Deep Reinforcement Learning (RL) is often criticised for being data inefficient and inflexible to changes in task structure. Part of the reason for these issues is that Deep RL typically learns end-to-end using backpropagation, which results in task-specific representations. One approach for circumventing these problems is to apply Deep RL to existing representations that have been learned in a more task-agnostic fashion. However, this only partially solves the problem as the Deep RL algorithm learns a function of all pre-existing representations and is therefore still susceptible to data inefficiency and a lack of flexibility. Biological agents appear to solve this problem by forming internal representations over many tasks and only selecting a subset of these features for decision-making based on the task at hand; a process commonly referred to as selective attention. We take inspiration from selective attention in biological agents and propose a novel algorithm called Selective Particle Attention (SPA), which selects subsets of existing representations for Deep RL. Crucially, these subsets are not learned through backpropagation, which is slow and prone to overfitting, but instead via a particle filter that rapidly and flexibly identifies key subsets of features using only reward feedback. We evaluate SPA on two tasks that involve raw pixel input and dynamic changes to the task structure, and show that it greatly increases the efficiency and flexibility of downstream Deep RL algorithms.

The human–computer interaction (HCI) design of educational technologies influences cognitive behaviour, so it is imperative to assess how different HCI strategies support intended behaviour. We developed a neuroscience-inspired game that trains children's use of “stopping-and-thinking” (S&T)—an inhibitory control-related behaviour—in the context of counterintuitive science problems. We tested the efficacy of four HCI features in supporting S&T: (1) a readiness mechanic, (2) motion cues, (3) colour cues, and (4) rewards/penalties. In a randomised eye-tracking trial with 45 7-to-8-year-olds, we found that the readiness mechanic increased S&T duration, that motion and colour cues proved equally effective at promoting S&T, that combining symbolic colour with the readiness mechanic may have a cumulative effect, and that rewards/penalties may have distracted children from S&T. Additionally, S&T duration was related to in-game performance. Our results underscore the importance of interdisciplinary approaches to educational technology research that actively investigates how HCI impacts intended learning behaviours.

Complementary Learning Systems (CLS) theory suggests that the brain uses a 'neocortical' and a 'hippocampal' learning system to achieve complex behaviour. These two systems are complementary in that the 'neocortical' system relies on slow learning of distributed representations while the 'hippocampal' system relies on fast learning of pattern-separated representations. Both of these systems project to the striatum, which is a key neural structure in the brain's implementation of Reinforcement Learning (RL). Current deep RL approaches share similarities with a 'neocortical' system because they slowly learn distributed representations through backpropagation in Deep Neural Networks (DNNs). An ongoing criticism of such approaches is that they are data inefficient and lack flexibility. CLS theory suggests that the addition of a 'hippocampal' system could address these criticisms. In the present study we propose a novel algorithm known as Complementary Temporal Difference Learning (CTDL), which combines a DNN with a Self-Organizing Map (SOM) to obtain the benefits of both a 'neocortical' and a 'hippocampal' system. Key features of CTDL include the use of Temporal Difference (TD) error to update a SOM and the combination of a SOM and DNN to calculate action values. We evaluate CTDL on Grid World, Cart-Pole and Continuous Mountain Car tasks and show several benefits over the classic Deep Q-Network (DQN) approach. These results demonstrate (1) the utility of complementary learning systems for the evaluation of actions, (2) that the TD error signal is a useful form of communication between the two systems and (3) that our approach extends to both discrete and continuous state and action spaces.

This manuscript surveys computational modeling efforts by researchers in developmental psychology. Developmental psychology is ready to blossom into a modern science that focuses on causal mechanistic explanations of development rather than just describing and classifying behaviors. Computational modeling is the key to this process. However, to be effective, models must not only mimic observed data. They must also be transparent, grounded, and plausible to be accepted by the developmental psychology community. Connectionist model provides one such example. Many developmental features of typical and atypical perception, cognition, and language have been modeled using connectionist methods. Successful models are closely tied to the details of existing empirical studies and make concrete testable predictions. The success of such a project relies on the close collaboration of computational scientists with empirical psychologists

This paper presents a modular connectionist network model of the development of seriation (sorting) in children. The model uses the cascade-correlation generative connectionist algorithm. These cascade-correlation networks do better than existing rule-based models at developing through soft stage transitions, sorting more correctly with larger stimulus size increments and showing variation in seriation performance within stages. However, the full generative power of cascade-correlation was not found to be a necessary component for successfully modelling the development of seriation abilities. Analysis of network weights indicates that improvements in seriation are due to continuous small changes instead of the radical restructuring suggested by Piaget. The model suggests that seriation skills are present early in development and increase in precision during later development. The required learning environment has a bias towards smaller and nearly ordered arrays. The variability characteristic of children's performance arises from sorting subsets of the total array. The model predicts better sorting moves with more array disorder, and a dissociation between which element should be moved and where it should be moved.