Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

https://github.com/decile-team/spear

SPEAR: Programmatically label and build training data quickly.
https://github.com/decile-team/spear

data-efficient-learning data-programming deep-learning labeling-functions machine-learning semi-supervised-learning unsupervised-learning weak-supervision

Last synced: about 2 months ago
JSON representation

SPEAR: Programmatically label and build training data quickly.

Host: GitHub
URL: https://github.com/decile-team/spear
Owner: decile-team
License: mit
Created: 2020-12-28T06:26:48.000Z (over 3 years ago)
Default Branch: main
Last Pushed: 2024-03-08T18:20:08.000Z (4 months ago)
Last Synced: 2024-05-09T05:19:25.811Z (2 months ago)
Topics: data-efficient-learning, data-programming, deep-learning, labeling-functions, machine-learning, semi-supervised-learning, unsupervised-learning, weak-supervision
Language: Jupyter Notebook
Homepage: https://spear-decile.readthedocs.io/
Size: 432 MB
Stars: 103
Watchers: 10
Forks: 20
Open Issues: 1
Metadata Files:
- Readme: README.md
- License: LICENSE.txt

Lists

awesome-stars - decile-team/spear - SPEAR: Programmatically label and build training data quickly. (Jupyter Notebook)

README

        

[![Lines of code](https://img.shields.io/tokei/lines/github/decile-team/spear)]()

![visitors](https://visitor-badge.glitch.me/badge?page_id=decile-team/spear)

![PyPI](https://img.shields.io/pypi/v/spear)

[![docs](https://readthedocs.org/projects/spear-decile/badge)](https://spear-decile.readthedocs.io/)

[![license](https://img.shields.io/badge/License-MIT-blue.svg)](https://github.com/decile-team/spear/blob/main/LICENSE.txt)

[![website](https://img.shields.io/badge/website-online-green)](https://decile.org/)



    


        

    



## Semi-Supervised Data Programming for Data Efficient Machine Learning

SPEAR is a library for data programming with semi-supervision. The package implements several recent data programming approaches including facility to programmatically label and build training data.

### Pipeline

* Design Labeling functions(LFs)

* generate pickle file containing labels by passing raw data to LFs

* Use one of the Label Aggregators(LA) to get final labels



    


        

    



### SPEAR provides functionality such as 

* development of LFs/rules/heuristics for quick labeling

* compare against several data programming approaches

* compare against semi-supervised data programming approaches

* use subset selection to make best use of the annotation efforts

* facility to store and save data in [pickle file](https://spear-decile.readthedocs.io/en/latest/index.html#spear.utils.data_editor.get_data)

#### Labelling Functions (LFs)

* discrete LFs - Users can define LFs that return discrete labels

* continuous LFs - return continuous scores/confidence to the labels assigned

#### Approaches Implemented

You can read [this paper](https://arxiv.org/pdf/2008.09887.pdf) to know about below approaches

* Only-L 

* Learning to Reweight

* Posterior Regularization

* Imply Loss

* CAGE

* Joint Learning

Data folder for SMS & TREC can be found [here](https://drive.google.com/file/d/1CJZ73nNa7Ho0BOSDgGx9CRvXoepVSpet/view?usp=sharing). This folder needs to be placed in the same directory as notebooks folder is in, to run the notebooks or examples.

Direct download of the zip file can be done via wget using `gdown` library . 

```bash

pip install gdown

gdown 1CJZ73nNa7Ho0BOSDgGx9CRvXoepVSpet

```

## Installation

* Install Submodlib library

`pip install -i https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ submodlib`

#### Method 1

To install latest version of SPEAR package using PyPI:

 ```python 

 pip install decile-spear

 ```

#### Method 2    

SPEAR requires Python 3.6 or later. First install [submodlib](https://github.com/decile-team/submodlib#setup). Then install SPEAR:

```bash

git clone https://github.com/decile-team/spear.git

cd spear

pip install -r requirements/requirements.txt

