This repo contains research and production tooling for using Machine Learning to automate document annotation on ToS;DR
The process before Docbot:
- We crawl documents
- We wait for volunteers to submit Points, which are privacy policy quotations that highlight evidence that a Case (privacy-related statement) is true.
- We wait for curators, trusted volunteers, to approve or reject Points
- We score the service A-F based on the approved Points
The process after is the same, but now Docbot does the initial document analysis and submits Points, along with a confidence score, to curators. We achieve this by fine-tuning large language models into binary classifiers.
Each week an automated job runs 123 case models in total, analyzing any new documents that were added to ToS;DR's database since the last run.
If you would like to have some privacy policies or T&Cs analyzed with docbot, the best way is to add them to the ToS;DR platform on edit.tosdr.org and wait until our automated system picks them up.
If you have a particular need to run the models on standalone documents, or have any other questions, please get in touch with us at [email protected].
We welcome contributions to the engineering or research to improve our models.
We also plan to release our datasets used for training/evaluation, which could be of value to NLP researchers.
In a pip or conda environment, run:
pip install requirements.txt
python -m spacy download en_core_web_md
We create training corpora from database dumps. The first step is to convert from sql to pandas.
- Start postgresql (on a mac:
brew services restart postgresql
) - Run
createdb tosdr
- In the interactive REPL (
psql -d tosdr
), runcreate user phoenix with superuser; ALTER ROLE "phoenix" WITH LOGIN;
- Comment out any foreign key setups, found at the bottom of the sql files.
- Log into psql with
psql -d tosdr -U phoenix
\i services.sql
\i topics.sql
\i documents.sql
\i cases.sql
\i points.sql
- Confirm all 5 tables exist with
\dt public.*
(you might have to log out and back in). Tables can be inspected with i.e.\d+ public.topics
- Run
sql_to_pandas.py
to load the tables and save them as pickled pandas DataFrames indata
Run explore.ipynb
on the output of sql_to_pandas.py
to clean the data, and then see make_classification_datasets.py
to turn that into a classification
dataset apt for training or eval.
Run train.py --help
to see options. Dockerfile.train
is available for convenience, train.py
args can be added to the end of docker run
.
CUDA will be used if available.
There are two modes, one to train all case models serially on a single host, and with --parallel_key
to parallelize across several containers using AWS SQS.
Exploratory data analysis and data cleaning, saves new versions as data/{DATASET_VERSION}_clean.pkl
Data that was removed:
- Services and Documents marked as
deleted
, and associated Points - Services that lack any Documents, and associated Points
- Points marked as
[disputed, changes-requested, pending, draft]
(only ~60) - Documents without text (~1.5k) and associated Points
- A handful of Points that have a
quoteStart
but noquoteText
- Points with
quoteText
that no longer matchesdocument.text[point.quoteStart:point.quoteEnd]
, likely due to re-crawled text that changed. About 2k, down from 3k (1k were saved by re-searching for the quote and updatingquoteStart
/quoteEnd
) - Service 502, Case 235, Topic 53, Document 1378, and all associated Points
Data that was kept for now in case they're useful:
- Points that don't refer to docs
- Docs without any points (there are many)
- Non-english docs and points (can be filtered by doing
documents = documents[documents.lang == 'en']
)
The highlights of EDA from explore.ipynb
, like graphs and dataset size
An early notebook used to look at points for brainstorming, and help decide whether sentence classification vs sentence spans is the right paradigm. Finds that a lot of points span multiple sentences, so that's ideal, but over 5 sentences is rare.
Tests whether positive predictions using our inference strategy of sentence expansion (inference.apply_sent_span_model()
) yields spans about the
same length as human submitted points.
Plots ROC curves, precision recall curves, and helps find optimal thresholds.