CHOLAN: A Modular Approach for Neural Entity Linking on Wikipedia and Wikidata
-
Updated
Jan 21, 2022 - Python
CHOLAN: A Modular Approach for Neural Entity Linking on Wikipedia and Wikidata
A vector database for querying meaningfully similar data.
Combining Linking Techniques (CLiT) is an entity linking combination and execution framework, allowing for the seamless integration of EL systems and result exploitation for the sake of system reusability, result reproducibility, analysis and continuous improvement. (We hate waste. Especially wasting time. So let's reuse instead!)
Offline Yelp ranking stack with route-aware recall, structured XGBoost rerank, and reward-model rescue reranking.
A scalable two-stage news recommender that retrieves relevant candidates and reranks them using hybrid lexical and semantic features to optimize top-K recommendation quality.
Deterministic invoice extraction using native PDF text layers. No OCR nonsense, no brittle rules that break at scale, no vendor lock-in paying exorbitant prices for creative interpretations of financial documents. This is my battle, I pick this hill!
Two-stage recommender system: PyTorch Two-Tower retrieval + FAISS ANN candidate generation + XGBoost LambdaMART re-ranking, served via FastAPI with a Streamlit demo and MLflow experiment tracking.
Hybrid product recommender system combining SVD-based collaborative filtering, TF-IDF content-based filtering, candidate generation, offline evaluation, and an interactive Streamlit interface.
Large-scale e-commerce recommendation engine with memory-efficient co-visitation, leakage-safe evaluation, feature engineering, and LightGBM LambdaRank reranking.
A high-precision homolog screening and candidate prioritization tool integrating reciprocal BLAST evidence with interpretable expression ranking.
Production-ready recommender system suite: serving API, pipelines, algorithm SDK, and evaluation tooling.
A reproducible, resource-aware solution for the Kaggle OTTO multi-objective recommendation competition, featuring co-visitation retrieval, target-aware candidates, LambdaMART ranking, and neural retrieval experiments.
Zero-Click Search recommender for ONCF Voyages — XGBoost two-stage ranking, FastAPI, HR@1=76%. PFA internship UM5/ONCF.
Add a description, image, and links to the candidate-generation topic page so that developers can more easily learn about it.
To associate your repository with the candidate-generation topic, visit your repo's landing page and select "manage topics."