Learning to assess the quality of stroke rehabilitation exercises
Open accessThe DOI matches its Crossref record; 51 of the 51 references it cites were found in a registry or an index.
Abstract
Due to the limited number of therapists, task-oriented exercises are often prescribed for post-stroke survivors as in-home rehabilitation. During in-home rehabilitation, a patient may become unmotivated or confused to comply prescriptions without the feedback of a therapist. To address this challenge, this paper proposes an automated method that can achieve not only qualitative, but also quantitative assessment of stroke rehabilitation exercises. Specifically, we explored a threshold model that utilizes the outputs of binary classifiers to quantify the correctness of a movements into a performance score. We collected movements of 11 healthy subjects and 15 post-stroke survivors using a Kinect sensor and ground truth scores from primary and secondary therapists. The proposed method achieves the following agreement with the primary therapist: 0.8436, 0.8264, and 0.7976 F1-scores on three task-oriented exercises. Experimental results show that our approach performs equally well or better than multi-class classification, regression, or the evaluation of the secondary therapist. Furthermore, we found a strong correlation (R2 = 0.95) between the sum of computed exercise scores and the Fugl-Meyer Assessment scores, clinically validated motor impairment index of post-stroke survivors. Our results demonstrate a feasibility of automatically assessing stroke rehabilitation exercises with the decent agreement levels and clinical relevance.
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
“Although previous research describes the potential of AI/ML systems to outperform human experts on prediction tasks [18, 32, 40]”
From Understanding the Effect of Counterfactual Explanations on Trust and Reliance on AI for Human-AI Collaborative Clinical Decision Making (2023) · cites this paper as [40]
Citations
Cites
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016 · cited 2×
“We utilize ‘Scikit-learn’ [34] and ‘PyTorch’ [31] libraries to implement Decision Trees (DTs), Linear Regression (LR), Support Vector Machine (SVM), Neural Networks (NNs), and Long Short Term Memory (LSTM) Network.”
Method(2017) · Cited by 11,625
(2006) · Springer eBooks · Cited by 8,123
Billinger, S. A., Arena, R., Bernhardt, J., Eng, J. J., Franklin, B. A., Johnson, C. M., MacKay-Lyons, M., Macko, R. F., Mead, G. E., Roth, E. J., et al. (2014) · Stroke · doi:10.1161/str.0000000000000022 · Cited by 1,529
Technology probes: inspiring design for and with families
Hutchinson, H., Mackay, W., Westerlund, B., Bederson, B. B., Druin, A., Plaisant, C., Beaudouin-Lafon, M., Conversy, S., Evans, H., Hansen, H., et al. (2003) · Proceedings of the SIGCHI Conference on Human Factors in Computing Systems · doi:10.1145/642611.642616 · Cited by 1,471 · cited 1×
“Instead, this paper focuses on presenting a method of assessing the quality of stroke rehabilitation to collect patient’s performance data and further demonstrating how well it can assess compare to therapists as a technology probe [17, 18].”
MethodEvidence for prescribing exercise as therapy in chronic disease
Pedersen, B. K., & Saltin, B. (2006) · Scandinavian Journal of Medicine & Science in Sports · doi:10.1111/j.1600-0838.2006.00520.x · Cited by 1,384 · cited 1×
“A therapist prescribes to repeat few exercises with major muscle groups during in-home rehabilitation [2, 33].”
BackgroundTransforming classifier scores into accurate multiclass probability estimates
Zadrozny, B., & Elkan, C. (2002) · Proceedings of the eighth ACM SIGKDD international conference on Knowledge discovery and data mining · doi:10.1145/775047.775151 · Cited by 1,280 · cited 1×
“These binary classifiers can predict the correctness of performance components and generate a confidence score to estimate performance score [13, 14, 23, 48] as shown in Figure 4.”
