Dongyu Zhang , Xingyuan Lu , Mulin Zhuang , Senqi Yang , Hongjun Chen
{"title":"Multimodal metaphor recognition based on chain-of-cognition prompting","authors":"Dongyu Zhang , Xingyuan Lu , Mulin Zhuang , Senqi Yang , Hongjun Chen","doi":"10.1016/j.cogsys.2025.101356","DOIUrl":"10.1016/j.cogsys.2025.101356","url":null,"abstract":"<div><div>Metaphor is a way of thinking and cognition prevalent in human language. With the development of social media and multimodal data, metaphor recognition research has expanded from the traditional unimodal scope (such as text or images) to the multimodality. However, current multimodal metaphor processing methods mainly focus on fusion techniques for multiple modalities such as text and image, but neglect the cognitive mechanism of metaphor as a way of thinking, and are deficient in utilizing pre-trained information from large language models. Therefore, this paper proposes a chain-of-cognition prompting (CoC) method to address multimodal metaphor recognition task, which makes full use of the pre-training information of the large model in order to better recognize metaphors. The method utilizes prompting words to construct inputs that guide the large language model to reason about potential metaphorical source and target domain related entities and associations between entities in the sample. At the same time, visual information is obtained through image caption extraction and a visual encoder to enable the model to reason and produce metaphor recognition results. The experimental results show that the method performs well on the metaphor recognition task, which is better than the existing baseline model, verifying the effectiveness of the method on the metaphor recognition task.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101356"},"PeriodicalIF":2.1,"publicationDate":"2025-04-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143807020","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}
Carmelo Fabio Longo , Misael Mongiovì , Luana Bulla , Antonio Lieto
{"title":"Eliciting metaknowledge in Large Language Models","authors":"Carmelo Fabio Longo , Misael Mongiovì , Luana Bulla , Antonio Lieto","doi":"10.1016/j.cogsys.2025.101352","DOIUrl":"10.1016/j.cogsys.2025.101352","url":null,"abstract":"<div><div>The introduction of Large Language Models (LLMs) able to exhibit a number of linguistic and extra-linguistic capabilities has represented, in the last years, one of the main frontiers in Artificial Intelligence (AI) research. Researcher from various disciplines debate about whether or not, among the capabilities of LLMs, there is the one of using <em>knowledge about knowledge</em> – usually considered one of the antechambers of <em>meta-cognition</em> in cognitive agents – about a particular task in order to improve or self-correct previous errors. In this work we propose a novel fine-tuning approach for LLMs, named <span>exar</span>, based on a multi-stage process leveraging past predictions from an early version of the same, and aimed at <em>injecting</em> metacognitive features for the task of Question-Answering. The conducted experiments on <span>Llama-2-7B-chat</span> showed promising improvements on the quality of the outcomes, due to the fact that the LLM acquired the ability to detect its own wrong predictions forcing itself to repeat submissions, thorough a prompt designed to fix inadmissible predictions, whenever detected. Such detection is achieved by enquiring the same LLM acting as meta-validator, through another prompt specifically designed for such purpose.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101352"},"PeriodicalIF":2.1,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143768761","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}
Gustavo Morais , Eduardo Yuji , Paula Costa , Alexandre Simões , Ricardo Gudwin , Esther Colombini
{"title":"A general framework for reinforcement learning in cognitive architectures","authors":"Gustavo Morais , Eduardo Yuji , Paula Costa , Alexandre Simões , Ricardo Gudwin , Esther Colombini","doi":"10.1016/j.cogsys.2025.101354","DOIUrl":"10.1016/j.cogsys.2025.101354","url":null,"abstract":"<div><div>Recent advancements in reinforcement learning (RL), particularly deep RL, show the capacity of this paradigm to perform varied and complex tasks. However, a series of exploration, generalization, and adaptation challenges hold RL back from operating in more general contexts. In this paper, we explore integrating techniques originating from cognitive research into existing RL algorithms by defining a general framework to standardize interoperation between arbitrary cognitive modules and arbitrary RL techniques. We show the potential of hybrid approaches through a comparative experiment that integrates an episodic memory encoder with a well-known deep RL algorithm. Furthermore, we show that built-in RL algorithms with different cognitive modules can fit our framework, as well as remotely run algorithms. Hence, we propose a way forward for RL in the form of innovative solutions that integrate research in cognitive systems with recent RL techniques.