Co-learning system for humans and machines using a weighted majority-based method

Abstract

Artificial intelligence systems are frequently used to solve various problems in our daily lives. However, these systems require problem-specific big data to facilitate their learning processes. Unfortunately, for unknown environments, there are no previous instances available for learning. To support such learning in unknown environments, we propose a novel hybrid learning system that facilitates collaborative learning between humans and artificial intelligence systems. In this study, we verified that the proposed system accelerated both human and machine learning by employing a simplified color design task. Moreover, we also improved the system to enable it to select the best answer from the solution candidates by using masters to evaluate these solution candidates. The system performance was evaluated using both a simulation and a psychological test comprising a color design task.

Keywords

Collaborative learning accelerated learning human skill general regression neural network

References

Price

and Boutilier

, Accelerating reinforcement learning through implicit imitation, Journal of Artificial Intelligence Research 19 (2003), 569-629.

, Barto

, Powell

and Wunsch

, Supervised Actor-Critic Reinforcement Learning, A Wiley-Interscience Publication, May 2012.

Ito

, Noda

, Hoshino

and Tani

, Dynamic and interactive generation of object handling behaviors by a small humanoid robot using a dynamic neural network model, Neural Networks 19 (2006), 323-337.

Fernàndez Anta

, Georgiou

, Mosteiro

M.A.

and Pareja

, Algorithmic mechanisms for reliable crowdsourcing computation under collusion, PLoS ONE 10(3) (2015), 1-22.

Ogiso

, Yamauchi

, Ishii

and Suzuki

, An ensemble learning system for correcting user skills and its application for education -proposal of skill collector-, In Proceedings of the 62th annual conference of JSSD, 2015, pp. 150-151.

Ogiso

, Yamauchi

, Ishii

and Suzuki

, A co-learning system for human and machines. ≠wblock In Proceedings of the 2015 Seventh International Conference of Soft Computing and Pattern Recognition (SoCPaR2015), The Institute of Electrical and Electronics Engineers, Inc. New York, New York, November 2015, pp. 363-368.

Anta

A.F.

and López

A.S.L.

, Reliable internet-based master-worker computing in the presence of malicious workers, Parallel Processing Letters 22(1) (2010), 1250002.

Konwar

K.M.

, Rajasekaran

and Shvartsman

A.A.

, Robust network supercomputing with unreliable workers, Journal of Parallel and Distributed Computing 75 (2015), 81-92.

Golle

and Mironov

, Uncheatable distributed computations, In Topics in Cryptology- CT-RSA 2001 Volume 2020 of the series Lecture Notes in Computer Science, Springer-Verlag, 2001, pp. 425-440.

10.

Forges

, An approach to communication equilibria, ECONOMETRICA 54(6) (1986), 1375-1385.

11.

Conforth

and Meng

, Reinforcement learning using swarm intelligence-trained neural networks, Journal of Experimental & Theoretical Artificial Intelligence 22(3) (2010), 197-218.

12.

Whitehill

, Fan Wu

, Bergsma

, Movellan

J.R.

and Ruvolo

P.L.

, Whose vote should count more: Optimal integration of labels from labelers of unknown expertise, in: Advances in Neural Information Processing Systems 22 (NIPS 2009), Y. Bengio, D. Schuurmans, J.D. Lafferty, C.K.I. Williams and A. Culotta, eds, Curran Associates, Inc., 2009, pp. 2035-2043.

13.

A modified general regression neural network (mgrnn) with a new efficient training algorithm as a robust.

14.

Igushi

, Ogiso

and Yamauchi

, Acceleration of reinforcement learning via game-based renewal energy management system, In SCISISIS2014, The Institute of Electrical and Electronics Engineers, Inc. New York, New York, December 2014, pp. 415-420.

15.

Breiman

, Bagging predictors, Machine Learning 26 (1996), 123-140.