Proposal to add the MedSTS dataset

[FremyCompany]

Adding a Dataset

• Name: MedSTS
• Description: 1,068 sentence pairs annotated by two medical experts with semantic similarity scores of 0-5 (low to high similarity).
• Task: STS
• Paper: https://arxiv.org/abs/1808.09397
• Data: (must be asked by email to Mayo Clinic)
• License: (probably depends on agreement with Mayo Clinic)
• Motivation: one of the largest clinical text similarity dataset

[hakunanatasha]

@FremyCompany I recognized you proposed this, will you consider implementing it?