The Workshop on Computational Linguistics and Clinical Psychology
CLPsych 2021, held in conjunction with NAACL 2021
June 11, 2021
In the era of COVID-19, when many people are experiencing unprecedented pressure — due to financial strain, health concerns, struggles with online work and education, and social isolation — experts report serious concerns about mental health. CLPsych has an important role to play in bringing people together to discuss and exchange their recent work and results. Together, we hope to be able to advance the common goals of using human language as a tool to better understand emotional and mental states, and reducing emotional suffering and the potential for self-harm.
This year, like the NAACL conference, the workshop is being held entirely remotely. It’s unfortunate that the community won’t be getting together in person, but a silver lining is that this will expand the reach of the workshop to people who might not typically attend in person.
In addition to great accepted papers, we are delighted to have a superb slate of invited speakers and panelists.
Keynote talks will be given by Dr. Munmun De Choudhury (Georgia Tech) and Dr. Matthew Nock (Harvard), with additional invited talks by Dr. Glen Coppersmith (Qntfy), Dr. Carol Espy-Wilson (University of Maryland), and Dr. Lyle Ungar (University of Pennsylvania). A panel discussion with all five speakers will be moderated by Dr. Lorenzo Norris, who is Medical Director of Psychiatric & Behavioral Services at George Washington University Hospital and host of MDedge Psychcast, the official psychiatry podcast of the MDedge Network.
Register for the Workshop
Student Registration Awards
Students can apply for financial support to cover or offset their cost of registration. The deadline for these applications has been extended to 5pm U.S. Central time on June 9, 2021.
Note: all times are Mexico City (U.S. Central) timeTime in Central Time:
Welcome and Keynotes
Keynote speaker session introduction
Keynote and Q&A
Introduction: Nazli Goharian
Keynote and Q&A
Dr. Amanda Purnell
Paper session 1
Session chair: Andrew Yates
Understanding who uses Reddit: Profiling individuals with a self-reported bipolardisorder diagnosis
On the State of Social Media Data for Mental Health Research
Individual Differences in the Movement-Mood Relationship in Digital Life Data
Commentary and discussion
Lunch / Zoom poster session
Intro to poster session: Molly Ireland
Breakout A: Social media analysis
Discussant: Dr. April Foreman
Qualitative Analysis of Depression Models by Demographics
Understanding Patterns of Anorexia Manifestations in Social Media Data with Deep Learning
Automatic Detection and Prediction of Psychiatric Hospitalizations From Social
Towards Understanding the Role of Gender in Deploying Social Media-Based Mental Health Surveillance Models
Breakout B: Speech/language analysis
Discussants: Molly Ireland and Steven Bedrick
Automated coherence measures fail to index thought disorder in individuals at risk for psychosis
Safeguarding against spurious AI-based predictions: The case of automated verbal
Towards the Development of Speech-Based Measures of Stress Response in Individuals
Evaluating Automatic Speech Recognition Quality and Its Impact on Counselor Utterance Coding
Paper session 2
Session chair: Molly Ireland
Dissociating Semantic and Phonemic Search Strategies in the Phonemic Verbal Fluency Task in early Dementia
Demonstrating the Reliability of Self-Annotated Emotion Data
Hebrew Psychological Lexicons
Commentary and discussion
Invited talks and panel discussion
Session chair: Molly Ireland
Invited talk 1
Invited talk 2
Session chair: Philip Resnik
Community-level Research on Suicidality Prediction in a Secure Environment: Overview of the CLPsych 2021 Shared Task
Determining a Person’s Suicide Risk by Voting on the Short-Term History of Tweets for the CLPsych 2021 Shared Task
Using Psychologically-Informed Priors for Suicide Prediction in the CLPsych 2021 Shared Task
Team 9: A Comparison of Simple vs. Complex Models for Suicide Risk Assessment
Suicide Risk Prediction by Tracking Self-Harm Aspects in Tweets: NUS-IDS at the
Learning Models for Suicide Prediction from Social Media Posts
Break to make cocktails/obtain a drink
Zoom cocktail poster session
Intro to poster session: Molly Ireland
Breakout C: Therapeutic/Counseling Focused
Discussant: Dr. Kate Neiderhoffer
Detecting Cognitive Distortions from Patient-Therapist Interactions
Towards Low-Resource Real-Time Assessment of Empathy in Counselling
Analysis of Behavior Classification in Motivational Interviewing
Automatic Identification of Ruptures in Transcribed Psychotherapy Sessions
Breakout D: Shared Task Q&A and Discussion
Shared Task papers from 4pm session above
2021 Organizing Committee
Call for Papers
A key goal of this workshop is to foster the conversation with clinicians and clinical researchers, both at the workshop and when these papers are read in the future. We therefore include practicing clinicians and clinical researchers on our program committee, and the ability to communicate ideas, approaches, and results clearly to people who are not computational linguists will be as important as the technical quality of the work.
Submissions will be done using softconf, and detailed information will appear at clpsych.org. We will be taking submissions of both long papers (up to eight pages of core content) and short papers (up to four pages of core content); plus unlimited references. Up to an extra page can be added in the final camera-ready version to allow space for addressing the reviewers’ comments. We strongly encourage all authors to include relevant discussions of ethical considerations and impact in the body of the paper.
