Temporary repository for implementing tensor factorization algorithms on Apache Spark
Currently I am working on the following 3 algorithms:
- PARAFAC (parallel algorithms given here: GigaTensor, U. Kang's PhD thesis) Also check the HaTen2 paper that apparently improves upon GigaTensor.
- RESCAL RESCAL paper, M. Nickel's PhD thesis (Spark-based distributed algorithm will be designed for this)
- HolE Holographic Embeddings of Knowledge Graphs (Spark-based distributed algorithm will be designed for this)
This will be divided across the Spark-RDF (interface, I/O, storage) and Spark-Sem-ML (algorithm) repositories eventually.