Alexander Chernyavskiy , Alexey Skrynnik , Aleksandr Panov
{"title":"Applying opponent and environment modelling in decentralised multi-agent reinforcement learning","authors":"Alexander Chernyavskiy , Alexey Skrynnik , Aleksandr Panov","doi":"10.1016/j.cogsys.2024.101306","DOIUrl":"10.1016/j.cogsys.2024.101306","url":null,"abstract":"<div><div>Multi-agent reinforcement learning (MARL) has recently gained popularity and achieved much success in different kind of games such as zero-sum, cooperative or general-sum games. Nevertheless, the vast majority of modern algorithms assume information sharing during training and, hence, could not be utilised in decentralised applications as well as leverage high-dimensional scenarios and be applied to applications with general or sophisticated reward structure. Thus, due to collecting expenses and sparsity of data in real-world applications it becomes necessary to use world models to model the environment dynamics, using latent variables — i.e. use world model to generate synthetic data for training of MARL algorithms. Therefore, focusing on the paradigm of decentralised training and decentralised execution, we propose an extension to the model-based reinforcement learning approaches leveraging fully decentralised training with planning conditioned on neighbouring co-players’ latent representations. Our approach is inspired by the idea of opponent modelling. The method makes the agent learn in joint latent space without need to interact with the environment. We suggest the approach as proof of concept that decentralised model-based algorithms are able to emerge collective behaviour with limited communication during planning, and demonstrate its necessity on iterated matrix games and modified versions of StarCraft Multi-Agent Challenge (SMAC).</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"89 ","pages":"Article 101306"},"PeriodicalIF":2.1,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143144639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Ismail M. Gadzhiev , Alexander S. Makarov , Vadim L. Ushakov , Vyacheslav A. Orlov , Georgy A. Ivanitsky , Sergei A. Dolenko
{"title":"Creating A dynamic cognovisor – Brain activity recognition using principal Component analysis and Machine learning models","authors":"Ismail M. Gadzhiev , Alexander S. Makarov , Vadim L. Ushakov , Vyacheslav A. Orlov , Georgy A. Ivanitsky , Sergei A. Dolenko","doi":"10.1016/j.cogsys.2024.101314","DOIUrl":"10.1016/j.cogsys.2024.101314","url":null,"abstract":"<div><div>This study explores the feasibility of developing a dynamic cognovisor capable of recognizing cognitive states and transitions using fMRI data. Data were collected from 31 participants performing spatial and verbal tasks during fMRI scanning and were preprocessed using a nine-step algorithm for artifact removal and denoising. Three types of classification problems were examined, with machine learning methods and dimensionality reduction techniques applied to classify activity states. The best-performing models were identified for each classification problem, providing insights into their applicability. Notably, binary classification of resting versus active states achieved good quality with relatively simple methods. A key finding underscores the importance of accounting for temporal history of the signal prior to the prediction moment to improve model performance.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"89 ","pages":"Article 101314"},"PeriodicalIF":2.1,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143143564","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rubén Torres Agustín , Zareth Bonilla González , Mario A. Rodríguez Camacho , Sebastián Almonte , Wendy Fabiola Lara Galindo , Francisco Abelardo Robles Aguirre
{"title":"Corrigendum to “Detection of semantic inconsistencies of motor actions: From language to praxis” [Cognit. Syst. Res. 88 (2024) 1–13/101292]","authors":"Rubén Torres Agustín , Zareth Bonilla González , Mario A. Rodríguez Camacho , Sebastián Almonte , Wendy Fabiola Lara Galindo , Francisco Abelardo Robles Aguirre","doi":"10.1016/j.cogsys.2025.101323","DOIUrl":"10.1016/j.cogsys.2025.101323","url":null,"abstract":"","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"89 ","pages":"Article 101323"},"PeriodicalIF":2.1,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143143568","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mikhail Kiselev , Alexander Ivanitsky , Denis Larionov
