Pith. sign in

REVIEW 3 major objections 5 minor 39 references

Transparent but Powerful: Explainability, Accuracy, and Generalizability in ADHD Detection from Social Media Data

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A feature-based BiLSTM matches the best transformer for ADHD detection on Reddit, and feature models transfer more robustly to Twitter.

desk verdict A solid, honest applied NLP study whose central ADHD-detection claim is undermined by a control group that confounds ADHD with mental-health discourse in general. read the letter →

arxiv 2411.15586 v1 pith:VPXUN3XW submitted 2024-11-23 cs.CL

D. Wiechmann , E. Kempa , E. Kerz , Y. Qiao This is my paper · ORCID
classification cs.CL
keywords ADHDdetectionsocialmediaNLPinterpretablemachinelearningBiLSTMtransformergeneralizationfeatureablationRedditTwitter
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to show that ADHD can be picked out from ordinary social-media text with moderate accuracy, and that this does not force a choice between transparency and performance. On Reddit data, a BiLSTM trained on 344 interpretable linguistic features reaches an F1-score of 0.77, matching the domain-adapted transformer MentalRoBERTa. When the same models are tested on Twitter, the feature-based classifiers degrade less (F1 drop of 0.06–0.12) than transformers (drop of 0.13–0.15), suggesting that human-readable features are more portable across platforms. The paper also identifies which feature groups carry the signal, with readability, grammatical categories, and emotion leading the rankings. If correct, these results make a case for building explainable digital screening tools whose decisions clinicians can inspect.

What carries the argument

The load-bearing machinery is a sentence-level feature extraction system that computes 344 human-interpretable measures in eight groups: Syntactic Complexity, Lexical Richness/Complexity, Cohesion, Stylistics, Readability, Grammatical Categories, Topical Categories, and Emotion Categories. For each user, these sentence vectors are averaged for shallow classifiers or fed as sequences into a three-layer Bidirectional Long Short-Term Memory (BiLSTM) network. Interpretability comes from SP-LIME, a global aggregation of LIME local linear explanations that ranks the eight feature groups, and from random-forest mean decrease in impurity, which ranks individual features. The feature-first design is what lets the authors trace predictions back to specific linguistic dimensions.

What would settle it

Take the same feature pipeline and retrain the classifiers with a control group sampled from users who post actively in mental-health subreddits about conditions other than ADHD, matching the ADHD group on posting volume and mental-health topic mentions; if the F1-score drops to chance, the reported signal is mental-health discourse, not ADHD.

Watch

Extended reading notes

Core claim

The central claim is that linguistic features engineered from eight dimensions of verbal behavior are sufficient for ADHD detection at levels comparable to fine-tuned transformers, while remaining interpretable. The paper reports in-domain F1-scores of 0.74–0.77 across shallow and deep models, with BiLSTM and MentalRoBERTa both at 0.77; BiLSTM achieves the highest recall (0.79). In out-of-distribution transfer from Reddit to Twitter, Gradient Boosting is the most robust (F1 0.70, drop 0.06), whereas RoBERTa and MentalRoBERTa lose recall heavily and land at F1 0.63 and 0.62. Feature ablations using SP-LIME rank Readability, Grammatical Categories, and Topical Categories as the most informative groups, and the paper links these to measurable linguistic patterns: lower readability with higher syntactic and lexical complexity, greater use of self-referential pronouns, more negative emotion words, and more health-related topics. The conclusion is that interpretable feature-based models generalize better than black-box transformers for this task.

Load-bearing premise

The weakest assumption is that control users who never mention mental health are a fair comparison for ADHD users, who by definition discuss ADHD and frequently other mental-health topics; the models may be detecting mental-health discourse rather than ADHD-specific language.

