idopinto/ts-aspire-biomed-specter2

Model Card for Model ID

tsAspire model included in a paper for modeling fine-grained similarity between documents in the biomedical domain fine-tuned from Specter2-base.

Title: "Multi-Vector Models with Textual Guidance for Fine-Grained Scientific Document Similarity"

Authors: Sheshera Mysore, Arman Cohan, Tom Hope

Paper: https://arxiv.org/abs/2111.08366

Github: https://github.com/allenai/aspire

Note: In the context of the paper, this model is referred to as tsAspire and represents the papers proposed multi-vector model for fine-grained scientific document similarity.

Refer to https://huggingface.co/allenai/aspire-contextualsentence-singlem-biomed for more details and usage.

Base Model: https://huggingface.co/allenai/specter2_base

Evaluation Results

Differences might be in co-citations training data which constructed from scratch from different release of S2ORC (originaly, 2019-09-28, which I didn't have access to.)

Model	Specification	TRECCOVID-MAP	TRECCOVID-NDCG%20	RELISH-MAP	RELISH-NDCG%20
aspire-contextualsentence-singlem-biomed	TSASPIRE_spec_orig	26.68	57.21	61.06	77.20
ts-aspire-biomed-recon	TSASPIRE_Spec	29.26	60.45	62.2	78.7
ts-aspire-biomed-specter2	TSASPIRE_Spec2	31.16	62.43	63.24	79.89