{"title":"A purely spiking approach to reinforcement learning","authors":"Mikhail Kiselev , Alexander Ivanitsky , Denis Larionov","doi":"10.1016/j.cogsys.2024.101317","DOIUrl":"10.1016/j.cogsys.2024.101317","url":null,"abstract":"<div><div>At present, implementation of learning mechanisms in spiking neural networks (SNN) cannot be considered as a solved scientific problem despite plenty of SNN learning algorithms proposed. It is also true for SNN implementation of reinforcement learning (RL), while RL is especially important for SNNs because of its close relationship to the domains most promising from the viewpoint of SNN application such as robotics. In the present paper, an SNN structure is described which, seemingly, can be used in wide range of RL tasks. The distinctive feature of our approach is usage of only the spike forms of all signals involved — sensory input streams, output signals sent to actuators and reward/punishment signals. Besides that, selection of the neuron/plasticity models was determined by the requirement that they should be easily implemented on modern neurochips. The SNN structure considered in the paper includes spiking neurons described by a generalization of the LIFAT (leaky integrate-and-fire neuron with adaptive threshold) model and a simple spike timing dependent synaptic plasticity model (a generalization of dopamine-modulated plasticity). In this study, we use the model-free approach to RL but it is based on very general assumptions about RL task characteristics and has no visible limitations on its applicability (inside the class of model-free RL tasks). To test our SNN, we apply it to a simple but non-trivial task of training the network to keep a chaotically moving light spot in the view field of an emulated Dynamic Vision Sensor (DVS) camera. Successful solution of this RL problem can be considered as an evidence in favor of efficiency of our approach.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"89 ","pages":"Article 101317"},"PeriodicalIF":2.1,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143143567","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Higher-order adaptive dynamical system modelling of epigenetic mechanisms in infant temperament shaped by prenatal maternal stress","authors":"Labiba Aziz , Jan Treur","doi":"10.1016/j.cogsys.2024.101315","DOIUrl":"10.1016/j.cogsys.2024.101315","url":null,"abstract":"<div><div>Prenatal maternal stress (PNMS) has significant implications for infant temperament, primarily through alterations in the hypothalamic–pituitary–adrenal (HPA) axis and epigenetic mechanisms. This study explores the effects of PNMS on infant stress reactivity using a fifth-order adaptive dynamical system model. The model integrates genetic, epigenetic, and environmental factors, focusing on the downregulation of 11β-HSD-2, an enzyme responsible for converting active cortisol to its inactive form, and its subsequent influence on fetal cortisol exposure. The article also employs network-oriented modeling to represent epigenetic changes and their impact on infant temperament development, emphasizing the HPA axis’ role in stress regulation. Simulation experiments compare scenarios with PNMS, illustrating the long-term developmental consequences on temperament. This research highlights the importance of maternal well-being during pregnancy in shaping infant stress responses and provides insights into the developmental origins of health and disease.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"90 ","pages":"Article 101315"},"PeriodicalIF":2.1,"publicationDate":"2024-12-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143097275","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Integrated model of cerebellal supervised learning and basal ganglia’s reinforcement learning for mobile robot behavioral decision-making","authors":"Zhiqiang Wu , Dongshu Wang , Lei Liu","doi":"10.1016/j.cogsys.2024.101302","DOIUrl":"10.1016/j.cogsys.2024.101302","url":null,"abstract":"<div><div>Behavioral decision-making in unknown environments of mobile robots is a crucial research topic in robotics. Inspired by the working mechanism of different brain regions in mammals, this paper designed a new hybrid model integrating the functions of cerebellum and basal ganglia by simulating the memory replay of hippocampus, so as to realize the autonomous behavioral decision-making of robot in unknown environments. A reinforcement learning module based on Actor-Critic framework and a developmental network module are used to simulate the functions of the basal ganglia and cerebellum, respectively. Considering the different functions of D1 and D2 dopamine receptors in basal ganglia, an Actor network module with separate learning of positive and negative rewards is designed for the basal ganglia to realize efficient exploration of the environments by the agent. According to the characteristics of biological memory, a physiological memory priority index is designed for hippocampus memory replay, which improves the offline learning efficiency of cerebellum. The integrated model enables dynamic switching between decisions made by cerebellum and basal ganglia based on the agent’s cognitive level with respect to the environment. Finally, the effectiveness of the proposed model is verified through experiments on agent navigation in both simulation and real environments, as well as through performance comparison experiments with other learning algorithms.