A One Shot Learning using Bayesian Program Learning

[Art by Danqing Wang]

Here is a brief explanation of what BPL algorithm is.

---

Overview

This is a One-Shot Learning Handwritten Character Classifer written in Python.

---

Dependencies

The dependencies can be all install via pip

1. Numpy
2. Scipy
3. Pillow

---

Result

---

Disclaimer

The original BPL repo is here: written in MATLAB.

The original BPL paper can be found by searching:

Lake, B. M., Salakhutdinov, R., and Tenenbaum, J. B. (2015). Human-level concept learning through probabilistic program induction. Science, 350(6266), 1332-1338.