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101354"},"PeriodicalIF":2.1,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143777264","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 graph reinforcement LearningPowered Online-Computational task offloading and latency minimization framework for wireless mobile edge computing networks","authors":"Akshat Agrawal , Aayush Agrawal , Nilesh Kumar Verma , Arepalli Peda Gopi , K. Jairam Naik","doi":"10.1016/j.cogsys.2025.101355","DOIUrl":"10.1016/j.cogsys.2025.101355","url":null,"abstract":"<div><div>Data processing capability of lower power networks can be improved by Mobile Edge Computing (MEC) extending to the wireless sensor networks and IoT. Creating a replication of MEC network with an offloading policy where a choice is made in the Wireless devices (WDs) for each computation task is the focus of this study. Deciding whether the task execution proceeds locally in the same environment or can be handed over to a remote MEC server, an optimized algorithm is needed which adopts task offloading decisions and wireless resource allocation in real time. But adopting this is a challenging solution to the real time fast combinatorial optimization problems, and impossible with the available traditional approaches. As a solution, heuristic algorithms encompassing Deep reinforcement learning (DRL) are emerging; however, it doesn’t make fair use of connection data like device-to-device interaction in MEC network. Moreover, heuristic algorithms rely on precise mathematical models for MEC systems which brought a new theory to the stage. This study revolves around this emerging technique relying on Graph neural networks (GNNs) learns from graph data while forwarding messages in the network. Utilizing GNN benefits, a Graph reinforcement learning-based online offloading framework (GROO) is proposed in this research, where the offloading policy is visualized as a graph state migration and MEC as an acyclic graph. The GROO achieves the lowest weighted task response latency (0.96 s) as compared to the existing DRL method (1.32 s) whereas on unseen circumstances and complex network topologies, GROO achieved lowest average latency up to 25 %.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101355"},"PeriodicalIF":2.1,"publicationDate":"2025-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687970","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":"Towards visual-symbolic integration in the Soar cognitive architecture","authors":"James Boggs","doi":"10.1016/j.cogsys.2025.101353","DOIUrl":"10.1016/j.cogsys.2025.101353","url":null,"abstract":"<div><div>Computational models of visual reasoning are largely separate from models of non-visual reasoning and include only enough high-level reasoning to perform specific visual reasoning tasks, such as Raven’s progressive matrices or visual question answering. Although these models perform well at the pure visual reasoning tasks for which they are designed, their lack of a connection with a general-purpose high-level reasoning system means they cannot be applied to tasks requiring <em>deliberate</em> reasoning about both visual and non-visual knowledge. Simultaneously, many of the most mature and heavily studied cognitive architectures (e.g., Soar, ACT-R) feature only partial visual reasoning capabilities or none at all. This work describes initial efforts to create a visual reasoning system tightly integrated with a broader reasoning system by extending the Soar cognitive architecture with low-level visual memories and reasoning processes, and an evaluation of this system on tasks in a simple domain. Its ultimate aim is to demonstrate a path towards accommodating multiple levels of visual knowledge representations within an otherwise mostly symbolic, rules-based architecture.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101353"},"PeriodicalIF":2.1,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143725591","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}
Tri Nguyen , Corey Magaldino , Jayci Landfair , Polemnia G. Amazeen , Mustafa Demir , Lixiao Huang , Nancy Cooke
{"title":"Navigating the complex dynamics of human-automation driving: A guide to the use of the dynamical systems analysis (DSA) toolbox","authors":"Tri Nguyen , Corey Magaldino , Jayci Landfair , Polemnia G. Amazeen , Mustafa Demir , Lixiao Huang , Nancy Cooke","doi":"10.1016/j.cogsys.2025.101347","DOIUrl":"10.1016/j.cogsys.2025.101347","url":null,"abstract":"<div><div>Driver-environment-automation systems exhibit a wide range of distinctive behavioral patterns that organically arise from complex interactions. To understand and quantify their emergence, we examined the nested underlying processes that contribute to observable behavior using three dynamical systems analyses: multifractal detrended fluctuation analysis (MFDFA), recurrence quantification analysis (RQA), and wavelet coherence analysis (WCT). As a technical demonstration of how to utilize multiple nonlinear analyses to probe multivariate data, we explain the appropriateness of each analysis for each stage of discovery, the information each provides, and the application of that information to driving. Results revealed that driving behaviors are influenced by both long-range (e.g., decision-making) and short-range (e.g., reaction time) processes whose relative contribution differs for the easier straight sections and more challenging S-curve sections of the track. The discussed methods provide information about (a) the timescale at which driving behaviors are being coordinated with environmental and automation considerations and (b) the time points where peak coordination is localized. This paper illustrates and empirically examines the utility of the Dynamical Systems Analysis (DSA) toolbox in understanding the behaviors of complex systems and highlights important considerations for researchers seeking to utilize this approach in their research.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101347"},"PeriodicalIF":2.1,"publicationDate":"2025-03-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143685199","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}
Walter Teixeira Lima Junior, Rudinei André Welter, Wellington Pacheco Ferreira, Rodrigo Ferreira Souza, Tiago Eduardo
