Open-domain question answering requires retrieval systems able to cope with the diverse and varied nature of questions, providing accurate answers across a broad spectrum of query types and topics. To deal with such topic heterogeneity through a unique model, we propose DESIRE-ME, a neural information retrieval model that leverages the Mixture-of-Experts framework to combine multiple specialized neural models. We rely on Wikipedia data to train an effective neural gating mechanism that classifies the incoming query and that weighs the predictions of the different domain-specific experts correspondingly. This allows DESIRE-ME to specialize adaptively in multiple domains. Through extensive experiments on publicly available datasets, we show that our proposal can effectively generalize domain-enhanced neural models. DESIRE-ME excels in handling open-domain questions adaptively, boosting by up to 12% in NDCG@10 and 22% in P@1, the underlying state-of-the-art dense retrieval model.
Kasela, P., Pasi, G., Perego, R., Tonellotto, N. (2024). DESIRE-ME: Domain-Enhanced Supervised Information Retrieval Using Mixture-of-Experts. In Advances in Information Retrieval 46th European Conference on Information Retrieval, ECIR 2024, Glasgow, UK, March 24–28, 2024, Proceedings, Part II (pp.111-125). Springer Science and Business Media Deutschland GmbH [10.1007/978-3-031-56060-6_8].
DESIRE-ME: Domain-Enhanced Supervised Information Retrieval Using Mixture-of-Experts
Kasela, Pranav
Primo
;Pasi, Gabriella;
2024
Abstract
Open-domain question answering requires retrieval systems able to cope with the diverse and varied nature of questions, providing accurate answers across a broad spectrum of query types and topics. To deal with such topic heterogeneity through a unique model, we propose DESIRE-ME, a neural information retrieval model that leverages the Mixture-of-Experts framework to combine multiple specialized neural models. We rely on Wikipedia data to train an effective neural gating mechanism that classifies the incoming query and that weighs the predictions of the different domain-specific experts correspondingly. This allows DESIRE-ME to specialize adaptively in multiple domains. Through extensive experiments on publicly available datasets, we show that our proposal can effectively generalize domain-enhanced neural models. DESIRE-ME excels in handling open-domain questions adaptively, boosting by up to 12% in NDCG@10 and 22% in P@1, the underlying state-of-the-art dense retrieval model.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