Editorial extensions

If this is right

  • A screening tool built on these features could show clinicians which linguistic dimensions pushed a given user into the ADHD-positive group, unlike a transformer's attention weights.
  • The feature-ranking results give candidate digital biomarkers—low readability, high self-referential pronoun use, elevated negative-emotion words—that can be tested in clinical interviews or diary text.
  • The small OOD degradation of feature models (0.06–0.12 F1 drop) suggests the framework may transfer to other social platforms or messaging apps without full retraining.
  • The large recall drop of transformers on Twitter (0.49) indicates that transformer embeddings pick up platform-specific phrasing, so cross-platform deployment should favor feature-based models.
  • The reported 0.77 F1 can serve as a baseline for future multimodal or LLM-based ADHD detection on Reddit data.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the control group never discusses mental health, the model may be learning topic rather than ADHD; a stricter control that also posts about mental-health topics (e.g., anxiety or depression) would reveal how much of the signal is ADHD-specific.
  • The 0.77 ceiling may reflect noisy self-reported labels on Reddit; the same feature framework with verified clinical labels could plausibly score higher or lower, testing the validity of the digital phenotype.
  • The same eight feature groups could be applied to autism or anxiety detection; if the same groups (readability, emotion, self-reference) dominate, they may be shared distress markers rather than ADHD-specific.
  • A direct test of the readability finding would be to compare clause-per-sentence and vocabulary measures against word count alone, to see whether ADHD language is structurally different or just longer and more verbose.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper presents user-level ADHD detection from social media text, comparing shallow feature-based classifiers (Logistic Regression, Random Forest, SVM, Gradient Boosting), a BiLSTM trained on 344 hand-crafted linguistic features, and transformer baselines (RoBERTa, MentalRoBERTa). The authors report that BiLSTM and MentalRoBERTa achieve the best in-domain F1 (0.77) on a new Reddit dataset, that feature-based models degrade less than transformers in an out-of-domain Twitter evaluation, and that SP-LIME and MDI analyses identify interpretable linguistic markers of ADHD.

Significance. If the central claims hold, the paper would make a useful contribution to interpretable digital phenotyping for ADHD: it reimplements and extends the SMHD collection methodology, uses a relatively large user-level sample, includes an explicit out-of-distribution evaluation across platforms, and compares transparent models against strong transformer baselines. The manuscript is weakened, however, by a control-group construction that couples the ADHD label with the presence of mental-health-related discourse, and by the absence of confidence intervals or significance tests for the reported performance differences. The feature-attribution experiments are also mislabeled as ablations. These issues are addressable but require additional experiments and reanalysis.

major comments (3)
  1. [Section 3.1] The control group is defined as users who never posted or commented in mental health-related subreddits and never mentioned ADHD or similar mental health terms. Because the ADHD group is identified by self-reported diagnosis phrases and its retained posts deliberately include mental-health-related content (the paper states that posts directly referencing ADHD diagnoses were removed while mental health-related posts were retained), the label is coupled with the presence or absence of mental-health discourse. A model can therefore achieve high F1 by detecting mental-health-related language generally rather than ADHD-specific linguistic patterns. This confound also affects the feature-importance interpretation in Tables 3 and 5, where emotion categories and health topics dominate. The authors should evaluate against a control group of users who post about other mental health conditions, or otherwise demonstrate that the classifier separates ADHD from non-ADHD mental-health discourse.
  2. [Section 4.1 / Section 4.3, Tables 2 and 4] The key comparative claims, that BiLSTM and MentalRoBERTa lead with F1 0.77 and that feature-based models are more robust than transformers out-of-domain, are reported without confidence intervals or significance tests. The reported differences are small (0.01 in the in-domain comparison and 0.07-0.08 in the out-of-domain comparison) and, in the absence of uncertainty estimates on user-level predictions, do not establish model ordering or a robustness advantage. Bootstrap confidence intervals or paired significance tests should be added for both the Reddit and Twitter evaluations.
  3. [Section 4.2, Table 3] The experiments labeled feature ablation do not remove feature groups and retrain the models; SP-LIME is a post-hoc global explanation method, and MDI is a similar attribution measure from the Random Forest. The claimed contribution Feature Ablation and the conclusion that specific feature groups play a significant role are therefore not supported by the reported experiments. Either retrain the models with feature groups excluded and report the resulting performance changes, or rename the section and the corresponding claims to feature attribution.
minor comments (5)
  1. [Introduction] There is a typo in the citation Zhou et al., 2022, with a doubled comma and space; this should be corrected.
  2. [Section 7, Limitations] The Limitations section lists several threats to generalizability but does not acknowledge the control-group selection confound described in my major comment; this omission should be addressed explicitly in a revision.
  3. [Appendix, Pre-processing] The threshold that discards posts with fewer than three sentences is not accompanied by exclusion counts for the ADHD and control groups; because the ADHD group has substantially more texts per user (317,073 vs. 174,765), the threshold could affect the two groups differentially, and this should be reported.
  4. [Section 3.3] Hyperparameter tuning is described only as grid search and sequential tuning without listing the search ranges; reporting the grids would be necessary for reproducibility.
  5. [References] The reference Loshchilov, 2017 is incomplete; the paper should cite the actual AdamW publication (Loshchilov and Hutter, ICLR 2019).