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"88 ","pages":"Article 101302"},"PeriodicalIF":2.1,"publicationDate":"2024-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593526","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Cognitive biases in natural language: Automatically detecting, differentiating, and measuring bias in text","authors":"Kyrtin Atreides, David J. Kelley","doi":"10.1016/j.cogsys.2024.101304","DOIUrl":"10.1016/j.cogsys.2024.101304","url":null,"abstract":"<div><div>We examine preliminary results from the first automated system to detect the 188 cognitive biases included in the 2016 Cognitive Bias Codex, as applied to both human and AI-generated text, and compared to a human baseline of performance. The human baseline was constructed from the collective intelligence of a small but diverse group of volunteers independently submitting their detected cognitive biases for each sample in the task used for the first phase. This baseline was used as an approximation of the ground truth on this task, for lack of any prior established and relevant benchmark. Results showed the system’s performance to be above that of the average human, but below that of the top-performing human and the collective, with greater performance on a subset of 18 out of the 24 categories in the codex. This version of the system was also applied to analyzing responses to 150 open-ended questions put to each of the top 5 performing closed and open-source Large Language Models, as of the time of testing. Results from this second phase showed measurably higher rates of cognitive bias detection across roughly half of all categories than those observed when analyzing human-generated text. The level of model contamination was also considered for two types of contamination observed, where the models gave canned responses. Levels of cognitive bias detected in each model were compared both to one another and to data from the first phase.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"88 ","pages":"Article 101304"},"PeriodicalIF":2.1,"publicationDate":"2024-10-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142593527","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"A new perspective on Misbeliefs: A computational model for perceived control","authors":"Haokui Xu , Bohao Shi , Yiming Zhu , Jifan Zhou , Mowei Shen","doi":"10.1016/j.cogsys.2024.101305","DOIUrl":"10.1016/j.cogsys.2024.101305","url":null,"abstract":"<div><div>The discovery of various cognitive biases and social illusions indicates that people routinely have misbeliefs. Focusing on the illusion of control (IOC), this article argues that when time and cognitive resources are limited, and information is imperfect, misbeliefs can be generated naturally in a normal belief formation system, and these misbeliefs might help people adapt better to the environment.<!--> <!-->In this study, we present a computational model—the informativeness-weighting model (IWM)—describing how beliefs are revised by observed evidence. To be precise, IOC is the result of distinct types of evidence being endowed with different weights according to its informativeness in a belief revision process. To evaluate the model, we also designed two behavioral experiments to compare people’s sense of control with that predicted by the model.<!--> <!-->In both experiments, our model outperformed two alternative models in predicting and explaining the misestimation of people’s perceived control. Thus, we suggest that our model reflects an adaptive strategy for information processing, which helps to explain why misbeliefs, like IOC, are prevalent in human cognition.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"88 ","pages":"Article 101305"},"PeriodicalIF":2.1,"publicationDate":"2024-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142526108","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Inga Ibs , Claire Ott , Frank Jäkel, Constantin A. Rothkopf