MethodManagement of Adult Stroke Rehabilitation Care: A Clinical Practice Guideline
Duncan, P. W., Zorowitz, R., Bates, B., Choi, J. Y., Glasberg, J. J., Graham, G. D., Katz, R. C., Lamberty, K., & Reker, D. (2005) · Stroke · doi:10.1161/01.str.0000180861.54180.ff · Cited by 1,046 · cited 1×
“As a post-stroke survivor mainly interacts with a single therapist for the consistent assessment and guidance [11, 39], our study mainly explores how well our proposed method can be calibrated with the primary therapist’s assessment and applicable to multiple patients.”
MethodThe impact of physical therapy on functional outcomes after stroke: what's the evidence?
Van Peppen, R. P., Kwakkel, G., Wood-Dauphinee, S., Hendriks, H. J., Van der Wees, P. J., & Dekker, J. (2004) · Clinical Rehabilitation · doi:10.1191/0269215504cr843oa · Cited by 942 · cited 1×
“One of the effective stroke rehabilitation approaches is physical therapy intervention with task-oriented exercises [21, 39, 47].”
BackgroundAn Efficient Boosting Algorithm for Combining Preferences
Freund, Y., Iyer, R. D., Schapire, R. E., & Singer, Y. (1998) · Cited by 806
Motor recovery after stroke: A systematic review of the literature
Hendricks, H. T., van Limbeek, J., Geurts, A. C., & Zwarts, M. J. (2002) · Archives of Physical Medicine and Rehabilitation · doi:10.1053/apmr.2002.35473 · Cited by 779
Movement Smoothness Changes during Stroke Recovery
Rohrer, B., Fasoli, S., Krebs, H. I., Hughes, R., Volpe, B., Frontera, W. R., Stein, J., & Hogan, N. (2002) · The Journal of Neuroscience · doi:10.1523/jneurosci.22-18-08297.2002 · Cited by 767 · cited 5×
“The ‘Smoothness’ component can be indicated by velocity related features [5, 32, 40, 41, 46].”
BackgroundInfluentialReliability of the Fugl-Meyer Assessment for Testing Motor Performance in Patients Following Stroke
Sanford, J., Moreland, J., Swanson, L. R., Stratford, P. W., & Gowland, C. (1993) · Physical Therapy · doi:10.1093/ptj/73.7.447 · Cited by 732 · cited 19×
“In addition to recording exercise motions, a therapist assessed post-stroke survivor’s functional ability of using the Fugl Meyer Assessment (FMA).”
BackgroundMethodInfluentialA Simple Approach to Ordinal Classification
Frank, E., & Hall, M. (2001) · Lecture Notes in Computer Science · doi:10.1007/3-540-44795-4_13 · Cited by 606 · cited 1×
“Prior work showed that well-tuned binary classification approaches can be transformed into good ranking algorithms [12, 22].”
BackgroundOrdinal Regression Methods: Survey and Experimental Study
Gutierrez, P. A., Perez-Ortiz, M., Sanchez-Monedero, J., Fernandez-Navarro, F., & Hervas-Martinez, C. (2016) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2015.2457911 · Cited by 483 · cited 6×
“They aim to learning a function, f (x) that predicts the values of a latent variable [14] and estimating a categorical response with a set of threshold in the range of f (x).”
BackgroundMethodInfluentialQualitative analysis of stroke patients' motivation for rehabilitation
Maclean, N. (2000) · BMJ · doi:10.1136/bmj.321.7268.1051 · Cited by 458 · cited 2×
“A patient might feel uncertain whether he/she correctly performs an exercise and become unmotivated in rehabilitation regimens [2, 26].”
BackgroundDesign strategies to improve patient motivation during robot-aided rehabilitation
Colombo, R., Pisano, F., Mazzone, A., Delconte, C., Micera, S., Carrozza, M. C., Dario, P., & Minuco, G. (2007) · Journal of NeuroEngineering and Rehabilitation · doi:10.1186/1743-0003-4-3 · Cited by 385
Murphy, M. A., Willén, C., & Sunnerhagen, K. S. (2011) · Neurorehabilitation and Neural Repair · doi:10.1177/1545968310370748 · Cited by 360
A Survey of Applications and Human Motion Recognition with Microsoft Kinect
Lun, R., & Zhao, W. (2015) · International Journal of Pattern Recognition and Artificial Intelligence · doi:10.1142/s0218001415550083 · Cited by 300 · cited 1×
“Another approach is to apply gesture recognition [10, 24] that classifies incorrect movements.”
