|
--- |
|
language: |
|
- en |
|
pipeline_tag: token-classification |
|
widget: |
|
- text: "Let P be a G-poset. The strong compatibility graph of P, denoted by C_P, is the graph C_P with vertex set P, and two elements x, y∈ P are adjacent if there is an element g∈ G∖{e} such that x and g· y are comparable in P and y∉ [x], where [x]={g· x : g∈ G}." |
|
example_title: "Strong compatibility graph" |
|
- text: "A simplicial map f : X → Y between simplicial complexes X and Y is a map which sends vertices to vertices, and whenever vertices v_0, ..., v_k∈ X span a simplex σ of X then their images span a simplex τ of Y and we have f(σ) = τ. Therefore a simplicial map is determined by its values on the vertex set of X. A simplicial map is nondegenerate if it is injective on each simplex." |
|
example_title: "simplicial map; nondegenerate" |
|
- text: "A vertical strip is a skew shape (either partition or composition) whose diagram contains at most one cell per row. A horizontal strip is a skew shape whose diagram contains at most one cell per column." |
|
example_title: "vertical strip; horizontal strip" |
|
- text: "Permutations ω and π are C-equivalent , denoted ωπ, if ωQ∼π and (P(ω))=(P(π)).\n We denote the C-equivalence class of the permutation π by [π]_C. \n The rectified shape of [π]_C is (P(π)). \n Two SCT T and T' are C-equivalent T T' if they have the same skew shape and w_col(T) w_col(T'). \n We denote the C-equivalence class of T by [T]_C. \n The rectified shape of [T]_C is (T). \n We say that [T]_C is complete if \n { w_col(T') : T'∈ [T]_C } = [w_col(T)]_C." |
|
example_title: "C-equivalent; rectified shape; complete" |
|
- text: "A Young graph Y such that Y≃ Y(10, 9) is called a 1089 graph." |
|
example_title: "1089 graph" |
|
- text: "A subset 𝒢 is a Gröbner basis for I if the leading term of each member of I is divided by the leading term of a member of 𝒢. That is, 𝒢 is a Gröbner basis if in_≺(I)=LT_≺(g):g∈𝒢." |
|
example_title: "Gröbner basis" |
|
--- |
|
### How to use |
|
|
|
You can use this model directly with a pipeline for token classification. |
|
Given a mathematical definition from the text of an academic article, the model shall identify the term(s) defined with it. |
|
```python |
|
from transformers import pipeline |
|
# Here, we use the prediction of the first word piece of every word. |
|
pipe = pipeline(model="InriaValda/roberta-base_definiendum", |
|
aggregation_strategy="first") |
|
pipe("A Young graph Y such that Y≃ Y(10, 9) is called a 1089 graph.") |
|
|
|
``` |
|
|
|
### Our Paper |
|
|
|
[Extracting Definienda in Mathematical Scholarly Articles with Transformers](https://arxiv.org/abs/2311.12448) (Accepted by 2nd WIESP @ IJCNLP-AACL 2023) |
|
|
|
### An 8 minutes' video about our work |
|
|
|
https://youtu.be/tUioJooDDio?si=6EnTN_5l-9t86IKk |
