The LIGER Hybrid Model: Transforming Sequential Recommendation Systems
from Andrea Viliotti Hosted by Andrea Viliotti Independent AI Strategy Consultant & Researcher | Author of GDE
The episode delves into LIGER, a hybrid model designed for sequential recommendations that combines dense and generative retrieval techniques. Dense retrieval ensures high accuracy but comes with significant computational costs, while generative retrieval is more efficient but less precise, particularly with new items ("cold-start"). LIGER addresses these limitations by enhancing the performance of generative retrieval, especially in handling novel items, through the integration of dense ranking. This hybrid approach allows businesses to balance accuracy and scalability, making it particularly suitable for constantly evolving catalogs. Tests conducted on various datasets highlight LIGER's effectiveness.
Embed this episode
No transcript yet.