Mixed-Initiative Real-Time Topic Modeling & Visualization for Crisis Counseling
Open accessThe DOI matches its Crossref record; 24 of the 25 references it cites were found in a registry or an index.
Abstract
Text-based counseling and support systems have seen an increasing proliferation in the past decade. We present Fathom, a natural language interface to help crisis counselors on Crisis Text Line, a new 911-like crisis hotline that takes calls via text messaging rather than voice. Text messaging opens up the opportunity for software to read the messages as well as people, and to provide assistance for human counselors who give clients emotional and practical support. Crisis counseling is a tough job that requires dealing with emotionally stressed people in possibly life-critical situations, under time constraints. Fathom is a system that provides topic modeling of calls and graphical visualization of topic distributions, updated in real time. We develop a mixed-initiative paradigm to train coherent topic and word distributions and use them to power real-time visualizations aimed at reducing counselor cognitive overload. We believe Fathom to be the first real-time computational framework to assist in crisis counseling.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“For example, innovations in computer science (e.g., natural language processing and machine learning) are being used to assess language patterns of successful crisis interventions in text-based mental health conversations (18, 35).”
From Key Considerations for Incorporating Conversational AI in Psychotherapy (2019) · cites this paper as [35]
Citations
Cites
Blei, D. M. (2012) · Communications of the ACM · doi:10.1145/2133806.2133826 · Cited by 5,516 · cited 1×
“Modeling the thematic distribution of issues and emotional reactions -Topic models are mixture models [2].”
BackgroundBostock, M., Ogievetsky, V., & Heer, J. (2011) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2011.185 · Cited by 3,165
Wang, P. S., Lane, M., Olfson, M., Pincus, H. A., Wells, K. B., & Kessler, R. C. (2005) · Archives of General Psychiatry · doi:10.1001/archpsyc.62.6.629 · Cited by 2,695 · cited 2×
“Though the distribution of mental disorders is widespread in the population, they manifest severely in a much smaller and concentrated proportion; six percent or one in seventeen Americans are thought to suffer grievously from such disorders [26].”
BackgroundPrinciples of mixed-initiative user interfaces
Horvitz, E. (1999) · Proceedings of the SIGCHI conference on Human factors in computing systems the CHI is the limit - CHI '99 · doi:10.1145/302979.303030 · Cited by 1,490 · cited 1×
“We develop a mixed-initiative [17] probabilistic graphical topic model framework and a visualization framework to give real-time visual summaries to counselors of both ongoing and archived conversations.”
MethodOnline Learning for Latent Dirichlet Allocation
Hoffman, M. D., Blei, D. M., & Bach, F. (2010) · Cited by 1,469
Labeled LDA: a supervised topic model for credit attribution in multi-labeled corpora
Ramage, D., Hall, D., Nallapati, R., & Manning, C. D. (2009) · Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing Volume 1 - EMNLP '09 · doi:10.3115/1699510.1699543 · Cited by 1,190 · cited 1×
“This is achieved by using a labeled LDA model [21] trained over the set of topics derived from stage 1.”
MethodRethinking LDA: Why Priors Matter
Wallach, H., Mimno, D., & McCallum, A. (2009) · neural information processing systems · Cited by 815
TextFlow: Towards Better Understanding of Evolving Topics in Text
Cui, W., Liu, S., Tan, L., Shi, C., Song, Y., Gao, Z., Qu, H., & Tong, X. (2011) · IEEE Transactions on Visualization and Computer Graphics · doi:10.1109/tvcg.2011.239 · Cited by 349 · cited 1×
“TextFlow [6] is a visually innovative visualization technique that does address the temporal element of evolution of topics in a document.”
BackgroundStatistical topic models for multi-label document classification
Rubin, T. N., Chambers, A., Smyth, P., & Steyvers, M. (2012) · Machine Learning · doi:10.1007/s10994-011-5272-5 · Cited by 294 · cited 1×
“They have been shown to work comparably with supervised learning models in their ability to model personal teenage stories of distress [8, 9] and in document classification [22].”
BackgroundIMing, Text Messaging, and Adolescent Social Networks
Alison Bryant, J., Sanders-Jackson, A., & Smallwood, A. M. K. (2006) · Journal of Computer-Mediated Communication · doi:10.1111/j.1083-6101.2006.00028.x · Cited by 293 · cited 1×
“There is also growing evidence that distressed individuals are more comfortable sharing sensitive information using electronic-based media [1].”
