In this paper, we present Hetero-Rec, a framework for optimal deployment of embeddings for faster inference of recommendation model. The main idea is to cache frequently accessed embeddings on faster memories to reduce average latency during inference.
Hetero-Rec, a framework for optimal deployment of embeddings for faster inference of recommendation model, uses performance model-based optimization ...
Dec 9, 2024 · To capture the directed edges between nodes, existing methods mostly learn two embedding vectors for each node, source vector and target vector.
This work presents Hetro-Rec++ with heuristic based pre-optimizer and advanced formulation of optimizer's cost function, and demonstrates its effectiveness ...
Jun 2, 2023 · Bibliographic details on Hetero-Rec: Optimal Deployment of Embeddings for High-Speed Recommendations.
Best Paper Award. Hetero-Rec: Optimal Deployment of Embeddings for High-Speed Recommendations Chinmay N Mahajan (TCS Research); Ashwin Krishnan (TCS Research); ...
Hetero-Rec: Optimal Deployment of Embeddings for High-Speed Recommendations. AIMLSystems 2022: 11:1-11:9. [c45]. view. electronic edition via DOI; unpaywalled ...
Hetero-Rec: Optimal Deployment of Embeddings for High-Speed Recommendations. 2022-10-12 | Conference paper. DOI: 10.1145/3564121.3564134. Contributors: Chinmay ...
Hetero-Rec: Optimal Deployment of Embeddings for High-Speed Recommendations. Conference Paper. May 2023. Chinmay Mahajan · Ashwin Krishnan · Manoj ...
Abstract—Recommendation models rely on deep learning net- works and large embedding tables, resulting in computationally and memory-intensive processes.