Facebook
X
LinkedIn
ForwardHands-on AI in Precision Psychiatry
Artificial intelligence techniques are increasingly applied to precision psychiatry, assisting in the estimation of health risks, aiding in diagnosis, recommending treatment planning, and supporting the scientific publication process. This workshop will focus on AI techniques for routinely collected electronic health record data, as well as emerging generative AI methods for more complex modalities such as clinical notes and patient conversations.
In this 5-hour workshop, we will explore 3 core aspects of AI: supervised machine learning for risk prediction (e.g. boosting, SHAP, ensembling, and AutoML), deep learning for complex understanding of patients (e.g. Q&A of clinical notes, facial emotion tracking), and generative AI to support scientific research (e.g. AI research assistance, data science, and image generation).
Each of these sections will feature high-level overviews of essential techniques to build understanding, combined with hands-on running of pre-written code in small groups to make these concepts tangible. Participants will take home high-quality notebooks in R and Python for use in their own projects (or sharing with collaborators), a set of key articles to further explore each technique, as well as a more concrete understanding of emerging opportunities for AI in precision psychiatry.
Please note that additional details, including a breakdown of the workshop structure, information about course instructors, and supplementary workshop materials, will be provided.
Pricing
MDs/Doctoral-Level Professionals/Other Professionals: $300.00
Students/Fellows/Trainees/Interns: $150.00
*Note: You DO NOT need to be registered for the 5th Annual Conference on Precision Psychiatry: Innovation to Implementation on Sept 25 - Sept 26 to register for a workshop.
This workshop is NOT for CME.
This workshop is NOT eligible for discounts.
Workshop Demo Session/Q&A:
Workshop registrants are invited to participate in a demonstration session on Friday, September 19, 2025, from 12:00 PM - 2:00 PM ET. The session will provide an overview of the workshop expectations and include a live Q&A to address participant questions. Additional details and demo session meeting invites will be sent via email.
Target Audience
This workshop is intended for a broad audience, including graduate students, researchers, and clinicians with an interest in machine learning and generative AI applications in mental health. No prior experience is required.
Prerequisites
None. Small groups will be formed that include individuals with and without coding expertise.
Learning Objectives
- Understand the basic principles of supervised machine learning, including popular algorithms like boosting, SHAP variable importance, and hyperparameter tuning.
- Gain intuition for what is distinctive about generative AI and how it differs from standard machine learning.
- Recognize emerging opportunities and clinical use cases for deep learning in psychiatry.
- Gain familiarity with user-friendly AI tools that can assist in scientific research and associated ethical considerations.
Outcomes:
After completion of this workshop, participants will have developed intuition and confidence in how to conduct supervised machine learning for risk prediction, how deep learning can extract complex characteristics from psychiatric patients, and how generative AI can responsibly contribute to the scientific process.
Contact:
For any questions on registration, please email [email protected].
- Background: Overall intro to AI / ML (15 minutes)
- Section 1: Supervised machine learning (2 hours)
- Highest value machine learning algorithms (random forest, boosting)
- Interpretability (SHAP global and local variable importance)
- Performance evaluation (discrimination, calibration, net benefit)
- Averaging multiple prediction algorithms (ensembling)
- Choosing algorithm settings (hyperparameter optimization)
- Automated machine learning
- Coding challenge/questions/break/troubleshooting (30 minutes)
- Section 2: Deep learning for complex understanding of psychiatric patients (1 hour)
- Extracting structured data from clinical notes
- Visual analysis of patient photos and videos
- Audio interaction with patients
- Drafting clinical notes from visits
- Fairness evaluation
- Questions/break (15 minutes)
- Section 3: Generative AI for the scientific process (1 hour)
- Ethics and rules for grants, papers, etc.
- Offline AI chat
- Tools for AI writing assistance
- Literature reviews
- Image generation and editing
- Agentic data science
- Podcast generation
- Scientific co-piloting
- Networking in break-out groups (30 minutes) – Optional
Pre-workshop preparation:
Prior to the workshop, participants will receive instructions to download/install all of the necessary data/tools that will be needed during the workshop. Details will be shared closer to the workshop event via email.
Dr. Chris Kennedy: Chris Kennedy is a biostatistician, data scientist, and instructor in psychiatry at Massachusetts General Hospital / Harvard Medical School. His research focuses on machine/deep learning, psychometrics, and causal inference methods for mental health, with an emphasis on electronic health records. He is the PI of an NIMH K01 that develops machine learning and precision treatment models for emergency psychiatry. He also serves as a co-I of an NIMH-funded RF1 and an NCI-funded R01, the latter using deep learning for surveillance of e-cigarette promotion on social media, and is a co-author of the popular SuperLearner ensemble ML framework. He completed his PhD in biostatistics from the University of California, Berkeley; his postdoctoral fellowship was in the department of biomedical informatics at HMS.
Michael Steigman, MS: Michael is a software engineer with a background in healthcare informatics. He brings his experience building production software systems and infrastructure at MGH to the task of streamlining the Center’s data processes. Michael earned his MS at the University of Texas Health Science Center at Houston and his BA at the University of Miami.
Pratik Nitin Khadse, MSBA: Pratik is a data science expert with a Master’s in Business Analytics from USC Marshall and over five years of experience in predictive modeling, NLP, and applied analytics. At the Center for Precision Psychiatry (CPP), his current focus is on leveraging electronic health record (EHR) data to develop predictive models and apply causal inference methods to address challenges in suicidality, phenotyping, and other areas of healthcare research. Always looking for opportunities to apply his analytics skills in new ways, Pratik aims to continue making valuable contributions through his work.
Price
This optional workshop supplements the 5th Annual Conference on Precision Psychiatry: Innovation to Implementation. Space is limited, and the cost is in addition to the registration fees to attend the conference.
This workshop is NOT for CME.
This workshop is NOT eligible for further discounting than what is indicated above.