BackgroundChaney, A., & Blei, D. (2021) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v6i1.14321 · Cited by 201 · cited 1×
“Several prior projects have provided different graphical views on topic models, and have also provided means for associating topic models or component words back to the original corpus, either in particular examples or in aggregate [4].”
BackgroundImpact of Applied Suicide Intervention Skills Training on the National Suicide Prevention Lifeline
Gould, M. S., Cross, W., Pisani, A. R., Munfakh, J. L., & Kleinman, M. (2013) · Suicide and Life-Threatening Behavior · doi:10.1111/sltb.12049 · Cited by 200 · cited 1×
“…a proliferation of hotlines to help people in crises within the last decade, where a distressed individual experiencing a crisis, or the caller , is connected to a crisis counselor via telephone hotlines [12] and web-based chats [5] for crises ranging from suicide and self-harm to sexual assault.”
BackgroundGould, M. S., Munfakh, J. L. H., Kleinman, M., & Lake, A. M. (2012) · Suicide and Life-Threatening Behavior · doi:10.1111/j.1943-278x.2011.00068.x · Cited by 142 · cited 1×
“While successful psychotherapy is conditioned on a more longitudinal and detailed understanding of an individual’s plight, crisis counseling can help transition a caller to longitudinal psychotherapy [13].”
BackgroundData‐driven decision making for the enterprise: an overview of business intelligence applications
Hedgebeth, D. (2007) · VINE · doi:10.1108/03055720710838498 · Cited by 141 · cited 1×
“Visualization is an important component of many call center management tools, of which Dundas [14] is a typical example.”
BackgroundChristensen, H., Farrer, L., Batterham, P. J., Mackinnon, A., Griffiths, K. M., & Donker, T. (2013) · BMJ Open · doi:10.1136/bmjopen-2013-002886 · Cited by 80 · cited 1×
“or the caller, is connected to a crisis counselor via telephone hotlines [12] and web-based chats [5] for crises ranging from suicide and self-harm to sexual assault.”
BackgroundEducation and Debriefing: Strategies for Preventing Crises in Crisis-Line Volunteers
Kinzel, A., & Nanson, J. (2000) · Crisis · doi:10.1027//0227-5910.21.3.126 · Cited by 77
Modeling of Stylistic Variation in Social Media with Stretchy Patterns
Gianfortoni, P., Adamson, G. D., & Rosé, C. P. (2011) · Cited by 44
Suicide assessment and evaluation strategies: a primer for counselling psychologists
Schwartz, R. C., & Rogers, J. R. (2004) · Counselling Psychology Quarterly · doi:10.1080/09515070410001665712 · Cited by 28 · cited 1×
“Nearly a third of all practicing psychotherapists and counselors are expected to encounter grave crises of their clients at some point in their careers [23].”
BackgroundStacked Generalization Learning to Analyze Teenage Distress
Dinakar, K., Weinstein, E., Lieberman, H. A., & Selman, R. L. (2014) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v8i1.14527 · Cited by 28 · cited 1×
“They have been shown to work comparably with supervised learning models in their ability to model personal teenage stories of distress [8, 9] and in document classi-fication [22].”
BackgroundSelf-citationYou Too?! Mixed-Initiative LDA Story Matching to Help Teens in Distress
Dinakar, K., Jones, B., Lieberman, H. A., Picard, R. W., Rosé, C. P., Thoman, M., & Reichart, R. (2021) · Proceedings of the International AAAI Conference on Web and Social Media · doi:10.1609/icwsm.v6i1.14251 · Cited by 23 · cited 4×
“More details about this approach is fully described in our prior work [8].”
BackgroundMethodInfluentialSelf-citationReal-time Topic Models for Crisis Counseling
Dinakar, K., Chaney, A. J. B., Lieberman, H. A., & Blei, D. M. (2014) · Cited by 12
Self-citationInteraction Mining: the new Frontier of Call Center Analytics.