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: reported F1 scores are empirical measurements on held-out and external data, not artifacts of fitted inputs or self-citation.

full rationale

This is an empirical classification study. The headline numbers (F1 0.77 for BiLSTM and MentalRoBERTa; OOD F1s in Table 4) are measured on a held-out 10% Reddit test split and on the external Twitter-STMHD dataset, respectively, not obtained by substituting a fitted parameter into the claimed output. No equation in the paper defines a reported result in terms of its own inputs: the SP-LIME and MDI feature importances are post-hoc explanations of already-trained classifiers, and feature ablation does not feed back into model fitting or into the evaluation metrics. The only author self-citation (Kerz et al., 2023, cited as 'an exception' for explainability) is contextual and not load-bearing; the transformer baselines RoBERTa and MentalRoBERTa are externally published models, which grounds the comparison independently. The control-group construction (users who never posted in mental-health subreddits vs. users identified by ADHD diagnosis phrases) is a real construct-validity threat, because the retained mental-health-related posts could let a model detect mental-health discourse rather than ADHD-specific language, and the reader should treat the reported F1 with that caveat. However, that is a confounding/validity problem, not a circularity problem under the definitions used here: there is no step where a prediction is equivalent to an input by construction, nor where a fitted parameter is renamed as a prediction. Therefore the circularity score is 0.

Assumptions & free parameters 8 free parameters · 4 assumptions · 0 invented entities

The central claims rest on the dataset construction rules, the proprietary feature set, and standard ML procedures. The most consequential free parameters are the label proximity threshold and the control selection rule, which jointly define the learning problem. No new theoretical entities are introduced.

free parameters (8)
  • Diagnosis pattern proximity threshold = 40 characters
    A user is labeled ADHD if an ADHD keyword appears within 40 characters of a diagnosis phrase; this hand-chosen threshold affects dataset composition and labels.
  • Minimum sentences per post = 3
    Posts with fewer than three sentences are discarded to enable coherence measurement; this changes the distribution of texts used for training and testing.
  • SP-LIME kernel width sigma = 0.75*sqrt(d)
    Exponential kernel width for weighting perturbed samples in LIME, following the original LIME heuristic; affects feature group importance scores.
  • Logistic Regression regularization C = 0.02
    Selected by grid search on the development set; affects LR performance and feature coefficients.
  • Random Forest n_estimators = 56
    Tuned on development set; affects RF performance and MDI importance values.
  • Gradient Boosting learning rate = 0.4
    Tuned on development set; affects GB performance and the OOD generalization comparison.
  • BiLSTM hidden size = 256
    Model capacity choice tuned on development set; affects BiLSTM F1.
  • Transformer learning rate = 2e-5
    Standard fine-tuning learning rate for RoBERTa; affects all transformer results.
assumptions (4)
  • domain assumption Self-reported diagnosis on social media is a valid proxy for clinically confirmed ADHD.
    The dataset labels rely on users' self-reports of diagnosis (Section 3.1), which may be noisy or inaccurate.
  • domain assumption Users who never mention mental health topics are representative of the non-ADHD population.
    Controls are sampled from users with no mental-health-related posts; this may not represent the general population and creates a confound (Section 3.1).
  • domain assumption The 344-feature CYMO framework covers all linguistic dimensions relevant to ADHD detection.
    The feature set is fixed by a proprietary tool; the paper does not justify its completeness or independence from the target labels (Section 3.2).
  • standard math Standard supervised learning and neural fine-tuning procedures produce valid performance estimates.
    The training, validation, and testing protocols follow common practice in NLP classification (Section 3.3).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Transparent but Powerful: Explainability, Accuracy, and Generalizability in ADHD Detection from Social Media Data." pith.science (2026). https://pith.science/paper/VPXUN3XW

