Computable Variants of AIXI which are More Powerful than AIXItl

Katayama, Susumu

doi:10.2478/jagi-2019-0001

Abstract

This paper presents Unlimited Computable AI, or UCAI, that is a family of computable variants of AIXI. UCAI is more powerful than AIXItl, which is a conventional family of computable variants of AIXI, in the following ways: 1) UCAI supports models of terminating computation, including typed lambda calculi, while AIXItl only supports Turing machine with timeout ˜t, which can be simulated by typed lambda calculi for any ˜t; 2) unlike UCAI, AIXItl limits the program length to some ˜l .

References

Bird, R., and Wadler, P. 1988. An Introduction to Functional Programming. Prentice-Hall.
Search in Google Scholar Back to article
Boehm, H., and Cartwright, R. 1990. Exact Real Arithmetic Formulating Real Numbers As Functions. In Turner, D. A., ed., Research Topics in Functional Programming. Boston, MA, USA: Addison-Wesley Longman Publishing Co., Inc. 43–64.
Search in Google Scholar Back to article
Claessen, K., and Pałka, M. H. 2013. Splittable Pseudorandom Number Generators Using Cryptographic Hashing. In Proceedings of the 2013 ACM SIGPLAN Symposium on Haskell, Haskell ’13, 47–58. New York, NY, USA: ACM.10.1145/2503778.2503784
Search in Google Scholar Back to article
Hutter, M. 2007. Universal Algorithmic Intelligence: A Mathematical Top→Down Approach. In Goertzel, B., and Pennachin, C., eds., Artificial General Intelligence, Cognitive Technologies. Berlin: Springer. 227–290.10.1007/978-3-540-68677-4_8
Search in Google Scholar Back to article
Katayama, S. 2016. Ideas for a Reinforcement Learning Algorithm that Learns Programs. In Artificial General Intelligence - 9th International Conference, AGI 2016, AGI 2016, New York, USA, July 16–19, 2016, Proceedings, 354–362.10.1007/978-3-319-41649-6_36
Search in Google Scholar Back to article
Plume, D. 1998. A Calculator for Exact Real Number Computation. Ph.D. Dissertation, University of Edinburgh.
Search in Google Scholar Back to article
Sutton, R. S., and Barto, A. G. 1998. Introduction to Reinforcement Learning. Cambridge, MA, USA: MIT Press, 1st edition.
Search in Google Scholar Back to article
Veness, J.; Ng, K. S.; Hutter, M.; Uther, W.; and Silver, D. 2011. A Monte-Carlo AIXI Approximation. Journal of Artificial Intelligence Research 40:95–142.10.1613/jair.3125
Search in Google Scholar Back to article
Watkins, C. J. C. H., and Dayan, P. 1992. Q-learning. In Machine Learning, 279–292.10.1023/A:1022676722315
Search in Google Scholar Back to article
Willems, F. M. J.; Shtarkov, Y. M.; and Tjalkens, T. J. 1995. The Context Tree Weighting Method: Basic Properties. IEEE Transactions on Information Theory 41:653–664.10.1109/18.382012
Search in Google Scholar Back to article

Computable Variants of AIXI which are More Powerful than AIXItl

Abstract

Paradigm

My account