LeaPI

LeaPI is a zero-shot leanring method to automatically Learn a classifier from human need descriptions with Prototypical Instantiation. It first automatically generates weak labels by instantiating high-level concepts with prototypical instances and then trains a human needs classifier with the weakly labeled data. To filter noisy concepts, it uses a reinforced selection algorithm to choose high-quality concepts for instantiation.

For more details, check out our paper Learning to Classify Events from Human Needs Category Descriptions.

Install

conda create -n leapi -y python=3.7

conda activate leapi

./setup.sh

Prepare data

• GoogleNews-vectors-negative300.bin Download and put it in resources
• probase-data-concept-instance-relations.txt Download Probase and put it in resources
• event-dict.gz Download event dictionary developed by Ding&Riloff Download link
• event-vecs.bin Download event embeddings. An event embeddings is the average of embeddings of words in the event. Download link.
• SemEval2015-English-Twitter-Lexicon.pos.json Download twitter english sentiment lexicon.

Run the code

• First cd leapi
• Run the following command python main_learn.py --rl-max-epoch 200 --rl-sample-times 30 --rl-agent-lr 1e-3 --state-feat prodlabel --none-size 300 --wvfpath ../resources/GoogleNews-vectors-negative300.bin --seed 111 --rl-init default

Extract candidate concepts

The resources directory contains a pre-extracted candidate concepts. The user can also extract their own candidate concepts using the following scripts.

• Install and start redis server ''' >> redis-server '''

• Load Probase to redis

 cd leapi/prepare-data/
 python probase-redis-server.py

 python create_event_dict_redis_server.py

• Extract candidate concepts

    cd leapi/concepts
    python generate_concept_candidates.py