Faculty Award (2021),  Microsoft Research Faculty Fellowship (2021),  and NSF CAREER Award (2022). 

• 9:30-10:00am: Paper panel #1: HCI+NLP and human learning (in-person and Zoom, discussions via Discord text channels)

  • • Towards Automated Generation and Evaluation of Questions in Educational Domains

Huy Anh Nguyen, Shravya Angri Bhat, Steven Moore, and John Stamper

• • • Can Explanations Help Users with Text Comprehension?

Joon Sik Kim, Valerie Chen, Nihar B. Shah, and Ameet Talwalkar

• • • Augmenting Scientific Creativity with Retrieval across Knowledge Domains

Hyeonsu B. Kang, Sheshera Mysore, Kevin J. Huang, Haw-Shiuan Chang, Thorben Prein, Andrew McCallum, Niki Kittur, and Elsa Olivetti

10:00am - 10:30am: Break

10:30am - 12:00pm:

• 10:30-11:20am: Paper panel #2: Human-in-the-loop (in-person and Zoom, discussions via Discord text channels)

  • • Is a Question Decomposition Unit All We Need?

Pruthvi Patel, Swaroop Mishra, Mihir Parmar, and Chitta Baral

• • • Taxonomy Builder: a Data-driven and User-centric Tool for Streamlining Taxonomy Construction

Mihai Surdeanu, John Hungerford, Yee Seng Chan, Jessica MacBride, Benjamin Gyori, Andrew Lee Zupon, Zheng Tang, Haoling Qiu, Bonan Min, Yan Zverev, Caitlin Hilverman, Max Thomas, Walter Andrews, Keith Alcock, Zeyu Zhang, Michael Reynolds, Steven Bethard, Rebecca Sharp, and Egoitz Laparra

• • • An Interactive Exploratory Tool for the Task of Hate Speech Detection

Angelina McMillan-Major, Amandalynne Paullada, and Yacine Jernite

• • • Putting Humans in the Image Captioning Loop

Aliki Anagnostopoulou, Mareike Hartmann, and Daniel Sonntag

• • • Teaching Interactively to Learn Emotions in Natural Language

Rajesh Titung and Cecilia Alm

• 11:20-12:00pm: Breakout session #1 (in-person and via Discord voice channels)

  • • Table 1: Human-centered problem formulation / task definition

    • Table 2: Human-centered corpus development 

      • (e.g., dataset definition, selection, annotation, including human-in-the-loop paradigms)

    • Table 3: Human-centered evaluations 

      • (e.g., of the usability, usefulness, and other features of systems, including qualitative and quantitative evaluations and human-in-the-loop evaluation pipelines)

12:00pm - 1:30pm: Lunch

1:30pm - 3:00pm:

• 1:30-2:00pm: Paper panel #3: Language in specific application/sociocultural contexts (in-person and Zoom, discussions via Discord text channels)

  • • Human-centered computing in legal NLP - An application to refugee status determination

Claire Barale

• • • Towards a Deep Multi-layered Dialectal Language Analysis: A Case Study of African-American English

Jamell Dacon

• • • Let's Chat: Understanding User Expectations in Socialbot Interactions

Elizabeth Soper, Erin Pacquetet, Sougata Saha, Souvik Das, and Rohini Srihari

• 2:00-2:40pm: Paper panel #4: HCI+NLP in health domains (in-person and Zoom, discussions via Discord text channels)

  • • Design Considerations for an NLP-Driven Empathy and Emotion Interface for Clinician Training via Telemedicine

Roxana Girju and Marina Girju

• • • Evaluating the Quality of Machine Translations in Medical Settings

Nikita Mehandru, Sweta Agrawal, Niloufar Salehi, and Marine Carpuat

• • • Leveraging Pre-Trained Language Models to Streamline Natural Language Interaction for Self-Tracking

Young-Ho Kim, Sungdong Kim, Minsuk Chang, and Sang-Woo Lee

• • • Cross-lingual German Biomedical Information Extraction: from Zero-shot to Human-in-the-Loop

Siting Liang, Mareike Hartmann, and Daniel Sonntag

• 2:40-3:00pm: Paper panel #5: Tensions and insights at the intersection of HCI+NLP (in-person and Zoom, discussions via Discord text channels)

  • • Human-Centric Research for NLP: Towards a Definition and Guiding Questions

Bhushan Kotnis, Kiril Gashteovski, Julia Gastinger, Giuseppe Serra, Francesco Alesiani, Timo Sztyler, Ammar Shaker, Na Gong, Carolin Lawrence, and Zhao Xu

• • • Narrative Datasets through the Lenses of NLP and HCI

Sharifa Sultana, Renwen Zhang, Hajin Lim, and Maria Antoniak

3:00pm - 3:30pm: Break

3:30pm - 5:00pm

• 3:30-4:30pm: Keynote #2: Jeffrey Bigham (in-person and Zoom, Q&A via Discord text channels)

An HCI Approach to Dialog Systems Research

Abstract: Over the past couple of decades, my group and I have been developing dialog systems that center humans in their creation, operation, and use. These systems have drawn on a breadth of HCI methods, from computer science and design, to the behavioral sciences and psychology. In this talk, I’ll use examples of my work in dialog, along with a few cherry-picked examples from others, to illustrate how different aspects of HCI can influence NLP, and highlight some areas that seem especially fruitful for future work.

Bio: Jeffrey P. Bigham is an Associate Professor in the Human-Computer Interaction and Language Technologies Institutes in the School of Computer Science at Carnegie Mellon University. His research combines computation and crowds to make novel deployable interactive systems, and ultimately solve hard problems in computer science. These systems combine machine learning and real-time crowdsourcing in domains like (i) access technology, (ii) interactive dialog systems, and (iii) support for crowd/gig workers. Much of my work focuses on accessibility because I see the field as a window into the future, given that people with disabilities are often the earliest adopters of AI.

• 4:30-4:55pm: Breakout discussion #2 (in-person and via Discord voice channels)

  • • Table 1: Critical perspectives, including fairness, ethics, inclusivity

    • Table 2: Fostering more interdisciplinary HCI / NLP collaboration

    • Table 3: Future directions for HCI+NLP

• 4:55-5:00pm: Closing