BackgroundTask‐oriented training in rehabilitation after stroke: systematic review
Rensink, M., Schuurmans, M., Lindeman, E., & Hafsteinsdóttir, T. (2009) · Journal of Advanced Nursing · doi:10.1111/j.1365-2648.2008.04925.x · Cited by 256 · cited 3×
“One of the effective stroke rehabilitation approaches is physical therapy intervention with task-oriented exercises [21, 39, 47].”
BackgroundMethodOrdinal Regression by Extended Binary Classification
Li, L., & Lin, H.-T. (2007) · Advances in Neural Information Processing Systems 19 · doi:10.7551/mitpress/7503.003.0113 · Cited by 244 · cited 4×
“The state-of-the-art approaches to classify categorical scales assume that the categorical response is a measured latent continuous variable, which can be modeled with intervals on the real line [14, 22].”
BackgroundInfluentialReliability of the Melbourne Assessment of Unilateral Upper Limb Function
Randall, M., Carlin, J. B., Chondros, P., & Reddihough, D. (2001) · Developmental Medicine and Child Neurology · doi:10.1017/s0012162201001396 · Cited by 232 · cited 1×
“‘0: cannot perform’, ‘1: partially perform’, ‘2: fully perform’ [37]).”
Barriers and Facilitators to Exercise Among Stroke Survivors
Damush, T. M., Plue, L., Bakas, T., Schmid, A., & Williams, L. S. (2007) · Rehabilitation Nursing · doi:10.1002/j.2048-7940.2007.tb00183.x · Cited by 201 · cited 4×
“They described the need of a trainer, who provides motivation and coaching on their performance [9].”
BackgroundInfluentialA Kinect-based upper limb rehabilitation system to assist people with cerebral palsy
Chang, Y.-J., Han, W.-Y., & Tsai, Y.-C. (2013) · Research in Developmental Disabilities · doi:10.1016/j.ridd.2013.08.021 · Cited by 188 · cited 3×
“Range of Motion (ROM) The amount of movement around a specific joint [5, 6, 16, 29, 41, 45, 46, 49, 50] Incorrect 0 Does not or barely involve any movement 1 Attempt is made but limited to be normal Correct 2 Movement appears to be normal”
BackgroundMethodEvaluation of an inexpensive depth camera for in-home gait assessment
Stone, E., & Skubic, M. (2011) · Journal of Ambient Intelligence and Smart Environments · doi:10.3233/ais-2011-0124 · Cited by 169 · cited 1×
“This work applies a moving average filter with the window size of five frames to reduce noise of acquiring joint positions from a Kinect similar to [44].”
MethodBourke‐Taylor, H. M. (2003) · Developmental Medicine & Child Neurology · doi:10.1017/s0012162203000185 · Cited by 151
Kinect-enabled home-based rehabilitation system using Dynamic Time Warping and fuzzy logic
Su, C.-J., Chiang, C.-Y., & Huang, J.-Y. (2014) · Applied Soft Computing · doi:10.1016/j.asoc.2014.04.020 · Cited by 149 · cited 2×
“Range of Motion (ROM) The amount of movement around a specific joint [5, 6, 16, 29, 41, 45, 46, 49, 50] Incorrect 0 Does not or barely involve any movement 1 Attempt is made but limited to be normal Correct 2 Movement appears to be normal”
BackgroundAutomatic recognition of gait-related health problems in the elderly using machine learning
Pogorelc, B., Bosnić, Z., & Gams, M. (2012) · Multimedia Tools and Applications · doi:10.1007/s11042-011-0786-1 · Cited by 145 · cited 1×
“utilized k-nearest neighbors and neural network algorithms to recognize four gait related problems with body-worn tags and wall-mounted sensors [35].”
