Code and data for "Universal Approximation Functions for Fast Learning to Rank: Replacing Expensive Regression Forests with Simple Feed-Forward Networks"
-
Updated
Jul 11, 2018 - Python
Code and data for "Universal Approximation Functions for Fast Learning to Rank: Replacing Expensive Regression Forests with Simple Feed-Forward Networks"
Dataset and experiments from the CIKM 2020 Resource Track
Learning-to-Rank search engine combining BM25 retrieval with LambdaMART (LightGBM) for relevance-optimized document ranking. Achieves NDCG@10 of 0.9449 — a 6% improvement over baseline. Includes feature engineering pipeline and interactive Streamlit search UI.
UCL COMP0084 Information Retrieval and Data Mining (2023/24)
Learning-to-Rank system for e-commerce product search using LambdaMART, MiniLM, TF-IDF, and marketplace-specific relevance features.
Ranklib for .NET is an open source learning to rank library
Reciprocal ranking in a two-sided market: LambdaMART over early interaction signals, evaluated with NDCG@K and wave-based splits to prevent leakage.
Bachelor Thesis 2024 - Automatic Identification of Duplicate Questions in Indonesian Consumer Health Forums Using Learning-to-Rank
A personalised news ranking system that learns and adapts to each user's preferences through interaction feedback such as clicks, dwell-time, likes, shares, and bookmarks.
Learning-to-rank search for short-term rental listings, built on public Inside Airbnb data and deployed on Azure ML.
Permission-aware hybrid search engine with retrieval-time ACL filtering, BM25, dense vector search, and learned ranking
An AI-native recruitment pipeline built for the IndiaRuns Hackathon. Features multi-threaded offline sentence-embeddings, graph-based career velocity mapping, proactive honeypot sweeping, and dynamic query-grouped LambdaMART ranking trees to evaluate 100,000 candidates with strict deterministic fairness.
A command line tool for training and evaluating ranking models using LightGBM and FastTree
Two-stage book recommender over 103K titles: multi-source retrieval (ALS, item-kNN, BM25, sentence embeddings) → LightGBM LambdaMART ranking. Temporal split, cold-start routing, FastAPI + Docker, GCP/Vertex scaffolding.
To associate your repository with the lambdamart topic, visit your repo's landing page and select "manage topics."