Approaches to Capacity Building for Machine Learning and Artificial Intelligence Applications in Health
Notice bibliographique
Résumé
Many health systems and research institutes are interested in supplementing their traditional analyses of linked data with machine learning (ML) and other artificial intelligence (AI) methods and tools. However, the availability of individuals who have the required skills to develop and/or implement ML/AI is a constraint, as there is high demand for ML/AI talent in many sectors. The three organizations presenting are all actively involved in training and capacity building for ML/AI broadly, and each has a focus on, and/or discrete initiatives for, particular trainees. P. Alison Paprica, Vector Institute for artificial intelligence, Institute for Clinical Evaluative Sciences, University of Toronto, Canada. Alison is VP, Health Strategy and Partnerships at Vector, responsible for health strategy and also playing a lead role in “1000AIMs” – a Vector-led initiative in support of the Province of Ontario’s \$30 million investment to increase the number of AI-related master’s program graduates to 1,000 per year within five years. Frank Sullivan, University of St Andrews Scotland. Frank is a family physician and an associate director of HDRUK@Scotland. Health Data Research UK \url{https://hdruk.ac.uk/} has recently provided funding to six sites across the UK to address challenging healthcare issues through use of data science. A 50 PhD student Doctoral Training Scheme in AI has also been announced. Each site works in close partnership with National Health Service bodies and the public to translate research findings into benefits for patients and populations. Yin Aphinyanaphongs – INTREPID NYU clinical training program for incoming clinical fellows. Yin is the Director of the Clinical Informatics Training Program at NYU Langone Health. He is deeply interested in the intersection of computer science and health care and as a physician and a scientist, he has a unique perspective on how to train medical professionals for a data drive world. One version of this teaching process is demonstrated in the INTREPID clinical training program. Yin teaches clinicians to work with large scale data within the R environment and generate hypothesis and insights. The session will begin with three brief presentations followed by a facilitated session where all participants share their insights about the essential skills and competencies required for different kinds of ML/AI application and contributions. Live polling and voting will be used at the end of the session to capture participants’ view on the key learnings and take away points. The intended outputs and outcomes of the session are: Participants will have a better understanding of the skills and competencies required for individuals to contribute to AI applications in health in various ways Participants will gain knowledge about different options for capacity building from targeted enhancement of the skills of clinical fellows, to producing large number of applied master’s graduates, to doctoral-level training After the session, the co-leads will work together to create a resource that summarizes the learnings from the session and make them public (though publication in a peer-reviewed journal and/or through the IPDLN website)
Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.
Comment cette classification a été obtenuedéplier
Prédiction machine sur la base complète
Imitation des enseignantsNi prévalence calibrée, ni vérité terrain. Validation humaine à venir. Le volet Gemma est une étiquette directe du modèle pour chaque travail de la base, lue sur la notice réduite au titre. Le volet Codex est un classifieur appris des 10 348 étiquettes directes de Codex et calibré sur les taux pondérés de l'échantillon; les champs sans appui suffisant ne portent aucun appel Codex. Le mode candidate est l'union des deux volets; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont pas des étiquettes humaines.
Scores du classifieur distillé par catégorie (deux têtes)
| Catégorie | Codex | Gemma |
|---|---|---|
| Métarecherche | 0,029 | 0,046 |
| Méta-épidémiologie (sens strict) | 0,002 | 0,001 |
| Méta-épidémiologie (sens large) | 0,001 | 0,002 |
| Bibliométrie | 0,005 | 0,004 |
| Études des sciences et des technologies | 0,003 | 0,017 |
| Communication savante | 0,011 | 0,013 |
| Science ouverte | 0,007 | 0,015 |
| Intégrité de la recherche | 0,006 | 0,007 |
| Charge utile insuffisante (le modèle a refusé de juger) | 0,033 | 0,003 |
Scores machine (provisoires)
Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.
Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.
score_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découleClassification
machine, non validéePrédiction automatique; un appel candidat d’une seule source (Gemma direct ou Codex distillé), pas un consensus.
Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».