MethodJohnson, M. J., Feng, X., Johnson, L. M., & Winters, J. M. (2007) · Journal of NeuroEngineering and Rehabilitation · doi:10.1186/1743-0003-4-6 · Cited by 117
Joint angle estimation in rehabilitation with inertial sensors and its integration with Kinect
Bo, A. P. L., Hayashibe, M., & Poignet, P. (2011) · 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society · doi:10.1109/iembs.2011.6090940 · Cited by 116 · cited 1×
“joint angles [4]), count repetitions of exercises, and detect incorrect movements [16, 19, 30] to support in-home rehabilitation.”
BackgroundBarriers to achieving evidence-based stroke rehabilitation
Pollock, A., Legg, L. A., Langhorne, P., & Sellars, C. (2000) · Clinical Rehabilitation · doi:10.1191/0269215500cr369oa · Cited by 115 · cited 2×
“However, it is expensive and difficult for post-stroke survivors to receive the administration of an individualized rehabilitation session [36].”
BackgroundAutomated Assessment of Upper Extremity Movement Impairment due to Stroke
Olesh, E. V., Yakovenko, S., & Gritsenko, V. (2014) · PLoS ONE · doi:10.1371/journal.pone.0104487 · Cited by 113 · cited 1×
“described the feasibility of measuring the movement impairment with linear regression between two principal components of four joint angles and qualitative scores from eight subjects [28].”
BackgroundEvidence-based practice for stroke
Langhorne, P., Sandercock, P. A., & Prasad, K. (2009) · The Lancet Neurology · doi:10.1016/s1474-4422(09)70060-2 · Cited by 99
Das, S., Trutoiu, L., Murai, A., Alcindor, D., Oh, M., De la Torre, F., & Hodgins, J. (2011) · 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society · doi:10.1109/iembs.2011.6091674 · Cited by 84 · cited 2×
“applied a Support Vector Machine (SVM) classifier to distinguish mild and severe symptoms of Parkinson’s Disease using full body motion capture data from four Parkinson’s patients [10].”
BackgroundMethodPatterson, T. S., Bishop, M. D., McGuirk, T. E., Sethi, A., & Richards, L. G. (2011) · Journal of Motor Behavior · doi:10.1080/00222895.2010.548422 · Cited by 80
A Kinect-based rehabilitation exercise monitoring and guidance system
Zhao, W., Feng, H. L., Lun, R., Espy, D. D., & Reinthal, M. A. (2014) · IEEE International Conference on Software Engineering and Service Science (ICSESS) · doi:10.1109/icsess.2014.6933678 · Cited by 78
Statistical Learning from a Regression Perspective
Shalabh (2009) · Journal of the Royal Statistical Society Series A (Statistics in Society) · doi:10.1111/j.1467-985x.2009.00614_2.x · Cited by 75
Ozturk, A., Tartar, A., Hüseyinsinoğlu, B. E., & Ertas, A. H. (2015) · Measurement · doi:10.1016/j.measurement.2015.11.026 · Cited by 73 · cited 5×
“The ‘ROM’ component can be represented by joint angles [6, 16, 29, 41, 49, 50] or joint trajectory positions [5, 45, 46, 49].”
BackgroundInfluentialA technology probe of wearable in-home computer-assisted physical therapy
Huang, K. I., Sparto, P. J., Kiesler, S. B., Smailagic, A., Mankoff, J. C., & Siewiorek, D. P. (2014) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/2556288.2557416 · Cited by 53 · cited 1×
“Instead, this paper focuses on presenting a method of assessing the quality of stroke rehabilitation to collect patient’s performance data and further demonstrating how well it can assess compare to therapists as a technology probe [17, 18].”
