Topics
A controlled vocabulary rather than a tag cloud: every piece of work is filed under the methods it uses, the domain it applies to, and the tools it was built with.
Methods
Techniques and model families.Supervised and unsupervised learning, model selection, evaluation.
Deep learning 5Neural architectures, training dynamics, representation learning.
Reinforcement learning 5Agents learning control policies from interaction and reward.
Multi-agent systems 3Coordination, task allocation and load-balancing across many agents.
Natural language processing 2Transformers, sentiment analysis, text representation.
Computer vision 3Image understanding, segmentation, landmark detection.
Signal processing 3Time, frequency and time–frequency characterisation of signals.
Statistics 4Inference, experimental design, statistical modelling.
Regression 2Linear, piecewise and generalised regression models.
Clustering 2Unsupervised partitioning and mixture models.
Dimensionality reduction 2Manifold learning, UMAP, autoencoders as compression.
Time series 1Forecasting and sequential structure in data.
Optimization 1Search, allocation and numerical optimisation.
Algorithms & complexity 1Algorithm design and complexity analysis.
Domains
Fields the work is applied to.Mechanical physics, vibration, fatigue and damage.
Remote sensing 2UAV imagery, LiDAR, point clouds, geospatial data.
Robotics 1Simulated and physical agents acting in an environment.
Cybersecurity 3Secure systems, memory safety, exploitation and defence.
Health data 3Clinical and epidemiological data analysis.
Finance & geopolitics 2Quantitative finance and the political economy of capital.
AI governance 2Regulation, oversight and institutional design for artificial intelligence.
Cloud systems 1Cloud cost, commitment and capacity modelling.
Software engineering 1Design, architecture and delivery of working software.
Mathematics 5Differential calculus, dynamical systems, numerical analysis.