In recent years, various animal observation instruments have been developed to support long-term measurement and analysis of animal behaviors. This study proposes an automatic observation instrument that specializes for turning behaviors of pill bugs and aims to obtain new knowledge in the field of ethology. Pill bugs strongly tend to turn in the opposite direction of a preceding turn. This alternation of turning is called turn alternation reaction. However, a repetition of turns in the same direction is called turn repetition reaction and has been considered a malfunction of turn alternation. In this research, the authors developed an automatic turntable-type multiple T-maze device and observed the turning behavior of 34 pill bugs for 6 h to investigate whether turn repetition is a malfunction. As a result, most of the pill bug movements were categorized into three groups: sub-diffusion, Brownian motion, and Lévy walk. This result suggests that pill bugs do not continue turn alternation mechanically but elicit turn repetition moderately, which results in various movement patterns. In organisms with relatively simple nervous systems such as pill bugs, stereotypical behaviors such as turn alternation have been considered mechanical reactions and variant behaviors such as turn repetition have been considered malfunctions. However, our results suggest that a moderate generation of turn repetition is involved in the generation of various movement patterns. This study is expected to provide a new perspective on the conventional view of the behaviors of simple organisms.

In this study, we start by proposing a causal induction model that incorporates symmetry bias. This model has two parameters that control the strength of symmetry bias and includes conditional probability and conventional models of causal induction as special cases. It can reproduce causal induction of human judgment with high accuracy. We further propose a human-like Bayesian inference method to replace the conditional probability in Bayesian inference with the aforementioned causal induction model. In this method, two components coexist: the component of Bayesian inference, which updates the degree of confidence for each hypothesis, and the component of inverse Bayesian inference that modifies the model of each hypothesis.In other words, this method allows not only inference but also simultaneous learning. Our study demonstrates that the method with both Bayesian inference and inverse Bayesian inference enables us to deal flexibly with unsteady situations where the target of inference changes occasionally.

Human beings have adaptively rational cognitive biases for efficiently acquiring concepts from small-sized datasets. With such inductive biases, humans can generalize concepts by learning a small number of samples. By incorporating human cognitive biases into learning vector quantization (LVQ), a prototype-based online machine learning method, we developed self-incremental LVQ (SILVQ) methods that can be easily interpreted. We first describe a method to automatically adjust the learning rate that incorporates human cognitive biases. Second, SILVQ, which self-increases the prototypes based on the method for automatically adjusting the learning rate, is described. The performance levels of the proposed methods are evaluated in experiments employing four real and two artificial datasets. Compared with the original learning vector quantization algorithms, our methods not only effectively remove the need for parameter tuning, but also achieve higher accuracy from learning small numbers of instances. In the cases of larger numbers of instances, SILVQ can still achieve an accuracy that is equal to or better than those of existing representative LVQ algorithms. Furthermore, SILVQ can learn linearly inseparable conceptual structures with the required and sufficient number of prototypes without overfitting.