@misc{pith2026241115586,
  author       = {Pith},
  title        = {Pith review of: Transparent but Powerful: Explainability, Accuracy, and Generalizability in ADHD Detection from Social Media Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VPXUN3XW}},
  note         = {Machine review of arXiv:2411.15586}
}
read the original abstract

Attention-deficit/hyperactivity disorder (ADHD) is a prevalent mental health condition affecting both children and adults, yet it remains severely underdiagnosed. Recent advances in artificial intelligence, particularly in Natural Language Processing (NLP) and Machine Learning (ML), offer promising solutions for scalable and non-invasive ADHD screening methods using social media data. This paper presents a comprehensive study on ADHD detection, leveraging both shallow machine learning models and deep learning approaches, including BiLSTM and transformer-based models, to analyze linguistic patterns in ADHD-related social media text. Our results highlight the trade-offs between interpretability and performance across different models, with BiLSTM offering a balance of transparency and accuracy. Additionally, we assess the generalizability of these models using cross-platform data from Reddit and Twitter, uncovering key linguistic features associated with ADHD that could contribute to more effective digital screening tools.

Figures

Figures reproduced from arXiv: 2411.15586 by the authors.

Figure 1
Figure 1. Text Feature Dynamics Across Social Media Posts. The figure illustrates the dynamic contours of text [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 25 canonical work pages

  1. [1]

    online" 'onlinestring :=

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint eprinttype howpublished institution journal key month note number organization pages publisher school series title type volume year doi pubmed url lastchecked label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block STRING...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    American Psychiatric Association . 2022. https://doi.org/10.1176/appi.books.9780890425787 Diagnostic and Statistical Manual of Mental Disorders , 5 edition. American Psychiatric Association Publishing, Arlington, VA. Text Revision

    doi:10.1176/appi.books.9780890425787 2022

  4. [4]

    Getinet Ayano, Light Tsegay, Yitbarek Gizachew, Mogesie Necho, Kalkidan Yohannes, Mebratu Abraha, Sileshi Demelash, Tamrat Anbesaw, and Rosa Alati. 2023. Prevalence of attention deficit hyperactivity disorder in adults: umbrella review of evidence generated across the globe. Psychiatry Research, page 115449

    work page 2023

  5. [5]

    Aaron T Beck, A John Rush, Brian F Shaw, Gary Emery, Robert J DeRubeis, and Steven D Hollon. 2024. Cognitive therapy of depression. Guilford Publications

    work page 2024

  6. [6]

    Leo Breiman. 2001. Random forests. Machine learning, 45:5--32

    2001

  7. [7]

    Rafael A Calvo, David N Milne, M Sazzad Hussain, and Helen Christensen. 2017. Natural language processing in mental health applications using non-clinical texts. Natural Language Engineering, 23(5):649--685

    work page 2017

  8. [8]

    Samuel R Chamberlain, Konstantinos Ioannidis, Eric W Leppink, Faiza Niaz, Sarah A Redden, and Jon E Grant. 2017. Adhd symptoms in non-treatment seeking young adults: relationship with other forms of impulsivity. CNS spectrums, 22(1):22--30

    work page 2017

Show all 39 references
  1. [9]

    Winston Chung, Sheng-Fang Jiang, Diana Paksarian, Aki Nikolaidis, F Xavier Castellanos, Kathleen R Merikangas, and Michael P Milham. 2019. Trends in the prevalence and incidence of attention-deficit/hyperactivity disorder among adults and children of different racial and ethni...

