Update README.md
Browse files
README.md
CHANGED
@@ -1,20 +1,19 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
4 |
-
|
5 |
-
In this work, we trained a model to recognize 4 types of relationships between entities in drug review texts: ADR–Drugname, Drugname–Diseasename, Drugname–SourceInfoDrug, Diseasename–Indication. The input of the model is a review text and a pair of entities, between which it is required to determine the fact of a relationship and one of the 4 types of relationship, listed above.
|
6 |
|
7 |
Data
|
8 |
----
|
9 |
Proposed model is trained on a subset of 908 reviews of the [Russian Drug Review Corpus (RDRS)](https://arxiv.org/pdf/2105.00059.pdf). The subset contains the markup of the following types of entities: and contains pairs of entities marked with the 4 listed types of relationships:
|
10 |
- ADR-Drugname — the relationship between the drug and its side effects
|
11 |
-
- Drugname-SourceInfodrug — the relationship between the medication and
|
|
|
12 |
- Diseasename-Indication — the connection between the illness and its symptoms (e.g., “cough”, “fever 39 degrees”)
|
13 |
Also, this subset contains pairs of the same entity types between which there is no relationship: for example, a drug and an unrelated side effect that appeared after taking another drug; in other words, this side effect is related to another drug.
|
14 |
|
15 |
-
Model
|
16 |
----
|
17 |
-
Proposed model is based on the [XLM-RoBERTA-large](https://arxiv.org/abs/1911.02116) topology. After the additional training as a
|
18 |
|
19 |
How to use
|
20 |
----
|
|
|
1 |
+
pharm-relation-extraction
|
2 |
+
===
|
3 |
+
Model trained a model to recognize 4 types of relationships between significant pharmacological entities in russian-language reviews: ADR–Drugname, Drugname–Diseasename, Drugname–SourceInfoDrug, Diseasename–Indication. The input of the model is a review text and a pair of entities, between which it is required to determine the fact of a relationship and one of the 4 types of relationship, listed above.
|
|
|
|
|
4 |
|
5 |
Data
|
6 |
----
|
7 |
Proposed model is trained on a subset of 908 reviews of the [Russian Drug Review Corpus (RDRS)](https://arxiv.org/pdf/2105.00059.pdf). The subset contains the markup of the following types of entities: and contains pairs of entities marked with the 4 listed types of relationships:
|
8 |
- ADR-Drugname — the relationship between the drug and its side effects
|
9 |
+
- Drugname-SourceInfodrug — the relationship between the medication and the source of information about it (e.g., “was advised at the pharmacy”, e.g., “was advised at the pharmacy”, “the doctor recommended it”);
|
10 |
+
- Drugname-Diseasname — the relationship between the drug and the disease
|
11 |
- Diseasename-Indication — the connection between the illness and its symptoms (e.g., “cough”, “fever 39 degrees”)
|
12 |
Also, this subset contains pairs of the same entity types between which there is no relationship: for example, a drug and an unrelated side effect that appeared after taking another drug; in other words, this side effect is related to another drug.
|
13 |
|
14 |
+
Model topology and training
|
15 |
----
|
16 |
+
Proposed model is based on the [XLM-RoBERTA-large](https://arxiv.org/abs/1911.02116) topology. After the additional training as a language model on corpus of unmarked drug reviews, this model was trained as a classification model on 80% of the texts from subset of the corps described above.
|
17 |
|
18 |
How to use
|
19 |
----
|