AI for Autism: Bridging Neuroscience and AI
summary
Claims a pipeline that achieves reliable early detection of autism from MRI scans using the Autism Brain Imaging Data Exchange and standard computer vision models, argues detection between 12–18 months greatly increases chance of independent life, and shows working implementations and demos on GitHub and Hugging Face using Keras and Keras Core. Presents a second pipeline converting speech to mel spectrograms with a 27 distinct emotions dataset for emotion classification, and a facial-emotion branch; prototypes exist, code runs with good accuracy and is research-only access.
AI-written summary · about
chapters (8)
- 0:00Introduction to AI for Autism
- 3:40Early autism detection with MRI scans
- 7:30Emotion detection challenges for autistic individuals
- 10:29AI solutions for emotion recognition
- 14:09AI in behavioral and social skills training
- 15:56Robotic therapy for autism support
- 17:21Open source collaboration and resources
- 17:41Q&A and community engagement
related talks
01h 03mUsing Machine Learning and AI to serve people and communities
44mAI in the Clinic: Prediction is Better Than Cure
44mThe age of applied AI
39mAI + JavaScript Rocks
42mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics
54mAI: Actual Intelligence
06mTransitioning AI from the lab to real-life applications
49mBuilding AI That Learns and Adapts: A Case Study in MRI Diagnostics