MethodSelf-citationSouza, C., Cross, A., Gustafsson, A., Catalano, M. D., Voss, S., Zavvari, H., Messer, A., Markou, N., Fo40225, A., et al. (2017) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.1029480 · Cited by 36
Reid, D., & Hirji, T. (2003) · Physical & Occupational Therapy In Geriatrics · doi:10.1080/j148v21n04_01 · Cited by 31 · cited 2×
“According to the studies of [25, 38], they found that viewing their scores on a screen made them motivated to beat their previous scores.”
BackgroundArchitecture and Applications of Virtual Coaches
Siewiorek, D. P., Smailagic, A., & Dey, A. K. (2012) · Proceedings of the IEEE · doi:10.1109/jproc.2012.2200560 · Cited by 29 · cited 1×
“Recent advances in sensors and machine learning algorithms offer a potential of computer-based in-home rehabilitation [20, 42].”
BackgroundSelf-citationMacdonald, G. A., Kayes, N., & Bright, F. A. S. (2013) · Cited by 29
Intelligent Shoulder Joint Home-Based Self-Rehabilitation Monitoring System
Pan, J., Chung, H.-W., & Huang, J.-J. (2013) · International Journal of Smart Home · doi:10.14257/ijsh.2013.7.5.38 · Cited by 29 · cited 1×
“joint angles [4]), count repetitions of exercises, and detect incorrect movements [16, 19, 30] to support in-home rehabilitation.”
BackgroundLarge-Margin Thresholded Ensembles for Ordinal Regression: Theory and Practice
Lin, H.-T., & Li, L. (2006) · Lecture notes in computer science · doi:10.1007/11894841_26 · Cited by 27
de Souza, C. R., Cross, A., Gustafsson, A., et al. (2017) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.1029481 · Cited by 26
A Technology for Computer-Assisted Stroke Rehabilitation
Lee, M. H. (2018) · ACM International Conference on Intelligent User Interfaces (IUI) · doi:10.1145/3172944.3173146 · Cited by 15 · cited 1×
“Based on the discussion with therapists, three exercise are selected due to their correspondence with major motion patterns: Exercise 1 to elbow flexion, Exercise 2 to shoulder flexion, and Exercise 3 to elbow extension [22].”
BackgroundSelf-citationAffordable erehabilitation monitoring platform
Janidarmian, M., Fekr, A. R., Radecka, K., & Žilić, Ž. (2014) · IEEE Canada International Humanitarian Technology Conference · doi:10.1109/ihtc.2014.7147537 · Cited by 11 · cited 1×
“joint angles [4]), count repetitions of exercises, and detect incorrect movements [16, 19, 30] to support in-home rehabilitation.”
BackgroundA Kinect-Based Virtual Rehabilitation System through Gesture Recognition
Zhao, L., Lu, X., Tao, X., & Chen, X. (2016) · International Conference on Virtual Reality and Visualization · doi:10.1109/icvrv.2016.70 · Cited by 11
Binary ranking for ordinal class imbalance
Cruz, R., Fernandes, K., da Costa, J. P., Pérez‐Ortiz, M., & Cardoso, J. S. (2018) · Pattern Analysis and Applications · doi:10.1007/s10044-018-0705-4 · Cited by 5 · cited 1×
“Our approach may be utilized as a way to address imbalance samples in a healthcare application (including computerassisted stroke rehabilitation) [8], which has a costly process of collecting data samples due to widely distributed subjects.”
MethodEvaluation of an inexpensive depth camera for in-home gait assessment
StoneErik, & SkubicMarjorie (2011) · Journal of Ambient Intelligence and Smart Environments · doi:10.5555/2336236.2336241 · Cited by 3
Research works printed but not linked: 2
- 10.5555/645328.649997 · doi:10.5555/645328.649997
- Edward Taub , David M Morris , Jean Crago , Danna Kay King , Mary Bowman, Camille Bryson, Staci Bishop, Sonya Pearson, and Sharon E Shaw. 2011 . Wolf motor function test (WMFT) manual. Birmingham : University of Alabama, CI Therapy Research Group ( 2011). Edward Taub, David M Morris, Jean Crago, Danna Kay King, Mary Bowman, Camille Bryson, Staci Bishop, Sonya Pearson, and Sharon E Shaw. 2011. Wolf motor function test (WMFT) manual. Birmingham: University of Alabama, CI Therapy Research Group (2011).
