OSCAR
A novel query-dependent online soft compression method for RAG that reduces computational overhead while preserving performance. Unlike traditional hard compression methods, which shorten retrieved texts, or soft compression approaches, which map documents to continuous embeddings offline, OSCAR dynamically compresses retrieved information at inference time, eliminating storage overhead and enabling higher compression rates.
mRAG
Retrieval-augmented generation (RAG) in the multilingual setting (mRAG). Our findings highlight that despite the availability of high-quality off-the-shelf multilingual retrievers and generators, task-specific prompt engineering is needed to enable generation in user languages. Moreover, current evaluation metrics need adjustments for multilingual setting, to account for variations in spelling named entities.
Several releases: SPLADE V-2, SPLADE V-3, CoSPLADE etc.
SPLADE is sparse bi-encoder BERT-based model for effective and efficient first-stage ranking.
The expohedron is a polytope whose points represent all achievable exposures of items for a Position Based Model (PBM).
Code implementing the model introduced in Learning to Rank Images with Cross-Modal Graph Convolutions (ECIR’20).