Pallotta, V., Delmonte, R., Vrieling, L., & Walker, D. F. (2011) · Cited by 9
The Future of Volunteers in Crisis Hotline Work
Pratt, M. E. (2013) · D-Scholarship@Pitt (University of Pittsburgh) · Cited by 6
Hoff, L. A., Brown, L., & Hoff, M. R. (2011) · Routledge eBooks · doi:10.4324/9780203893234 · Cited by 3
Research works printed but not linked: 3
- Gardner , M. J. , Lutes , J. , Lund , J. , Hansen , J. , Walker , D. , Ringger , E. , and Seppi , K . The topic browser: An interactive tool for browsing topic models . In NIPS Workshop on Challenges of Data Visualization ( 2010 ). Gardner, M. J., Lutes, J., Lund, J., Hansen, J., Walker, D., Ringger, E., and Seppi, K. The topic browser: An interactive tool for browsing topic models. In NIPS Workshop on Challenges of Data Visualization (2010).
- 10.5555/2140533.2140539 · doi:10.5555/2140533.2140539
- 10.5555/1699510.1699543 · doi:10.5555/1699510.1699543
Cited by
A Systematic Review of Social Presence: Definition, Antecedents, and Implications
Oh, C. S., Bailenson, J. N., & Welch, G. F. (2018) · Frontiers in Robotics and AI · doi:10.3389/frobt.2018.00114 · Cited by 1,032
“In contrast, it is possible that people will prefer lower levels of social presence when they are feeling more vulnerable; the success of text-based counseling and support systems (Dinakar et al., 2015) lends some support to this conjecture.”
BackgroundSociety-in-the-loop: programming the algorithmic social contract
Rahwan, I. (2018) · Ethics and Information Technology · doi:10.1007/s10676-017-9430-8 · Cited by 553
“For instance, in a crisis counseling system, a machine learning system classifies messages sent by callers, and provides visualizations to a human counselor in real-time [23].”
BackgroundKey Considerations for Incorporating Conversational AI in Psychotherapy
Miner, A. S., Shah, N., Bullock, K. D., Arnow, B. A., Bailenson, J., & Hancock, J. (2019) · Frontiers in Psychiatry · doi:10.3389/fpsyt.2019.00746 · Cited by 166
“, natural language processing and machine learning) are being used to assess language patterns of successful crisis interventions in text-based mental health conversations (18, 35).”
BackgroundThe Effectiveness of Crisis Line Services: A Systematic Review
Hoffberg, A. S., Stearns-Yoder, K. A., & Brenner, L. A. (2020) · Frontiers in Public Health · doi:10.3389/fpubh.2019.00399 · Cited by 150
Park, A., & Conway, M. (2017) · Journal of Medical Internet Research · doi:10.2196/jmir.6826 · Cited by 113
“Although beyond the scope of this study, an investigation with respect to changes of emotion in users of different systems such as Internet-based interventions for depression [52] or text-based crisis counselor system [53] is important for individuals suffering from depression.”
Metadata
- Authors
- Dinakar, K., Chen, J., Lieberman, H., Picard, R., & Filbin, R.
- Affiliations
- Massachusetts Institute of Technology (US)
- Crisis Text Line, New York, NY, USA
- Venue
- Proceedings of the 20th International Conference on Intelligent User Interfaces
- Publisher
- ACM
- Published
- 2015-03-18
- Pages
- 417-426
- OpenAlex
- W2084551014
- Cited by
- 42
- References
- 25
- Citations by year
- 2026: 1 · 2025: 6 · 2024: 1 · 2023: 1 · 2022: 2 · 2021: 4 · 2020: 8 · 2019: 5 · 2018: 4 · 2017: 5 · 2016: 4 · 2015: 1
- FWCI
- 1.91
- Type
- conference paper
- Language
- en
- Open access
- green
- hdl.handle.net
- Licence
- Locations
- ACM International Conference on Intelligent User Interfaces (IUI) (publishedVersion)
- DSpace@MIT (Massachusetts Institute of Technology) (submittedVersion, cc-by-nc-sa)
- Funders
- Google Faculty Grant
- Reid Hoffman Fellowship
- Topics
- Physical Sciences › Computer Science › Computer Graphics and Computer-Aided Design › Data Visualization and Analytics
- Social Sciences › General Social Sciences › Computational and Text Analysis Methods
- Social Sciences › Psychology › Social Psychology › Mental Health via Writing
- Keywords
- crisis counseling, Crisis Text Line, topic modeling, mixed-initiative systems, real-time visualization, natural language interfaces
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.