Cited by
A review of computational approaches for evaluation of rehabilitation exercises
Liao, Y., Vakanski, A., Xian, M., Paul, D. R., & Baker, R. T. (2020) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2020.103687 · Cited by 172
“…New rehabilitation exercises require different rules to be designed [33]–[36], [49], [50], [61], [62], [73], [126], [127], [156], [182], [191]–[195] Template-based approaches Avoid the process of making rules; Can reflect the level of motor ability Only give a overall score for exercise…”
BackgroundA Human-AI Collaborative Approach for Clinical Decision Making on Rehabilitation Assessment
Lee, M. H., Siewiorek, D. P., Smailagic, A., Bernardino, A., & Bermúdez i Badia, S. B. (2021) · Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3411764.3445472 · Cited by 137
“…of rehabilitation monitoring systems that focuses on improving the performance of monitoring and understanding an activity with a complex deep learning model [19, 34, 46, 49], our work highlights the importance of creating a human-centered, interactive approach for better deployment in practice.”
BackgroundMethodInfluentialSelf-citationA Survey of Human Gait-Based Artificial Intelligence Applications
Harris, E., Khoo, I., & Demircan, E. (2022) · Frontiers in Robotics and AI · doi:10.3389/frobt.2021.749274 · Cited by 165
“, 2019), while an inhome rehabilitation system provides qualitative and quantitative feedback to post-stroke survivors (Lee et al., 2019).”
BackgroundA Review of Recent Deep Learning Approaches in Human-Centered Machine Learning
Kaluarachchi, T., Reis, A., & Nanayakkara, S. (2021) · Sensors · doi:10.3390/s21072514 · Cited by 124
“[124], [125], [128], [175], [217] [69], [112], [113], [118], [130], [131], [133], [135], [137], [140], [144], [152], [153], [155]”
MethodSardari, S., Sharifzadeh, S., Daneshkhah, A., Nakisa, B., Loke, S. W., Palade, V., & Duncan, M. J. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.106835 · Cited by 118
Metadata
- Authors
- Lee, M. H., Siewiorek, D. P., Smailagic, A., Bernardino, A., & Badia, S. B. I.
- Affiliations
- Carnegie Mellon University (US)
- Instituto Superior Técnico (PT)
- Madeira Interactive Technology Institute
- Venue
- Proceedings of the 24th International Conference on Intelligent User Interfaces
- Publisher
- ACM
- Published
- 2019-03-17
- Pages
- 218-228
- OpenAlex
- W2915322498
- Cited by
- 71
- References
- 51
- Citations by year
- 2026: 6 · 2025: 12 · 2024: 11 · 2023: 13 · 2022: 12 · 2021: 5 · 2020: 10 · 2019: 2
- FWCI
- 22.36
- Type
- conference paper
- Language
- en
- Open access
- Locations
- ACM International Conference on Intelligent User Interfaces (IUI) (publishedVersion)
- Singapore Management University Institutional Knowledge (InK) (Singapore Management University) (submittedVersion, other-oa)
- Funders
- Fundação para a Ciência e a Tecnologia
- Topics
- Health Sciences › Medicine › Rehabilitation › Stroke Rehabilitation and Recovery
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Context-Aware Activity Recognition Systems
- Life Sciences › Neuroscience › Cognitive Neuroscience › Motor Control and Adaptation
- Keywords
- stroke rehabilitation, exercise quality assessment, task-oriented training, home-based rehabilitation, Kinect sensor, threshold model, Fugl-Meyer Assessment, motor impairment
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.