DDBP

Branches and experiments

1. comparison
   This branch contains the code for an experiment with a simple autoencoder.
2. comparison_fixed
   This branch is a further modification of comparision, during the fine tuning phase the weights in the already pretrained autoencoder are frozen and cannot be longer modified.

3. altered_comparison
   This experiment is based on the approach described in [1], the network was modified accordingly to use the advantages of the dedicated connections described in the paper.

4. altered_comparison_fixed
   This branch contains a similar modification as branch comparison_fixed but based on the experiment altered_comparison, the pretrained weights in the encoding layer are no longer modified during the last phase.

Running an experiment

In order to run an experiment please ensure you have TensorFlow installed. The version used during the project was TensorFlow 1.7. Then checkout the commit corresponding to the desired experiment:

$ git checkout hash

Architecture	Neuron count	Fixed	Experiment type	Hash
dedicated	104	No	No trump	511533f6bfec023990a9ce195092489eb2107a64
dedicated	104	No	Trump	a011ead0e055c032ec72c1872cd4885fc818c71e
dedicated	104	Yes	No trump	54c3de297e77c2afa8eb552fa2d623b09a0e5393
dedicated	104	Yes	Trump	47af6f0881c58e1c5f6b60f1ff3433638b7041b5
dedicated	156	No	No trump	d17cd31b47c0e9283bfd31b9d06a9d22781a4506
dedicated	156	No	Trump	2de6cd452c07cc801a3aa44b7dd4e9786f055788
dedicated	156	Yes	No trump	8f668eba9c178b1f4afacf01a4b531711587f209
dedicated	156	Yes	Trump	b9b545f2cf940ce63f35aa0cfad69278d4cc9808
full	104	No	No trump	8dc064ea6303057d53ceac12263376589c1d4f80
full	104	No	Trump	24f63e4dabe9359f2cd3956cbdc6a9321d54b8e3
full	104	Yes	No Trump	71d82d06a73a5d07d777284932fb3f43b9570795
full	104	Yes	Trump	22d3874cd5c48dbd360268d1d67957bb6f724486
full	156	No	No trump	83954aaa8ab0d4722465340c617103dacfb54da8
full	156	No	Trump	5cb485510f4e04bfc6bb3123a022bf5d2bfcb4bc
full	156	Yes	No trump	74f897b469e7e0f7a8808cd365ca6476a51edae3
full	156	Yes	Trump	3b30f93a4f939df64c488214cbecbdae195af29d

Then move to the directory DDBP/DDBP/run and execute the DDBP.py script

$ cd DDBP/DDBP/run
$ python3 DDBP.py

References

1. Amit, A., Markovitch, S.: Learning to bid in bridge. Machine Learning 63(3), 287– 327 (Jun 2006), https://doi.org/10.1007/s10994-006-6225-2
2. Beling, P.: Partition search revisited. IEEE Transactions on Computational Intelligence and AI in Games 9(1), 76–87 (March 2017)
3. David O.E., Netanyahu N.S., W.L.: DeepChess: End-to-End Deep Neural Network for Automatic Learning in Chess. Artificial Neural Networks and Machine Learning ICANN 2016 9887, 88–96 (2016)
4. Dharmalingam, M., Amalraj, R.: Articifial neural network architecture for solving the double dummy bridge problem in contract bridge. International Journal of Advanced Research in Computer and Communication Engineering Vol. 2, Issue 12, December 2013 2, 4683–4691 (2013)
5. Dharmalingam, M., Amalraj, R.: A solution to the double dummy contract bridge problem influenced by supervised learning module adapted by artificial neural network. ICTACT Journal on Soft Computing 5, 836–843 (2014)
6. Dharmalingam, M., Amalraj, R.: Supervised Elman Neural Network Architecture for Solving Double Dummy Bridge Problem in Contract Bridge. International Journal of Science and Research (IJSR) 3, 2745–2750 (2014)
7. Francis, H., Truscott, A., Francis, D. (eds.): The Official Encyclopedia of Bridge. American Contract Bridge League Inc, Memphis, TN, fifth edn. (1994)
8. Ginsberg, M.L.: http://www.gibware.com
9. Ginsberg, M.L.: Library of double-dummy results, http://www.cirl.uoregon. edu/ginsberg/gibresearch.html
10. Ho, C.Y., Lin, H.T.: Contract bridge bidding by learning. In: AAAI Workshop: Computer Poker and Imperfect Information (2015)
11. Man´dziuk, J., Mossakowski, K.: Example-based estimation of hand’s strength in the game of bridge with or without using explicit human knowledge. In: Proceedings of the IEEE Symposium on Computational Intelligence in Data Mining (CIDM 2007). pp. 413–420. IEEE Press, Honolulu, Hawaii, USA (2007)
12. Man´dziuk, J., Mossakowski, K.: Neural networks compete with expert human players in solving the double dummy bridge problem. In: 2009 IEEE Symposium on Computational Intelligence and Games. pp. 117–124 (Sept 2009)
13. Mossakowski, K., Man´dziuk, J.: Artificial neural networks for solving double dummy bridge problems. In: Artificial Intelligence and Soft Computing - ICAISC 2004. LNAI, vol. 3070, pp. 915–921. Springer (2004)
14. Mossakowski, K., Man´dziuk, J.: Neural networks and the estimation of hands strength in contract bridge. In: Rutkowski, L., et al. (eds.) Artificial Intelligence and Soft Computing ICAISC 2006. Lecture Notes in Artificial Intelligence, vol. 4029, pp. 1189–1198. Springer (2006)
15. Mossakowski, K., Man´dziuk, J.: Learning Without Human Expertise: A Case Study of the Double Dummy Bridge Problem. IEEE Transactions on Neural Networks 20(2), 278–299 (2009)
16. Muthusamy, D.: Double Dummy Bridge Problem in Contract Bridge: An Overview. Artificial Intelligence Systems and Machine Learning 10(1), 1–7 (2018)
17. Ng, A., Ngiam, J., Foo, C.Y., Mai, Y., Suen, C.: Ufldl tutorial, http://ufldl. stanford.edu/wiki/index.php/UFLDL_Tutorial
18. Yegnanarayana, B., Khemani, D., Sarkar, M.: Neural networks for contract bridge bidding. Sadhana 21(3), 395–413 (Jun 1996)