Authors may optionally include appendices, but these constitute additional information and might not be looked at by reviewers. If anything in the appendix is an important part of the contribution, or important for the reviewers to assess the work, they should be a part of the main paper, and not appear in the appendix.
All paper submissions must describe substantial, original, completed, and unpublished work. In addition to papers describing algorithms, models, or experimentation, we are happy to receive carefully argued and supported position papers, insightful reviews or synthesis of relevant literature, or informative descriptions of real-world experiences deploying language technology (including prototypes) in relevant clinical settings (Topics include, but are not limited to, the ones listed in the Intent To Submit form — see under Important Dates below) .
All submissions must be fully anonymized to preserve the double-blind reviewing policy. Insufficiently anonymized submissions will be considered for desk-reject.
Authors should adhere to the policies under Submission Types & Requirements in the NAACL 2021 final call for papers at https://2021.naacl.org/calls/papers/ in terms of author guidelines, double submission, anonymity period calculation, double blind review, data management, human subjects discussion, and referencing prior work.
Specifically regarding dual submissions, papers submitted to this workshop must not have been accepted for publication elsewhere or be under review for another workshop, conference or journal.
Paper formatting information, including Word and LaTeX templates, can be found at https://acl-org.github.io/ACLPUB/formatting.html. In order for the paper to appear in the proceedings, at least one author must register for the workshop by the early registration deadline.
Authors of accepted papers will be asked to provide a pre-recorded video presenting their work and to participate in discussant-led breakout sessions during the workshop, in which their paper will be discussed.
This year, we are adopting the theme “Improving Access” to encourage submission and discussion of work that has the potential to help identify and serve people at risk, to create stronger links between patients and providers, and to address the needs of underrepresented communities. We particularly welcome papers that seek to understand people who are difficult to reach, who are traditionally less likely to seek and receive help, or who may be socially or digitally excluded because of conventional measurement/diagnosis or care models.
Contact organizers at: firstname.lastname@example.org
Nazli Goharian, Georgetown University (co-chair)
Philip Resnik, University of Maryland (co-chair)
Andrew Yates, Max Planck Institute for Informatics
Molly Ireland, Texas Tech University
Kate Niederhoffer, Knowable Research
Rebecca Resnik, Rebecca Resnik and Associates, LLC
Shared Task Organizers
Sean MacAvaney, Georgetown University & University of Glasgow
Anjali Mittu, University of Maryland
Philip Resnik, University of Maryland
February 15, 2021: Authors are encouraged to send in an optional “intent to submit” at https://forms.gle/bfF6GUFBMz2hH2eT7 by this date, to help in planning. Although this date has passed, you are still welcome to send an “intent to submit” if you’d like; the more information we have, the better. February 18, 2021: Anonymity period begins March 15, 2021 Extended to March 18, 2021 at 11:59 pm UTC-12: Workshop paper submissions are due at https://www.softconf.com/naacl2021/clpsych2021/ April 15, 2021: Notification of acceptance April 23, 2021: Camera-ready papers due
- June 11, 2021: Remote workshop was held as part of NAACL
Announcement of the shared task was delayed but we have put together an activity that is worth the wait! We have created an opportunity for secure and ethical access to sensitive data in order to work as a community on the problem of predicting suicide risk from social media, and the dataset for the task includes de-identified Twitter posts and ground-truth outcomes from individuals who have attempted or succeeded in a suicide attempt, along with control individuals who have not. These data were donated for research purposes on Qntfy’s OurDataHelps platform. Teams participating in the shared task will do their experimentation on the UMD/NORC Mental Health Data Enclave, a secure computing environment that brings researchers to the data rather than vice-versa. Please see the shared task page for details on registration and participation.
Nick Allen, University of Oregon
Steven Bedrick, Oregon Health & Science University
Arman Cohan, Allen Institute for AI
Glen Coppersmith, Qntfy
Tyler Davis, Texas Tech University
Bart Desmet, National Institutes of Health
Samuel Dooley, UMD
Ophir Frieder, Georgetown University
Pranav Goel, University of Maryland
David Hancock, Weill Cornell Medicine
Kristy Hollingshead, IHMC
Loring Ingraham, George Washington University
Andrew Littlefield, Texas Tech University
Sean MacAvaney, IR Lab, Georgetown University
Adam Miner, Stanford University School of Medicine
Taleen Nalabandian, Texas Tech University
Yaakov Ophir, Technion – Israel Institute of Technology
Ted Pedersen, University of Minnesota, Duluth
Daniel Preotiuc-Pietro, Bloomberg LP
Emily Prud’hommeaux, Boston College
Masoud Rouhizadeh, Johns Hopkins University
Jonathan Schler, HIT
H. Andrew Schwartz, Stony Brook University
Han-Chin Shing, University of Maryland at College Park
Luca Soldaini, Amazon
Amelia Talley, Texas Tech University
Jason Van Allen, Texas Tech University
Sarah Wayland, Guiding Exceptional Parents, LLC
Eugene Yang, Georgetown University
Jessica Yu, Livongo
Ayah Zirikly, Johns Hopkins University