nocotan / orakl

Active Learning for TensorFlow

Orakl

---

Orakl is an active learning library for TensorFlow.

System	3.7
Linux CPU

Installation

Installation Requirements

• python >= 3.7
• tensorflow>=2.0.0

Manual

If you'd like to try our latest features, you can install the latest master directly from GitHub. For a basic install, run:

git clone https://github.com/nocotan/orakl.git
cd orakl
pip install -e .

Tests

To execute unit tests from a manual install, run:

python3 -m unittest tests/attr/core/strategies/test_expected_model_change.py

or recursive test execution via the command line is built-in:

python3 -m unittest discover tests

Strategies

• Random Samplling (baseline)
• Expected Model Change Maximization [1]
• Margin Sampling [2]

LICENSE

Orakl is Apache-2.0 licensed, as found in the LICENSE file.

References

• [1] Cai, Wenbin, Ya Zhang, and Jun Zhou. "Maximizing expected model change for active learning in regression." 2013 IEEE 13th International Conference on Data Mining. IEEE, 2013.
• [2] Wang, Dan, and Yi Shang. "A new active labeling method for deep learning." 2014 International joint conference on neural networks (IJCNN). IEEE, 2014.