README.md

metadata

pretty_name: PubMed Cognitive Control Abstracts
license:
  - cc-by-4.0
language:
  - en
multilinguality:
  - monolingual
task_categories:
  - text-classification
task_ids:
  - topic-classification
  - semantic-similarity-classification
size_categories:
  - 100K<n<1M
paperswithcode_id: linking-theories-and-methods-in-cognitive
inference: false
model-index:
  - name: cogtext-pubmed
    results: []
source_datasets:
  - original
language_creators:
  - found
  - expert-generated
configs:
  - pubmed
  - pubmed20pct
  - lexicon
  - pubmed_gp3ada

Dataset Description

• Homepage: ArXiv preprint
• Repository: Linking Theories and Methods of Cognitive Control
• Point of Contact: Morteza Ansarinia

We performed automated text analyses on a large body of scientific texts (385705 scientific abstracts) and created a joint representation of cognitive control tasks and constructs.

Abstracts were first mapped into an embedding space using GPT-3 and Top2Vec models. Document embeddings were then used to identify a task-construct graph embedding that grounds constructs on tasks and supports nuanced meaning of the constructs by taking advantage of constrained random walks in the graph.

CogText dataset contains a collection of PubMed abstracts, along with their GPT-3 embeddings and topic embeddings. See CogText on GitHub for the details and codes.

Citation

To cite the paper use the following entry:

@misc{cogtext2022,
  author = {Morteza Ansarinia and
            Paul Schrater and
            Pedro Cardoso-Leite},
  title = {Linking Theories and Methods in Cognitive Sciences via Joint Embedding of the Scientific Literature: The Example of Cognitive Control},
  year = {2022},
  url = {https://arxiv.org/abs/2203.11016}
}