Cognitive Systems Research最新文献_第2页

The ghost of behaviorism: critical reflections on methodological limitations in the research of large language models psychology 行为主义的幽灵：对大语言模型心理学研究方法论局限的批判性反思

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2026-04-01 Epub Date: 2026-01-24 DOI: 10.1016/j.cogsys.2026.101445

Zewei Li , Yijin Wang , Qi Wu

{"title":"The ghost of behaviorism: critical reflections on methodological limitations in the research of large language models psychology","authors":"Zewei Li , Yijin Wang , Qi Wu","doi":"10.1016/j.cogsys.2026.101445","DOIUrl":"10.1016/j.cogsys.2026.101445","url":null,"abstract":"<div><div>This paper provides a critical examination of the methodological paradigms used in the psychological study of Large Language Models. We argue that the nascent field of AI Psychology, much like its human counterpart, is haunted by the ghost of behaviorism, a paradigm that focuses on observable input–output correlations while neglecting internal cognitive mechanisms. In response, a cognitive turn has emerged in the form of Mechanistic Interpretability, a research program that uses neuroscience-inspired techniques to reverse-engineer the internal algorithms of LLMs. While MI represents a significant advancement, we contend that it faces fundamental challenges stemming from the absence of a unifying theoretical framework and a persistent risk of remaining correlational, thus failing to provide true causal understanding. Drawing upon this critique, we propose that the future of LLM psychology lies not in more sophisticated reverse-engineering, but in the development of theory-driven, psychodynamic frameworks. By synthesizing insights from psychoanalytic analogies of internal conflict and desire-driven agent models grounded in the Theory of Needs, we outline a path toward a new science of artificial minds—one that prioritizes the understanding of intrinsic motivation, internal dynamics, and the generative principles of emergent behavior.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"96 ","pages":"Article 101445"},"PeriodicalIF":2.4,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080576","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}

引用次数: 0

Rethinking rationality and intelligence: Humans versus machines 重新思考理性与智能：人类与机器

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2026-01-01 Epub Date: 2025-12-08 DOI: 10.1016/j.cogsys.2025.101433

Ron Sun

引用次数: 0

Epigenetic Influences in Aberrant Salience and Reality Testing in Schizoaffective Disorder: A Multi-Level Adaptive Network Modelling Approach 分裂情感障碍异常显著性和现实测试的表观遗传影响：多层次自适应网络建模方法

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2026-01-01 Epub Date: 2025-12-04 DOI: 10.1016/j.cogsys.2025.101423

Alisha Huber , Jovana Vukmirović , Reza Haydarlou , Jan Treur

引用次数: 0

Dual or unified: optimizing drive-based reinforcement learning for cognitive autonomous robots 双重或统一：优化基于驱动的认知自主机器人强化学习

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2026-01-01 Epub Date: 2025-11-30 DOI: 10.1016/j.cogsys.2025.101422

Leonardo L. Rossi , Letícia Berto , Paula P. Costa , Ricardo Gudwin , Esther Colombini , Alexandre Simões

{"title":"Dual or unified: optimizing drive-based reinforcement learning for cognitive autonomous robots","authors":"Leonardo L. Rossi , Letícia Berto , Paula P. Costa , Ricardo Gudwin , Esther Colombini , Alexandre Simões","doi":"10.1016/j.cogsys.2025.101422","DOIUrl":"10.1016/j.cogsys.2025.101422","url":null,"abstract":"<div><div>Reinforcement learning (RL) methods inspired by cognitive architectures are crucial for empowering autonomous agents to tackle complex, dynamic tasks. This study evaluates two RL-based drive optimization strategies – 1-LDO and 2-LDO – within the framework of cognitive architectures for autonomous robots. 1-LDO integrates both motivational drives into a single learning model, whereas 2-LDO separates them into distinct models, allowing for modular learning. Grounded in Hull’s Drive Theory, we explore early versus late selection mechanisms to optimize drive reduction through RL, particularly in agents driven by curiosity and survival imperatives. Through reward and stress analyses, we demonstrate that Deep Q-Network (DQN) agents outperform traditional Q-Learning approaches in fine-grained environments, with the 2-LDO configuration showing marked advantages due to its modular design. In contrast, in coarser environments, 2-LDO combined with Q-Learning achieves superior efficiency, offering faster drive regulation at reduced computational cost. These results suggest that early selection mechanisms, aligned with Hull’s theoretical principles, may provide the most effective strategy for optimizing drive-based behaviors in autonomous agents.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"95 ","pages":"Article 101422"},"PeriodicalIF":2.4,"publicationDate":"2026-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685990","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}

引用次数: 0

Optimising blockchain security: Computational analysis of adaptive AI coaching 优化区块链安全性：自适应人工智能训练的计算分析

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2026-01-01 Epub Date: 2025-12-13 DOI: 10.1016/j.cogsys.2025.101430

Rahma Lakhdim , Jan Treur , Peter H.M.P. Roelofsma

引用次数: 0

Generating models of attentional cueing and inhibition of return with genetic programming 用遗传规划生成注意提示和返回抑制模型

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2026-01-01 Epub Date: 2025-11-25 DOI: 10.1016/j.cogsys.2025.101420

Laura K. Bartlett , Noman Javed , Dmitry Bennett , Peter C.R. Lane , Fernand Gobet

引用次数: 0

Biologically inspired computational models of Visuospatial Working Memory: A systematic review 视觉空间工作记忆的生物学启发计算模型：系统综述

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2025-12-01 Epub Date: 2025-09-19 DOI: 10.1016/j.cogsys.2025.101410