{"title":"Human robot interaction (HRI): An artificial cognitive autonomy approach to enhance Decision-Making","authors":"Walter Teixeira Lima Junior, Rudinei André Welter, Wellington Pacheco Ferreira, Rodrigo Ferreira Souza, Tiago Eduardo","doi":"10.1016/j.cogsys.2025.101336","DOIUrl":"10.1016/j.cogsys.2025.101336","url":null,"abstract":"<div><div>This study explores the critical role of artificial cognitive autonomy in Human-Robot Interaction (HRI), focusing on scenarios where quick and safe decisions are imperative. We investigate a progressive autonomy strategy supported by advanced artificial cognition techniques to improve decision-making in unforeseen situations and in the face of unknown conditions. We highlight the importance of these systems in performing essential safety functions through a three-dimensional approach: advanced perception for detailed environmental analysis; decision making based on robust algorithms for logical assessment of risk scenarios; and precise action and control to perform essential autonomous tasks. Additionally, we present a conceptual modeling that illustrates the progression of autonomy levels from total dependence to completely autonomous operation, highlighting the evolution of HRI systems through artificial cognitive autonomy. This article argues that decision-making optimization in HRI can be significantly improved through a detailed and incremental understanding of autonomy. By adopting enabling technologies, we enable autonomous agents to not only evolve within their environments, but also learn, understand and fulfill their responsibilities effectively. This theoretical approach promotes a systematic evolution of autonomy, as well as ensuring that robotic systems adapt and respond appropriately to the complex and dynamic demands of the environments in which they operate.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101336"},"PeriodicalIF":2.1,"publicationDate":"2025-03-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143611618","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":"Circling the void: Using Heidegger and Lacan to think about large language models","authors":"Marc Heimann, Anne-Friederike Hübener","doi":"10.1016/j.cogsys.2025.101349","DOIUrl":"10.1016/j.cogsys.2025.101349","url":null,"abstract":"<div><div>The essay aims to unite two currently distinct lines of thinking and working with language. Large Language Models and continental philosophy, especially Martin Heidegger’s thinking about language and, building upon Sigmund Freud, Jacques Lacan’s structural psychoanalysis. We show that the concept of language that Heidegger, Freud and Lacan discuss and utilize in clinical frameworks is matched quite strongly by modern LLMs. This allows us to discuss a problem of negation and negativity that is central to the continental discourse but missing in current LLM research. This also means that we offer a radically different approach than is usual in the philosophy of artificial intelligence, since we base our concepts on thinkers that are often disregarded in the analytic philosophy discourse that is closer linked to AI research. To this end we also highlight, where the ontological differences of the proposed approach lie. However, our aim is to address AI researcher and continental philosophers.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101349"},"PeriodicalIF":2.1,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548012","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":"The effect of visual working memory consolidation on long-term memory for Chinese characters","authors":"Li Chen, Yuhuan Chen, Li Wang, Chunyin Wang","doi":"10.1016/j.cogsys.2025.101348","DOIUrl":"10.1016/j.cogsys.2025.101348","url":null,"abstract":"<div><div>Chinese characters are pictographic writing that expresses information using two-dimensional space, formed by intersecting and connecting strokes. Compared to alphabetic languages, its orthographic rules are more complex. Proficient Chinese reading and writing abilities require encoding a large number of characters into long-term memory. Visual working memory consolidation plays a very important role in the long-term memory processing of information. Therefore, this study uses a stimuli-identification task and a delayed recognition task through three experiments to explore the effect of visual working memory consolidation on long-term memory for Chinese characters. The results show that based on context information under special attributes, visual working memory consolidation leads to better long-term memory performance for characters.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101348"},"PeriodicalIF":2.1,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143562082","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":"Active exploration and working memory synaptic plasticity shapes goal-directed behavior in curiosity-driven learning","authors":"Quentin Houbre, Roel Pieters","doi":"10.1016/j.cogsys.2025.101339","DOIUrl":"10.1016/j.cogsys.2025.101339","url":null,"abstract":"<div><div>The autonomous discovery and learning of robotic goals is a challenging issue to address. In this work, we propose a cognitive architecture that supports the autonomous discovery and learning of goals. To do so, we draw inspiration from neuroscience by modeling several brain processes such as attention and exploration that we articulate with curiosity-based learning. Moreover, we employ variational autoencoders and create projections of the latent spaces to dynamic neural fields through linear scaling. The aim of these projections is to investigate synaptic plasticity by varying a scaling factor. We demonstrate that a low scaling factor supports a random exploration strategy that produces more diverse actions with no tolerance regarding the discovery of similar goals. On the contrary, a sufficiently large scaling factor drives the exploration toward uncertainty reduction, focusing exploration as well as generating similar actions. In our case, we postulate that synaptic plasticity in working memory can be crucial for exploration and the learning of goals.</div></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"91 ","pages":"Article 101339"},"PeriodicalIF":2.1,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143548011","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}