It is a novel approach for one-shot rule learning using a logic program declarative bias which is a special case of Top-Directed Hypothesis Derivation (TDHD)[1]

We apply OSHD to the challenging task of Malayalam character recognition. This is a challenging task due to spherical and complex structure of Malayalam hand-written language. We compare our results with a state-of-the- art Deep Learning approach, called Siamese Network, which has been developed for one-shot learning.

Malayalam is an official language for Kerala, a state of India. Unlike for other languages, there is currently no efficient algorithm for Malayalam handwritten recognition. The spherical feature of Malayalam characters are the main reason for this problem.

We selected the hand-written characters provided by Omniglot [2,3] dataset. Feature extraction is done by using a set of advanced geometrical features and directional features.

ILP - This folder contains the program codes and instructions to execute OSHD using ILP.

Deep_Learning_Siamese_Net - This folder contains the program codes and instructions to execute Deep Learning Siamese Network.

datasets - The Dataset folder contains;

1. Character_Malayalm_Dataset.csv : Holds the feature values which is extracted from the 'Omniglot' dataset fro Malayalam Characters.

2. feature.py : This script will generate the prolog file(OSHD.pl) for the OSHD experiment which contains Background Knowledge, Positive & Negative examples, Setup details for OSHD

3. OSHD.pl : prolog file. By Default, all the examples added as negative examples with weight -3.

4. Malayalam_Alphabets(Zip File) : This files contains the Malayalam alphabets from 'Omniglot' dataset. Sample Malayalam character is displayed below.

Malayalam Alphabet 'Aha'

• [1] Muggleton, S.H., Santos, J., Tamaddoni-Nezhad, A.: TopLog: ILP using a logic program declarative bias. In: Proceedings of the International Conference on Logic Programming 2008. pp. 687–692. LNCS 5366, Springer-Verlag (2010)
• [2] Lake, B., Salakhutdinov, R., Gross, J., Tenenbaum, J.: One shot learning of simple visual concepts. In: Proceedings of the 33rd Annual Conference of the Cognitive Science Society. pp. 2568–2573 (2011)
• [3] Lake, B.M., Salakhutdinov, R., Tenenbaum, J.B.: Human-level concept learning through probabilistic program induction. Science 350(6266), 1332–1338 (2015)