Proxy A-Distance

This is an implementation of an algorithm discussed in Ganin et. al (2015), Glorot et. al (2011), and Ben-David et. al (2007). It has been adapted for use with machine translation datasets, and released to the public under the MIT license.

This algorithm computes the Proxy A-Distance (PAD) between two domain distributions. PAD is a measure of similarity between datasets from different domains (e.g. newspapers and talk shows). Intuitively, similar domains => bigger error => smaller PAD. Dissimilar domains => smaller error => bigger PAD. The MAE error metric for binary classification between domains will bound PAD in the range [0, 2].

The algorithm is as follows:

1. Mix the two datasets. Apply label that indicate each example's origin.
2. Train a classifier on these merged data.
3. Measure the classifier's error e on a held-out test set.
4. Set PAD = 2 (1 − 2e)

We use a linear bag-of-words SVM for the underlying classifier.

Requirements

• numpy: pip install numpy
• sklearn: pip install sklearn

Usage

python main.py [corpusfile 1] [corpusfile 2] [vocab file]

• corpusfile 1 is a text file with one sentence per line.
• corpusfile 2 is another text file with one sentence per line.
• vocab is a text file with one token per line. These tokens represent a shared vocabulary for the above corpusfiles.

Example

python main.py test_data/europarl.en test_data/europarl.fr test_data/opensubtitles.en test_data/opensubtitles.fr test_data/vocab