Hello LTR
summary
A set of Jupyter notebooks and a client library that take TMDB or MS MARCO data through feature logging, training, deployment and search on Solr/Elasticsearch using Docker Compose and Jupyter. It trains a linear model with SVM rank on features like BM25 title and overview, returns learned weights (title weight dominates), uploads a Solr linear model with per-feature mean/std (title mean 2.65 std 4.34), and shows the model works but is biased toward title matches.
AI-written summary · about
chapters (5)
- 0:00Introduction to Hello LTR
- 1:13Project overview and setup
- 3:00Purpose of Hello LTR
- 4:03Learning to Rank workflow demo
- 11:14Conclusion and Q&A
related talks
46mJump-Start Your Data Science Learning Java & Jupyter Notebooks
57mGood Practices for Applied Machine Learning
24mLearning "Learning to Rank"
31mWhen humans teach machines
01h 30mBecome a data-driven Organization with Machine Learning- 57mGood Practices for Applied Machine Learning
45mData Science Apps: Beyond Notebooks
32mHow to be best in conveying the message by using NLP approach (Big Five modeling)