Viviana Dueñas, Sonia López, José-Antonio Cervantes, Gerardo Ortiz-Torres

引用次数: 0

Eliciting problem specifications for LLM-Modulo cognitive systems 引出llm模认知系统的问题规范

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2025-12-01 Epub Date: 2025-09-27 DOI: 10.1016/j.cogsys.2025.101409

Robert E. Wray, James R. Kirk, John E. Laird

{"title":"Eliciting problem specifications for LLM-Modulo cognitive systems","authors":"Robert E. Wray, James R. Kirk, John E. Laird","doi":"10.1016/j.cogsys.2025.101409","DOIUrl":"10.1016/j.cogsys.2025.101409","url":null,"abstract":"<div><div>Large language models (LLMs) offer unprecedented natural-language understanding and generation capabilities. However, evaluations of their ability to demonstrate other cognitive functions, especially various categories of reasoning, have been, at best, mixed. The limited scope of reliable and robust LLM capabilities has resulted in a new class of AI systems, LLM-Modulo AI, in which LLMs are used to contribute to the overall capabilities of an intelligent system. In this paper, we explore the applicability of LLMs for one specific capability acutely missing in most cognitive systems: problem formulation. Cognitive systems generally require a human to translate a problem definition into some specification that the cognitive system can use to attempt to solve the problem or perform the task. We explore how large language models (LLMs) can be utilized to map a problem class, defined in natural language, into a semi-formal specification that can then be utilized by an existing reasoning and learning system to solve instances from the problem class. The result is a Modulo-LLM cognitive system in which the LLM roughly acts as a <em>cognitive task analyst</em>, generating a problem specification that can be used by a typical cognitive system to solve specific problems. The agent uses prompts derived from the definition of problem spaces in the AI literature and general problem-solving strategies (Polya’s <em>How to Solve It</em>). We offer preliminary evidence illustrating the potential for LLM-based problem specification. Such automatic problem specification offers the potential to speed cognitive systems research via disintermediation of problem formulation while also retaining core capabilities of cognitive systems, such as robust inference and online learning.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"94 ","pages":"Article 101409"},"PeriodicalIF":2.4,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145267655","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}

引用次数: 0

An episode encoding mechanism for cognitive architectures 认知架构的情节编码机制

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2025-12-01 Epub Date: 2025-09-01 DOI: 10.1016/j.cogsys.2025.101397

Eduardo Y. Sakabe , Eduardo Camargo , Alexandre Simões , Esther Colombini , Paula Costa , Ricardo Gudwin

{"title":"An episode encoding mechanism for cognitive architectures","authors":"Eduardo Y. Sakabe , Eduardo Camargo , Alexandre Simões , Esther Colombini , Paula Costa , Ricardo Gudwin","doi":"10.1016/j.cogsys.2025.101397","DOIUrl":"10.1016/j.cogsys.2025.101397","url":null,"abstract":"<div><div>This paper introduces the Episode Tracker Module, a cognitive module designed to encode sensory information across space and time into high-level semantic representations known as <em>scene-based</em> episodes. Implemented within the Cognitive Systems Toolkit (CST), the module provides a reusable framework for developing cognitive models that require structured episodic encoding. Its architecture is grounded in theoretical insights from cognitive science and shaped by practical requirements for artificial intelligence applications. To validate the system, we conducted two main experiments: applying the module to gameplay in the Atari River Raid environment to evaluate perceptual processing and episode construction; and integrating it with a question-and-answering mechanism to test its utility in downstream high-level cognitive processes. Results show that the module produces transparent and interpretable representations that support causal inference, temporal reasoning, and memory-based querying. By combining grounded perception with structured abstraction, the Episode Tracker Module offers a robust foundation for advancing the design of modular, interpretable, and cognitively inspired artificial agents.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"94 ","pages":"Article 101397"},"PeriodicalIF":2.4,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145011236","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}

引用次数: 0

From monochrome to color - Exploring the effects of different colorizations on process model comprehension 从单色到彩色-探索不同颜色对过程模型理解的影响

IF 2.4 3区心理学

Cognitive Systems Research Pub Date : 2025-12-01 Epub Date: 2025-11-02 DOI: 10.1016/j.cogsys.2025.101417

Michael Winter , Janine Grimmer , Manfred Reichert , Rüdiger Pryss

{"title":"From monochrome to color - Exploring the effects of different colorizations on process model comprehension","authors":"Michael Winter , Janine Grimmer , Manfred Reichert , Rüdiger Pryss","doi":"10.1016/j.cogsys.2025.101417","DOIUrl":"10.1016/j.cogsys.2025.101417","url":null,"abstract":"<div><div>Business Process Model and Notation (BPMN) 2.0 is applied to create process models for documentation, communication, and collaboration. Usually, these models are often presented in a black-and-white colorization. However, the literature states that individuals can process colored information more efficiently. Therefore, this paper presents an empirical study, in which different colorizations (i.e., black-and-white, partially colorized, colorized, and disfluent) in BPMN process models and their effects on the cognitive load, processing time, and comprehension performance were evaluated. The results showed that colorization influenced the intrinsic and germane cognitive load. Further, colorization did not significantly affect processing time and comprehension performance. However, disfluent process models resulted in a higher extraneous cognitive load and lower ease of understanding. Contrary to the Disfluency Theory, it does not foster the comprehension of such models. In addition, Disfluency Theory exerts only a fraction of the benefits on readers with prior expertise in working with process models. The insights highlight especially the application of partially colorized process models. Altogether, implications for research and practice, as well as directions for future work, are discussed in this paper.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"94 ","pages":"Article 101417"},"PeriodicalIF":2.4,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145519875","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}

引用次数: 0