Glossary · E-commerce ML

Cold Start Problem

Definition

The cold-start problem describes recommendation and ranking systems' inability to serve new users or new items with no interaction history. Few-shot learning, meta-learning (MAML), and prototypical networks address it by learning initializations that adapt quickly from sparse signals.

Collaborative filtering fails for items with zero interactions. Remedies fall into three buckets: content-based fallbacks (use item features), transfer learning (borrow from similar items), and meta-learning (learn an initialization that adapts fast to new items with K interactions). MAML and Prototypical Networks are the canonical few-shot approaches in recommendation.

Essays on this concept