fmpr/CrowdLayer
A neural network layer that enables training of deep neural networks directly from crowdsourced labels (e.g. from Amazon Mechanical Turk) or, more generally, labels from multiple annotators with different biases and levels of expertise.
This tool helps data scientists and machine learning engineers train deep learning models using 'noisy' labels from multiple human annotators, like those gathered from crowdsourcing platforms. You input your raw data and the individual labels from each annotator; the tool outputs a more robust, trained model that accounts for varying annotator expertise and biases. This is for anyone building machine learning models that rely on human-generated labels, especially when perfect, single-source labels are unavailable.
No commits in the last 6 months.
Use this if you need to train a deep neural network but only have access to labels provided by multiple, potentially inconsistent, human annotators.
Not ideal if your dataset already has a single, definitive 'ground truth' label for each item, or if you are not working with deep neural networks.
Stars
69
Forks
20
Language
Jupyter Notebook
License
GPL-3.0
Category
Last pushed
Dec 13, 2021
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/fmpr/CrowdLayer"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
charles9n/bert-sklearn
a sklearn wrapper for Google's BERT model
jidasheng/bi-lstm-crf
A PyTorch implementation of the BI-LSTM-CRF model.
howl-anderson/seq2annotation
基于 TensorFlow & PaddlePaddle 的通用序列标注算法库（目前包含 BiLSTM+CRF, Stacked-BiLSTM+CRF 和...
kamalkraj/BERT-NER
Pytorch-Named-Entity-Recognition-with-BERT
kamalkraj/Named-Entity-Recognition-with-Bidirectional-LSTM-CNNs
Named-Entity-Recognition-with-Bidirectional-LSTM-CNNs