Solvi Arnold

Second Order Learning and the Evolution of Mental Representation

Mental representation is a fundamental aspect of advanced cognition. An understanding of the evolution of mental representation is essential to an understanding of the evolution of mind. However, being a decidedly mental phenomenon, its evolution is difficult to study. We hypothesize how interactions between adaptation levels may cause emergence of isomorphism between a cognitive system and its environment, and that mental representation may be understood as an instance of this effect. Specifically, we propose that selection for second order learning translates into selection for isomorphism-based implementation of first order learning ability, and that mental representation is (an aspect of) the environment-cognition isomorphism produced by such learning ability. We then give a reformulation of cognitive map ability, a paradigm case of mental representation, in terms of our hypothesis and explore it computationally by evolving a neural network species with the neural basics for second order plasticity (the basis for second order learning) in an environment composed of randomly generated maze tasks, including tasks generally believed to require mental representation (in the form of cognitive maps). The model is shown capable of evolving nets that solve these tasks, providing preliminary support for our hypothesis.

Unfolding of a rectangular cloth from unarranged starting shapes by a Dual-Armed robot with a mechanism for managing recognition error and uncertainty

Graphical Abstract We propose a method for unfolding a rectangular cloth placed on a table in an arbitrary unarranged shape, using a dual arm robot. There are many situations where the manipulation of fabric products by dual arm robots is slow due to operation complexity. Also, observation of fabric products in unarranged shapes can be fraught with uncertainty, posing further difficulties for robotic manipulation. In this article, we address these problems for our specific task, implementing a ‘pinch and slide motion’ to address the former issue, and an operation selection mechanism implemented as a partially observable Markov decision process to address the latter. We used this approach to let a robot unfold a rectangular cloth, thereby experimentally verifying the effectiveness of our approach.

Search Activities in Outdoor Environments by Cyber Rescue Dogs with Recognition Assistance

This paper reports results on our verification experiments for outdoor disaster response using rescue dogs. We have developed environment recognition system that especially focuses on constructing recognition functions quickly. We embedded this result into the system of cyber rescue dog. We designed a scenario that assumed disaster-stricken situation, and confirmed the effectiveness of the system.

The emergence of executive functions by the evolution of second-order learning

Mental representation (MR) is regarded as part of sophisticated cognitive processes. It has been argued that under selection pressure for second-order learning (learning how to learn), first-order learning evolves to facilitate second-order learning within lifetime by capturing inherent structures of changing environment as MR. Two hypotheses derive from this theory: (1) solving MR-dependent tasks should involve second-order plasticity at the neural level. (2) Solving MR-dependent tasks should involve internalization of structural features of environment into corresponding features of the cognitive system. In this paper, constructive approach was taken and the result was analyzed from the viewpoint of these two hypotheses. Executive functions, a collection of cognitive processes necessary for good performance on complex tasks, are the theme of our model. They are considered to be related to theory of mind, which is a typical example of MR. We conducted an evolutionary simulation where agents with recurrent neural networks tackled the Wisconsin card sorting test (WCST), a widely used task to measure abilities of executive functions. The results showed some agents were successfully able to achieve ideal scores in the WCST, hence the emergence of executive functions. In addition, we also discussed the hypotheses based on one of the evolved neural networks.

Implicit policies for deformable object manipulation with arbitrary start and end states: A novel evolutionary approach

We present a novel method for training (evolving) fully convolutional neural networks (CNNs) for deformable object manipulation. Instead of using a weight update rule, we evolve an ensemble of compositional pattern generating networks (CPPNs) by means of a genetic algorithm (GA). These ensembles generate the convolutional kernels that comprise the CNN. This allows the GA to search for fit kernels in the space of spatial 2D patterns, which allows for evolution of larger networks than is feasible with direct evolution of the connection weight vector. We apply this method to a thread manipulation task, in an attempt to let the CNN grasp the deformation dynamics of the thread. We report results both on single-manipulation tasks, and on tasks requiring a sequence of two manipulations, establishing the viability of this approach.