    2019

  2. [10]

    Arman Cohan, Bart Desmet, Andrew Yates, Luca Soldaini, Sean MacAvaney, and Nazli Goharian. 2018. https://aclanthology.org/C18-1126 SMHD : a large-scale resource for exploring online language usage for multiple mental health conditions . In Proceedings of the 27th International...

    2018

  3. [11]

    Melissa L Danielson, Angelika H Claussen, Rebecca H Bitsko, Samuel M Katz, Kimberly Newsome, Stephen J Blumberg, Michael D Kogan, and Reem Ghandour. 2024. Adhd prevalence among us children and adolescents in 2022: Diagnosis, severity, co-occurring disorders, and treatment. Jou...

    2024

  4. [12]

    Moshe Fried, Eteri Tsitsiashvili, Yoram S Bonneh, Anna Sterkin, Tamara Wygnanski-Jaffe, Tamir Epstein, and Uri Polat. 2014. Adhd subjects fail to suppress eye blinks and microsaccades while anticipating visual stimuli but recover with medication. Vision research, 101:62--72

    2014

  5. [13]

    Muskan Garg. 2023. Mental health analysis in social media posts: a survey. Archives of Computational Methods in Engineering, 30(3):1819--1842

    2023

  6. [14]

    Sharath Chandra Guntuku, J Russell Ramsay, Raina M Merchant, and Lyle H Ungar. 2019. Language of adhd in adults on social media. Journal of attention disorders, 23(12):1475--1485

    2019

  7. [15]

    Keith Harrigian, Carlos Aguirre, and Mark Dredze. 2020. https://doi.org/10.18653/v1/2020.findings-emnlp.337 Do models of mental health based on social media data generalize? In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3774--3788, Online. Ass...

    2020 doi

  8. [16]

    Martine Hoogman, Ryan Muetzel, Joao P Guimaraes, Elena Shumskaya, Maarten Mennes, Marcel P Zwiers, Neda Jahanshad, Gustavo Sudre, Thomas Wolfers, Eric A Earl, et al. 2019. Brain imaging of the cortex in adhd: a coordinated analysis of large-scale clinical and population-based ...

    2019

  9. [17]

    Shaoxiong Ji, Tianlin Zhang, Luna Ansari, Jie Fu, Prayag Tiwari, and Erik Cambria. 2021. Mentalbert: Publicly available pretrained language models for mental healthcare. arXiv preprint arXiv:2110.15621

    2021 arXiv

  10. [18]

    Niloofar Kalantari, Amirreza Payandeh, Marcos Zampieri, and Vivian G Motti. 2023. Understanding the language of adhd and autism communities on social media. In 2023 IEEE International Conference on Big Data (BigData), pages 2188--2195. IEEE

    2023

  11. [19]

    Tim Kendall, Eric Taylor, Alejandra Perez, and Clare Taylor. 2008. Diagnosis and management of attention-deficit/hyperactivity disorder in children, young people, and adults: summary of nice guidance. Bmj, 337

    2008

  12. [20]

    Elma Kerz, Sourabh Zanwar, Yu Qiao, and Daniel Wiechmann. 2023. Toward explainable ai (xai) for mental health detection based on language behavior. Frontiers in psychiatry, 14:1219479

    2023

  13. [21]

    Evelyn MR Lake, Emily S Finn, Stephanie M Noble, Tamara Vanderwal, Xilin Shen, Monica D Rosenberg, Marisa N Spann, Marvin M Chun, Dustin Scheinost, and R Todd Constable. 2019. The functional brain organization of an individual allows prediction of measures of social abilities ...

