Completed
Regression based pointwise ranking Input (4.x) feature vector responding to the query and a document, Label: y relevance of the document
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Deep Learning for Search and Recommender Systems in Practice
Automatically move to the next video in the Classroom when playback concludes
- 1 System Overview Document Retrieval Scoring and Ranking . Personalization and Re-ranking
- 2 Document Retrieval • Simple regex based retrieval . Traditional inverted index based retrieval Embedding based retrieval
- 3 Metrics for Evaluation • Multiple level of relevance NDCG (Normalized Discounted Cumulative Gain) . Binary relevance DMAP (Mean Average Precision) MRR Meon Reciprocal Ronk
- 4 Normalized Discounted Cumulative Gain Discounted Cumulative Goin
- 5 Mean Average Precision Precision: Relevant documents up to rank K/K
- 6 Mean Reciprocal Rank Reciprocal Rank
- 7 Learning to Rank
- 8 Pointwise Ranking Loss function is based on a single (query, document) pair
- 9 Regression based pointwise ranking Input (4.x) feature vector responding to the query and a document, Label: y relevance of the document
- 10 Classification based pointwise ranking
- 11 Ordinal regression based pointwise ranking
- 12 Summary of pointwise ranking Pros • Simple, considering one document at a time. • Available algorithms are rich. Most regression/classification algorithms can be used.
- 13 Pairwise Ranking Loss function is based on query and a pair of documents.
- 14 Listwise Ranking Loss function is based on the query and a list of documents
- 15 AdaRank Motivation: commonly used evaluation metrics are not differentiable. So it is not easy to optimize directly. AdaRank minimizes the exponential loss. El below can be NDCG.
- 16 List Net / ListMLE Map list of scores to a probability distribution by Plockett-Luce model. • Permutation probability, where 5() is the scoring function.
- 17 Summary of listwise ranking Pros
- 18 DeText: a Deep Learning Ranking Framework
