Trial2Vec

Wang, Zifeng and Sun, Jimeng. (2022). Trial2Vec: Zero-Shot Clinical Trial Document Similarity Search using Self-Supervision. Findings of EMNLP'22.

News

• 12/8/2022: Support download_embedding that obtains the pretrained embedding only. It saves a lot of GPU/CPU memory! Please refer this example for detailed use cases.

from trial2vec import download_embedding
t2v_emb = download_embedding()

• 10/27/2022: Support word_vector and sentence_vector!

# sentence vectors
inputs = ['I am a sentence', 'I am another sentence']
outputs = model.sentence_vector(inputs)
# torch.tensor w/ shape [2, 128]

# word vectors
inputs = ['I am a sentence', 'I am another sentence abcdefg xyz']
outputs = model.word_vector(inputs)
# {'word_embs': torch.tensor w/ shape [2, max_token, 128], 'mask': torch.tensor w/ shape [2, max_token]}

Usage

Get pretrained Trial2Vec model in three lines:

from trial2vec import Trial2Vec

model = Trial2Vec()

model.from_pretrained()

A jupyter example is shown at https://github.com/RyanWangZf/Trial2Vec/blob/main/example/demo_trial2vec.ipynb.

How to install

Install the correct PyTorch version by referring to https://pytorch.org/get-started/locally/.

Then install Trial2Vec by

# Recommended because it is update to date, small bugs will be kept fixed
pip install git+https://github.com/RyanWangZf/Trial2Vec.git

or

pip install trial2vec

Search similar trials

Use Trial2Vec to search similar clinical trials:

# load demo data
from trial2vec import load_demo_data
data = load_demo_data()

# contains trial documents
test_data = {'x': data['x']} 

# make prediction
pred = model.predict(test_data)

Encode trials

Use Trial2Vec to encode clinical trial documents:

test_data = {'x': df} # contains trial documents

emb = model.encode(test_data) # make inference

# or just find the pre-encoded trial documents
emb = [model[nct_id] for test_data['x']['nct_id']]

Continue training

One can continue to train the pretrained models on new trials as

# just formulate trial documents as the format of `data`
data = load_demo_data()

model.fit(
    {
    'x':data['x'], # document dataframe
    'fields':data['fields'], # attribute field columns
    'ctx_fields':data['ctx_fields'], # context field columns
    'tag': data['tag'], # nct_id is the unique tag for each trial
    },
    valid_data={
            'x':data['x_val'],
            'y':data['y_val']
        },
)

# save
model.save_model('./finetuned-trial2vec')