EMNLP 2022 Papers

Visually explore all conference papers. Embeddings created with SciNCL. 2D reduction with UMAP.

Huggingface Space: https://huggingface.co/spaces/malteos/emnlp2022-papers

Download data

wget https://2022.emnlp.org/downloads/Accepted-Papers-20221027.xls

Install dependecies

pip install -r requirements.txt

Generate embeddings and reduce to 2D

export CUDA_VISIBLE_DEVICES=0

python embed_papers.py --input_path ./Accepted-Papers-20221027.xls \
    --json_output_path ./papers.json --js_output_path ./papers.js \
    --model_name_or_path /data/datasets/huggingface_transformers/pytorch/scincl --limit 10

View Web page

# from local FS
open index.html

# via local web server at http://localhost
python -m http.server 80