To Comprehend the New: On Measuring the Freshness of a Document by Tirthankar Ghosal, Abhishek Shukla, Asif Ekbal and Pushpak Bhattacharyya accepted as a full paper in the 37th International Joint Conference on Neural Networks (IJCNN 2019) to be held at Budapest, Hungary (CORE rank A/H-Index: 31).
PyTorch (for training the sentence encoder and inferring sentence embeddings)
Keras
Numpy
Pandas
Description of important files in each directory
input_generation.py Produces pre-trained sentence embeddings for dlnd data also produces document matrix based on sentence embeddings for input to CNN.
model.py This is the main CNN program.It creates the output file which has the predictions for each target and source document pair.
NER_Result.py This file is for Named-Entities similarity scoring for Phase-I (Search and Retrieval) in Premise Selection
wmd_from_ner.py This file is for Word Mover’s Distance for Phase-II (Recognition) in Premise Selection
TAPNew.zip This is training Data.
to-comprehend-the-new-on-measuring-the-freshness-of-a-document's People
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent