Abstract
Universal induction relies on some general search procedure that is doomed to be inefficient. One possibility to achieve both generality and efficiency is to specialize this procedure w.r.t. any given narrow task. However, complete specialization that implies direct mapping from the task parameters to solutions (discriminative models) without search is not always possible. In this paper, partial specialization of general search is considered in the form of genetic algorithms (GAs) with a specialized crossover operator. We perform a feasibility study of this idea implementing such an operator in the form of a deep feedforward neural network. GAs with trainable crossover operators are compared with the result of complete specialization, which is also represented as a deep neural network. Experimental results show that specialized GAs can be more efficient than both general GAs and discriminative models.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Solomonoff, R.: A formal theory of inductive inference, part 1 and part 2. Inf. Control 7, 1–22, 224–254 (1964)
Graves, A., Wayne, G., Danihelka, I.: Neural turing machines. arXiv:1410.5401 [cs.NE] (2014)
Graves, A., et al.: Hybrid computing using a neural network with dynamic external memory. Nature 538, 471–476 (2016)
Riedel, S., Bošnjak, M., Rocktäschel, T.: Programming with a differentiable forth interpreter. arXiv:1605.06640 (2016)
Reed, S., Freitas, N.: Neural Programmer-interpreters. arXiv:1511.06279 [cs.LG] (2015)
Kaiser, Ł., Sutskever, I.: Neural GPUs learn algorithms. arXiv:1511.08228 [cs.LG] (2015)
Andrychowicz, M., et al.: Learning to learn by gradient descent by gradient descent. arXiv:1606.04474 [cs.NE] (2016)
Wang, J.X., et al.: Learning to reinforcement learn. arXiv:1611.05763 [cs.LG] (2016)
Duan, Y., et al.: RL2: fast reinforcement learning via slow reinforcement learning. arXiv:1611.02779 [cs.AI] (2016)
Hochreiter, S., Younger, A.S., Conwell, P.R.: Learning to learn using gradient descent. In: International Conference on Artificial Neural Networks, pp. 87–94 (2001)
Stuhlmüller, A., Goodman, N.D.: A dynamic programming algorithm for inference in recursive probabilistic programs. arXiv:1206.3555 [cs.AI] (2012)
Potapov, A., Rodionov, S.: Making universal induction efficient by specialization. In: Goertzel, B., Orseau, L., Snaider, J. (eds.) AGI 2014. LNCS, vol. 8598, pp. 133–142. Springer, Cham (2014). doi:10.1007/978-3-319-09274-4_13
Potapov, A., Batishcheva, V., Rodionov, S.: Optimization framework with minimum description length principle for probabilistic programming. In: Bieger, J., Goertzel, B., Potapov, A. (eds.) AGI 2015. LNCS, vol. 9205, pp. 331–340. Springer, Cham (2015). doi:10.1007/978-3-319-21365-1_34
Batishcheva, V., Potapov, A.: Genetic programming on program traces as an inference engine for probabilistic languages. In: Bieger, J., Goertzel, B., Potapov, A. (eds.) AGI 2015. LNCS, vol. 9205, pp. 14–24. Springer, Cham (2015). doi:10.1007/978-3-319-21365-1_2
Solomonoff, R.J.: The discovery of algorithmic probability. J. Comput. Syst. Sci. 55(1), 73–88 (1997)
Özkural, E.: An application of stochastic context sensitive grammar induction to transfer learning. In: Goertzel, B., Orseau, L., Snaider, J. (eds.) AGI 2014. LNCS, vol. 8598, pp. 121–132. Springer, Cham (2014). doi:10.1007/978-3-319-09274-4_12
Le, T.A., Baydin, A.G., Wood, F.: Inference compilation and universal probabilistic programming. arXiv:1610.09900 [cs.AI] (2016)
Goertzel, B.: From Complexity to Creativity: Explorations in Evolutionary, Autopoietic, and Cognitive Dynamics. Springer, New York (1997)
Zhang, C., Bengio, S., Hardt, M., Recht, B., Vinyals, O.: Understanding deep learning requires rethinking generalization. arXiv:1611.03530 [cs.LG] (2017)
Futamura, Y.: Partial evaluation of computation process – an approach to a compiler-compiler. Syst. Comput. Controls 2(5), 45–50 (1971)
Acknowledgements
This work was supported by Government of Russian Federation, Grant 074-U01.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
Potapov, A., Rodionov, S. (2017). Genetic Algorithms with DNN-Based Trainable Crossover as an Example of Partial Specialization of General Search. In: Everitt, T., Goertzel, B., Potapov, A. (eds) Artificial General Intelligence. AGI 2017. Lecture Notes in Computer Science(), vol 10414. Springer, Cham. https://doi.org/10.1007/978-3-319-63703-7_10
Download citation
DOI: https://doi.org/10.1007/978-3-319-63703-7_10
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-63702-0
Online ISBN: 978-3-319-63703-7
eBook Packages: Computer ScienceComputer Science (R0)