Training Dynamics
Training Dynamics groups the core ideas a learner needs at the 7 · modelling & training stage. Work through the lessons in order when new to the area, or use them independently as a reference when implementing an analysis.
Learn the mechanism one decision at a time
Work through the lessons in order if the topic is new. If you already know the basics, open the specific leaf lesson that matches the operation, diagnostic or failure mode you need.
1Definition→
2Mechanism→
3Example→
4Diagnostic→
5Decision
Loss functionsLoss functions is an evaluation quantity that compresses a particular aspect of predictive behaviour into a number. Its usefulness depends on whether that aspect matches the real decision cost, class prevalence and intended model output.
02Gradient descentGradient descent is a practical concept within Training Dynamics. It helps turn the broader workflow stage “7 · Modelling & Training” into an explicit analytical decision that can be explained, implemented and checked. The concept should be understood in terms of purpose, mechanism, assumptions, evidence and downstream consequences.
03Learning rateLearning rate is a practical concept within Training Dynamics. It helps turn the broader workflow stage “7 · Modelling & Training” into an explicit analytical decision that can be explained, implemented and checked. The concept should be understood in terms of purpose, mechanism, assumptions, evidence and downstream consequences.
04Batch sizeBatch size belongs to the operational phase where an analytical result becomes a maintained system. Production quality requires the data contract, preprocessing, model, decision logic and monitoring to remain consistent over time.
05Epochs and iterationsEpochs and iterations is a practical concept within Training Dynamics. It helps turn the broader workflow stage “7 · Modelling & Training” into an explicit analytical decision that can be explained, implemented and checked. The concept should be understood in terms of purpose, mechanism, assumptions, evidence and downstream consequences.
06ConvergenceConvergence is a practical concept within Training Dynamics. It helps turn the broader workflow stage “7 · Modelling & Training” into an explicit analytical decision that can be explained, implemented and checked. The concept should be understood in terms of purpose, mechanism, assumptions, evidence and downstream consequences.
07RegularisationRegularisation is a practical concept within Training Dynamics. It helps turn the broader workflow stage “7 · Modelling & Training” into an explicit analytical decision that can be explained, implemented and checked. The concept should be understood in terms of purpose, mechanism, assumptions, evidence and downstream consequences.
08Early stoppingEarly stopping is a practical concept within Training Dynamics. It helps turn the broader workflow stage “7 · Modelling & Training” into an explicit analytical decision that can be explained, implemented and checked. The concept should be understood in terms of purpose, mechanism, assumptions, evidence and downstream consequences.