Virtual Skinner Box for the Test of Operant Conditioning
Deokgun Park, Md Ashaduzzaman Rubel Mondol · 2022 13th International Conference on Information and Communication Technology Convergence (ICTC) · 2022
In this paper, authors evaluate the progress of the artificial intelligence research according to the levels of intelligence. Current model-free reinforcement learning is similar to the Level 1 intelligence or instincts that is refined with evolution. Animals with Level 2 intelligence can learn with operant conditioning where behavior policy can be updated with a few trials. Therefore, authors claim that next mile-stone should be the operant conditioning and propose an experiment framework motivated by the skinner box experiment.