    2019

  14. [22]

    Dong Yun Lee, Yunmi Shin, Rae Woong Park, Sun-Mi Cho, Sora Han, Changsoon Yoon, Jaheui Choo, Joo Min Shim, Kahee Kim, Sang-Won Jeon, et al. 2023. Use of eye tracking to improve the identification of attention-deficit/hyperactivity disorder in children. Scientific Reports, 13(1):14469

    2023

  15. [23]

    Yunji Liang, Xiaolong Zheng, and Daniel D Zeng. 2019. A survey on big data-driven digital phenotyping of mental health. Information Fusion, 52:290--307

    2019

  16. [24]

    Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692

    2019 arXiv

  17. [25]

    I Loshchilov. 2017. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101

    2017 arXiv

  18. [26]

    Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. 2019. Pytorch: An imperative style, high-performance deep learning library. Advances in neural information processing systems, 32

    2019

  19. [27]

    Guilherme Polanczyk, Maur \' cio Silva De Lima, Bernardo Lessa Horta, Joseph Biederman, and Luis Augusto Rohde. 2007. The worldwide prevalence of adhd: a systematic review and metaregression analysis. American journal of psychiatry, 164(6):942--948

    2007

  20. [28]

    why should i trust you?

    Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2016. " why should i trust you?" explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, pages 1135--1144

    2016

  21. [29]

    Carmen Schiweck, Gara Arteaga-Henriquez, Mareike Aichholzer, Sharmili Edwin Thanarajah, Sebastian Vargas-C \'a ceres, Silke Matura, Oliver Grimm, Jan Haavik, Sarah Kittel-Schneider, Josep Antoni Ramos-Quiroga, et al. 2021. Comorbidity of adhd and adult bipolar disorder: A syst...

    2021

  22. [30]

    Philip Shaw, Argyris Stringaris, Joel Nigg, and Ellen Leibenluft. 2014. Emotion dysregulation in attention deficit hyperactivity disorder. American Journal of Psychiatry, 171(3):276--293

    2014

  23. [31]

    Asmit Kumar Singh, Udit Arora, Somyadeep Shrivastava, Aryaveer Singh, Rajiv Ratn Shah, Ponnurangam Kumaraguru, et al. 2022. Twitter-stmhd: An extensive user-level database of multiple mental health disorders. In Proceedings of the International AAAI Conference on Web and Socia...

    2022

  24. [32]

    Yosephine Susanto, Andrew G Livingstone, Bee Chin Ng, and Erik Cambria. 2020. The hourglass model revisited. IEEE Intelligent Systems, 35(5):96--102

    2020

  25. [33]

    Lisa B Thorell, Hanna Tilling, and Douglas Sj \"o wall. 2020. Emotion dysregulation in adult adhd: Introducing the comprehensive emotion regulation inventory (ceri). Journal of Clinical and Experimental Neuropsychology, 42(7):747--758

    2020

  26. [34]

    S \'e verine Van De Voorde, Herbert Roeyers, and Jan Roelf Wiersema. 2010. Error monitoring in children with adhd or reading disorder: An event-related potential study. Biological psychology, 84(2):176--185

    2010

  27. [35]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, ukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. In Advances in Neural Information Processing Systems, volume 30. Curran Associates, Inc

    2017

  28. [36]

    Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, R \'e mi Louf, Morgan Funtowicz, et al. 2019. Huggingface's transformers: State-of-the-art natural language processing. arXiv preprint arXiv:1910.03771

    2019 arXiv

  29. [37]

    Zhang, A

    T. Zhang, A. Schoene, and S. Ananiadou. 2022 a . https://doi.org/10.1038/s41746-022-00589-7 Natural language processing applied to mental illness detection: a narrative review . NPJ Digital Medicine, 5:46. Overview of what's been used/done and sources etc

    2022 doi

  30. [38]

    Tianlin Zhang, Annika M Schoene, Shaoxiong Ji, and Sophia Ananiadou. 2022 b . Natural language processing applied to mental illness detection: a narrative review. NPJ digital medicine, 5(1):1--13

    2022

  31. [39]

    Binggui Zhou, Guanghua Yang, Zheng Shi, and Shaodan Ma. 2022. Natural language processing for smart healthcare. IEEE Reviews in Biomedical Engineering

    2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.