{"title":"From human explanations to explainable AI: Insights from constrained optimization","authors":"Inga Ibs , Claire Ott , Frank Jäkel, Constantin A. Rothkopf","doi":"10.1016/j.cogsys.2024.101297","DOIUrl":"10.1016/j.cogsys.2024.101297","url":null,"abstract":"<div><div>Many complex decision-making scenarios encountered in the real-world, including energy systems and infrastructure planning, can be formulated as constrained optimization problems. Solutions for these problems are often obtained using white-box solvers based on linear program representations. Even though these algorithms are well understood and the optimality of the solution is guaranteed, explanations for the solutions are still necessary to build trust and ensure the implementation of policies. Solution algorithms represent the problem in a high-dimensional abstract space, which does not translate well to intuitive explanations for lay people. Here, we report three studies in which we pose constrained optimization problems in the form of a computer game to participants. In the game, called Furniture Factory, participants manage a company that produces furniture. In two qualitative studies, we first elicit representations and heuristics with concurrent explanations and validate their use in post-hoc explanations. We analyze the complexity of the explanations given by participants to gain a deeper understanding of how complex cognitively adequate explanations should be. Based on insights from the analysis of the two qualitative studies, we formalize strategies that in combination can act as descriptors for participants’ behavior and optimal solutions. We match the strategies to decisions in a large behavioral dataset (<span><math><mrow><mo>></mo><mn>150</mn></mrow></math></span> participants) gathered in a third study, and compare the complexity of strategy combinations to the complexity featured in participants’ explanations. Based on the analyses from these three studies, we discuss how these insights can inform the automatic generation of cognitively adequate explanations in future AI systems.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"88 ","pages":"Article 101297"},"PeriodicalIF":2.1,"publicationDate":"2024-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142572379","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Technology roadmap toward the completion of whole-brain architecture with BRA-driven development","authors":"Hiroshi Yamakawa , Yoshimasa Tawatsuji , Yuta Ashihara , Ayako Fukawa , Naoya Arakawa , Koichi Takahashi , Yutaka Matsuo","doi":"10.1016/j.cogsys.2024.101300","DOIUrl":"10.1016/j.cogsys.2024.101300","url":null,"abstract":"<div><div>The development of brain-morphic software holds significant promise for creating artificial general intelligence that exhibits high affinity and interpretability for humans and also offers substantial benefits for medical applications. To facilitate this, creating Brain Reference Architecture (BRA) data, serving as a design specification for brain-morphic software is imperative. BRA-driven development, which utilizes Brain Information Flow (BIF) diagrams based on mesoscale brain anatomy and Hypothetical Component Diagrams (HCD) for corresponding computational functionalities, has been proposed to address this need. This methodology formalizes identifying possible functional structures by leveraging existing, albeit insufficient, neuroscientific knowledge. However, applying this methodology across the entire brain, thereby creating a Whole Brain Reference Architecture (WBRA), represents a significant research and development challenge due to its scale and complexity. Technology roadmaps have been introduced as a strategic tool to guide discussion, management, and distribution of resources within such expansive research and development activities. These roadmaps proposed a manual, anatomically based approach to incrementally construct BIF and HCD, thereby systematically expanding brain organ coverage toward achieving a complete WBRA. Large Language Model (LLM) technologies have introduced a paradigm shift, substantially automating the BRA-driven development process. This is largely due to the BRA data being structured around the brain’s anatomy and described in natural language, which aligns well with the capabilities of LLMs for supporting and automating the construction and verification processes. In this paper, we propose a novel technology roadmap to largely automate the creation of WBRA, leveraging neuroscientific insights. This roadmap includes 12 activities for automating BIF construction, notably extracting anatomical structures from scholarly articles. Furthermore, it details 11 activities aimed at enhancing the integration of Hypothetical Component Diagrams (HCD) into the WBRA, focusing on automating checks for functional consistency. This roadmap aims to establish a cost-effective and efficient design process for WBRA, ensuring the availability of brain-morphic software design specifications that are continually validated against the latest neuroscientific knowledge.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"88 ","pages":"Article 101300"},"PeriodicalIF":2.1,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142560689","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"心理学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}