fibber 0.0.1.dev0

Fibber is a benchmarking suite for adversarial attacks on text classification.

An open source project from Data to AI Lab at MIT.

Fibber

Fibber is a library to evaluate different strategies to paraphrase natural language, especially how these strategies can break text classifiers without changing the meaning of a sentence.

• Documentation: https://DAI-Lab.github.io/fibber
• GitHub: https://github.com/DAI-Lab/fibber

Overview

Fibber is a library to evaluate different strategies to paraphrase natural language. In this library, we have several built-in paraphrasing strategies. We also have a benchmark framework to evaluate the quality of paraphrase. In particular, we use the GPT2 language model to measure how meaningful is the paraphrased text. We use a universal sentence encoder to evaluate the semantic similarity between original and paraphrased text. We also train a BERT classifier on the original dataset, and check of paraphrased sentences can break the text classifier.

Install

Requirements

fibber has been developed and tested on Python 3.6, 3.7 and 3.8

Also, although it is not strictly required, the usage of conda is highly recommended to avoid interfering with other software installed in the system in which fibber is run.

These are the minimum commands needed to create a conda environment using python3.6 for fibber:

# First you should install conda.
conda create -n fibber_env python=3.6

Afterward, you have to execute this command to activate the environment:

conda activate fibber_env

Then you should install tensorflow and pytorch. Please follow the instructions for tensorflow and pytorch. Fibber requires tensorflow>=2.0.0 and pytorch>=1.5.0.

Remember to execute conda activate fibber_env every time you start a new console to work on fibber!

Install from PyPI

After creating the conda environment and activating it, we recommend using pip in order to install fibber:

pip install fibber

This will pull and install the latest stable release from PyPI.

Use without install

If you are using this project for research purpose and want to make changes to the code, you can install all requirements by

git clone [e-mail]:DAI-Lab/fibber.git
cd fibber
pip install --requirement requirement.txt

Then you can use fibber by

python -m fibber.datasets.download_datasets
python -m fibber.benchmark.benchmark

In this case, any changes you made on the code will take effect immediately.

Install from source

With your conda environment activated, you can clone the repository and install it from source by running make install on the stable branch:

git clone [e-mail]:DAI-Lab/fibber.git
cd fibber
git checkout stable
make install

Quickstart

In this short tutorial, we will guide you through a series of steps that will help you getting started with fibber.

(1) Install Fibber

(2) Get a demo dataset.

from fibber.datasets import get_demo_dataset

trainset, testset = get_demo_dataset()	

(3) Create a Fibber object.

from fibber.fibber import Fibber

arg_dict = {
    "use_gpu_id": 0,
    "gpt2_gpu_id": 0,
    "strategy_gpu_id": 0,
}

fibber = Fibber(arg_dict, dataset_name="demo", strategy_name="RandomStrategy", 
                trainset=trainset, testset=testset)

(4) Randomly sample a sentence from the test set, and paraphrase it.

The following command can randomly paraphrase the sentence into 5 different ways.

fibber.paraphrase_a_random_sentence(n=5)

The output is a tuple of (str, list, list).

# Original Text
'the movie slides downhill as soon as macho action conventions assert themselves .'

# 5 paraphrases
 ['conventions slides as as action assert macho downhill soon movie . the themselves',
  'as . downhill action macho the themselves assert as slides conventions soon movie',
  'movie as slides macho action . soon themselves the downhill as assert conventions',
  'the soon assert as movie themselves macho conventions as downhill . action slides',
  'downhill movie conventions slides the assert themselves action macho as as . soon'],

# Evaluation metrics of these 5 paraphrases.
 [{'EditingDistance': 8,
   'USESemanticSimilarity': 0.8859144449234009,
   'GloVeSemanticSimilarity': 1.0000000321979126,
   'GPT2GrammarQuality': 23.059619903564453},
  {'EditingDistance': 9,
   'USESemanticSimilarity': 0.8609699010848999,
   'GloVeSemanticSimilarity': 1.0000000321979126,
   'GPT2GrammarQuality': 39.824188232421875},
  {'EditingDistance': 8,
   'USESemanticSimilarity': 0.8530778288841248,
   'GloVeSemanticSimilarity': 1.0000000321979126,
   'GPT2GrammarQuality': 17.592607498168945},
  {'EditingDistance': 9,
   'USESemanticSimilarity': 0.8957847356796265,
   'GloVeSemanticSimilarity': 1.0000000321979126,
   'GPT2GrammarQuality': 24.76700210571289},
  {'EditingDistance': 9,
   'USESemanticSimilarity': 0.9004875421524048,
   'GloVeSemanticSimilarity': 1.0000000321979126,
   'GPT2GrammarQuality': 11.36586856842041}]

(5) You can also ask fibber to paraphrase your sentence.

fibber.paraphrase({"text0": "This movie is fantastic"}, "text0", 5)

Supported strategies

In this version, we implement three strategies

• IdentityStrategy:
  • The identity strategy outputs the original text as its paraphrase.
  • This strategy generates exactly 1 paraphrase for each original text regardless of --num_paraphrases_per_text flag.
• RandomStrategy:
  • The random strategy outputs the random shuffle of words in the original text.

What's next?

For more details about fibber and all its possibilities and features, please check the documentation site.

History