The future of machine learning is not merely a progression towards larger models. It is a search for systems that learn from less supervision, design themselves more efficiently, operate closer to users and remain accountable when their decisions affect society.
Google Colab notebook: Open in Google Colab
This seminar exercise develops a recommender system in stages, from simple baselines to biased matrix factorisation, before asking whether SHAP explanations make individual predictions more understandable. The main lesson: better rating prediction did not automatically produce better top-10 ranking.