Research topic
Learning predictive models and evaluating generalization without causal overclaiming.
Independent educational guide. Not affiliated with or endorsed by the universities, professors or laboratories described here. This collection reflects the material currently mapped on Socratic Learn.
Human Motor Control and Neuromodulation Laboratory: linking quantitative movement measurements with implanted subthalamic local field potentials and neuromodulation to understand Parkinson’s disease and develop personalized adaptive therapies.
5 lessons · ~27 minutes
Explore the labBerkeley RAIL research studies robot learning from demonstrations, reward, prior datasets, and interaction: policies, reinforcement learning, manipulation, and generalization. The independent Levine primer also covers his co-founder work at Physical Intelligence, a separate company.
7 lessons · ~37 minutes
Explore the labRecommended through the concepts linked to this topic. Each lesson belongs to an existing learning series.
Decode Life: AI & Computational Biology
Train, validation and test setsTraining fits model parameters; validation guides choices; a held-out test estimates performance on unseen cases.
Supports: Machine learning
Decoding and Restoring Movement in Parkinson’s Disease
AI, gait, and the future of personalized neuromodulationInterpret neural and wearable gait biomarkers within their measurement limits.
Supports: Machine learning
How Robots Learn
From programming robots to teaching robotsExplain a policy and the closed perception-action loop.
Supports: Machine learning
How Robots Learn
Imitation learning versus reinforcement learningExplain behavioral cloning and compounding error under distribution shift.
Supports: Machine learning
Decode Life: AI & Computational Biology
Regression and classificationA regression output is numerical, such as biomarker concentration; a classifier estimates a class or class probability.
Supports: Machine learning
No source-backed papers mapped here yet.