HannaAbiAkl commited on
Commit
d489979
1 Parent(s): ddc2008

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +2 -0
README.md CHANGED
@@ -19,6 +19,8 @@ tags:
19
 
20
  We present FinISH, a [SRoBERTa](https://huggingface.co/sentence-transformers/nli-roberta-base-v2) base model fine-tuned on the [FIBO ontology](https://spec.edmcouncil.org/fibo/) dataset for domain-specific representation learning on the [**Semantic Search**](https://www.sbert.net/examples/applications/semantic-search/README.html) downstream task.
21
 
 
 
22
  ## SRoBERTa Model Architecture
23
  Sentence-RoBERTa (SRoBERTa) is a modification of the pretrained RoBERTa network that uses siamese and triplet network structures to derive semantically meaningful sentence embeddings that can be compared using cosine-similarity. This reduces the effort for finding the most similar pair from 65 hours with RoBERTa to about 5 seconds with SRoBERTa, while maintaining the accuracy from RoBERTa. SRoBERTa has been evaluated on common STS tasks and transfer learning tasks, where it outperforms other state-of-the-art sentence embeddings methods.
24
 
 
19
 
20
  We present FinISH, a [SRoBERTa](https://huggingface.co/sentence-transformers/nli-roberta-base-v2) base model fine-tuned on the [FIBO ontology](https://spec.edmcouncil.org/fibo/) dataset for domain-specific representation learning on the [**Semantic Search**](https://www.sbert.net/examples/applications/semantic-search/README.html) downstream task.
21
 
22
+ The model is an implementation of the following paper: [Yseop at FinSim-3 Shared Task 2021: Specializing Financial Domain Learning with Phrase Representations](https://arxiv.org/abs/2108.09485)
23
+
24
  ## SRoBERTa Model Architecture
25
  Sentence-RoBERTa (SRoBERTa) is a modification of the pretrained RoBERTa network that uses siamese and triplet network structures to derive semantically meaningful sentence embeddings that can be compared using cosine-similarity. This reduces the effort for finding the most similar pair from 65 hours with RoBERTa to about 5 seconds with SRoBERTa, while maintaining the accuracy from RoBERTa. SRoBERTa has been evaluated on common STS tasks and transfer learning tasks, where it outperforms other state-of-the-art sentence embeddings methods.
26