Artificial Intelligence¶
The shared foundation of AI: learning principles, representations and architectures; generation, perception and decision-making; engineering and evaluation. Foundation models and embodied intelligence have their own sections; this one keeps the methods they share.
Start here¶
Knowledge¶
-
Foundations & Classical ML
Probability and optimisation, generalisation, data; linear, tree and ensemble methods.
-
Representation Learning
Embeddings and transfer; self-supervised, contrastive and masked learning.
-
Architectures
Trade-offs across MLP, CNN, Transformer, Mamba and graph networks.
-
Geometry & Generative Modelling
Manifolds and latent spaces; diffusion and flow matching.
-
Vision & 3D Space
Detection, segmentation, depth, point clouds and 3D reconstruction.
-
Search, Planning & Decisions
MDPs, reinforcement, imitation and offline learning.
-
Time Series & Scientific ML
Forecasting and anomaly detection; PINNs and neural operators.
-
Efficient Training & Engineering
Reproducibility, profiling, quantisation, distillation and deployment.
-
Evaluation & Trustworthy AI
Data quality, calibration, uncertainty and out-of-distribution behaviour.