```

## Citation

```bibtex

@inproceedings{abhishek-etal-2022-spear,

    title = "{SPEAR} : Semi-supervised Data Programming in Python",

    author = "Abhishek, Guttu  and

      Ingole, Harshad  and

      Laturia, Parth  and

      Dorna, Vineeth  and

      Maheshwari, Ayush  and

      Ramakrishnan, Ganesh  and

      Iyer, Rishabh",

    booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing: System Demonstrations",

    month = dec,

    year = "2022",

    address = "Abu Dhabi, UAE",

    publisher = "Association for Computational Linguistics",

    url = "https://aclanthology.org/2022.emnlp-demos.12",

    pages = "121--127",

}

```

### Quick Links

* [SPEAR tutorials](https://github.com/decile-team/spear/tree/main/notebooks)

* [SPEAR documentation](https://spear-decile.readthedocs.io/)

* [Demonstration of Cage and Joint Learning using SPEAR](https://youtu.be/qdukvO3B8YU)

* [Demonstration of Imply Loss, Learn2Reweight using SPEAR](https://youtu.be/SN9YYK4FlU0)

* SMS SPAM: [CAGE colab](https://colab.research.google.com/drive/1vec-Q-xO9wQtM3p_CZ7237gCq0xIR9b9?usp=sharing), [JL colab](https://colab.research.google.com/drive/1HqkqQ8ytWjP9on3du-vVB07IQvo8Li3W?usp=sharing)

* [DECILE website](https://decile.org)

* [SubModLib - Summarize massive datasets using submodular optimization](https://github.com/decile-team/submodlib)

* [DISTIL- Deep Diversified Interactive Learning](https://github.com/decile-team/distil)

* [CORDS- COResets and Data Subset Selection](https://github.com/decile-team/cords)

### Acknowledgment

SPEAR takes inspiration, builds upon, and uses pieces of code from several open source codebases. These include [Snorkel](https://github.com/snorkel-team/snorkel), [Snuba](https://github.com/HazyResearch/reef)  & [Imply Loss](https://github.com/awasthiabhijeet/Learning-From-Rules). Also, SPEAR uses [SUBMODLIB](https://github.com/decile-team/submodlib) for subset selection, which is provided by [DECILE](https://decile.org/) too.

### Team

SPEAR is created and maintained by [Ayush](https://www.cse.iitb.ac.in/~ayusham), [Abhishek](https://www.linkedin.com/in/guttu-sai-abhishek/), [Vineeth](https://www.cse.iitb.ac.in/~vineethdorna/), [Harshad](https://www.cse.iitb.ac.in/~harshadingole/), [Parth](https://www.cse.iitb.ac.in/~parthlaturia/), [Pankaj](https://www.linkedin.com/in/pankaj-singh-b000894a/), [Rishabh Iyer](https://www.rishiyer.com), and [Ganesh Ramakrishnan](https://www.cse.iitb.ac.in/~ganesh/). We look forward to have SPEAR more community driven. Please use it and contribute to it for your research, and feel free to use it for your commercial projects. We will add the major contributors here.

## Publications

[1] Abhishek et al. [SPEAR : Semi-supervised Data Programming in Python](https://arxiv.org/abs/2108.00373), Demonstration Paper.

[2] Maheshwari et al. [Learning to Robustly Aggregate Labeling Functions for Semi-supervised Data Programming](https://arxiv.org/abs/2109.11410), In Findings of ACL (Long Paper) 2022.

[3] Maheshwari, Ayush, et al. [Data Programming using Semi-Supervision and Subset Selection](https://arxiv.org/abs/2008.09887), In Findings of ACL (Long Paper) 2021.

[4] Chatterjee, Oishik, Ganesh Ramakrishnan, and Sunita Sarawagi. [Data Programming using Continuous and Quality-Guided Labeling Functions](https://arxiv.org/abs/1911.09860), In AAAI 2020.

[5] Sahay, Atul, et al. [Rule augmented unsupervised constituency parsing](https://arxiv.org/abs/2105.10193), In Findings of ACL (Short Paper) 2021.