Cognitive Systems Research最新文献

{"title":"Preventive mental health care: A complex systems framework for ambient smart environments","authors":"Ben White , Inês Hipólito","doi":"10.1016/j.cogsys.2023.101199","DOIUrl":"10.1016/j.cogsys.2023.101199","url":null,"abstract":"","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"84 ","pages":""},"PeriodicalIF":3.9,"publicationDate":"2023-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S138904172300133X/pdfft?md5=0f8a535d260cff4e2692d459919764c5&pid=1-s2.0-S138904172300133X-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138692614","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":"Improving deep learning with prior knowledge and cognitive models: A survey on enhancing explainability, adversarial robustness and zero-shot learning","authors":"Fuseini Mumuni ,&nbsp;Alhassan Mumuni","doi":"10.1016/j.cogsys.2023.101188","DOIUrl":"10.1016/j.cogsys.2023.101188","url":null,"abstract":"<div><p>We review current and emerging knowledge-informed and brain-inspired cognitive systems for realizing adversarial defenses, eXplainable Artificial Intelligence (XAI), and zero-shot or few-shot learning. Data-driven machine learning models have achieved remarkable performance and demonstrated capabilities surpassing humans in many applications. Yet, their inability to exploit domain knowledge leads to serious performance limitations in practical applications. In particular, deep learning systems are exposed to adversarial attacks, which can trick them into making glaringly incorrect decisions. Moreover, complex data-driven models typically lack interpretability or explainability, i.e., their decisions cannot be understood by human subjects. Furthermore, models are usually trained on standard datasets with a closed-world assumption. Hence, they struggle to generalize to unseen cases during inference in practical open-world environments, thus, raising the zero- or few-shot generalization problem. Although many conventional solutions exist, explicit domain knowledge, brain-inspired neural networks and cognitive architectures offer powerful new dimensions towards alleviating these problems. Prior knowledge is represented in appropriate forms like mathematical relations, logic rules, knowledge graphs, and large language models (LLMs). and incorporated in deep learning frameworks to improve performance. Brain-inspired cognition methods use computational models that mimic the human brain to enhance intelligent behavior in artificial agents and autonomous robots. Ultimately, these models achieve better explainability, higher adversarial robustness and data-efficient learning, and can, in turn, provide insights for cognitive science and neuroscience—that is, to deepen human understanding on how the brain works in general, and how it handles these problems.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"84 ","pages":"Article 101188"},"PeriodicalIF":3.9,"publicationDate":"2023-11-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138531994","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 multi-adaptive network model for human Hebbian learning, synchronization and social bonding based on adaptive homophily","authors":"Yelyzaveta Mukeriia ,&nbsp;Jan Treur ,&nbsp;Sophie Hendrikse","doi":"10.1016/j.cogsys.2023.101187","DOIUrl":"10.1016/j.cogsys.2023.101187","url":null,"abstract":"<div><p>This paper present a multi-adaptive network model integrating multiple adaptation mechanisms, specifically focusing on five types of such adaptation mechanisms. Two of them address first-order adaptation by learning of responding on others and first-order adaptation by bonding with others based on homophily. Three other adaptation mechanisms addressed are second-order adaptation of the speed of both Hebbian learning and bonding by homophily, and second-order adaptation of the homophily tipping point. The paper provides a comprehensive explanation of these concepts and their role in controlled adaptation within the diverse contextual scenarios of the paper.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"84 ","pages":"Article 101187"},"PeriodicalIF":3.9,"publicationDate":"2023-11-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1389041723001213/pdfft?md5=94183a8e8b500637904a9126d9bb7289&pid=1-s2.0-S1389041723001213-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138532052","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":"XAI Transformer based Approach for Interpreting Depressed and Suicidal User Behavior on Online Social Networks","authors":"Anshu Malhotra,&nbsp;Rajni Jindal","doi":"10.1016/j.cogsys.2023.101186","DOIUrl":"10.1016/j.cogsys.2023.101186","url":null,"abstract":"<div><p><span><span><span><span><span><span>Online social networks can be used for mental healthcare monitoring using </span>Artificial Intelligence<span> and Machine Learning techniques for detecting various </span></span>mental health disorders and corresponding risk assessment. Recent research in this domain has primarily been focused on leveraging deep </span>neural networks<span> and various Transformer based Large Language Models, which have now become state-of-the-art for most </span></span>natural language processing<span><span> and computational linguistic<span><span><span> tasks due to their unmatched prediction accuracy. Unlike conventional machine learning algorithms, these deep neural networks are black box architectures, where it is difficult to interpret and explain their predicted outcome. However, a black box classification outcome is insufficient for healthcare applications. Such systems will not be widely adopted and trusted by healthcare practitioners if they are not able to understand and explain the reasoning behind the predicted decisions made by an AI and ML based healthcare diagnostic system. The key objective of our research is to demonstrate the applications of model agnostic, post hoc surrogate </span>XAI techniques for providing explainability to classification decisions of pretrained LLMs (Transformers) based mental healthcare diagnostic systems fine-tuned (or trained) to detect depressive and suicidal </span>behavior using UGC from online social networks. For this, we have used the two most recent and popular techniques, SHAP and LIME. We have conducted extensive and in-depth experiments with four datasets and six pretrained LLMs, three of which have already been domain-adapted using mental health related datasets. We have also performed </span></span>Few Shot Learning experiments with these three pretrained mental health domain-adapted LLMs. The results of qualitative and </span></span>descriptive data analysis in this paper demonstrate that in order to build a comprehensive understanding of a person’s psychological state, emotion, and behavior and to discover the causes, symptoms, and triggers of mental health issues, it is essential to utilize e</span><em>X</em>pl<em>AI</em>nable <em>(XAI)</em> techniques with Transformer based LLMs (supervised). Alternatively, Transformer based unsupervised topic modeling technique BERTopic may be used for mental health risk monitoring and cause or symptom extraction when supervised training of LLMs is not feasible due to dataset annotation or availability challenges.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"84 ","pages":"Article 101186"},"PeriodicalIF":3.9,"publicationDate":"2023-11-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"138532064","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":"Relationship-specific and relationship-independent behavioural adaptivity in affiliation and bonding: A multi-adaptive dynamical systems approach","authors":"Sophie C.F. Hendrikse ,&nbsp;Jan Treur ,&nbsp;Sander L. Koole","doi":"10.1016/j.cogsys.2023.101182","DOIUrl":"10.1016/j.cogsys.2023.101182","url":null,"abstract":"<div><p>Humans often adapt their behaviour toward each other when they interact. From a neuroscientific perspective, such adaptivity can involve mechanisms based on adaptive connections (synaptic plasticity) and adaptive excitability thresholds (nonsynaptic plasticity) within the mental or neural network concerned. It is, however, often left unaddressed which of the types of adaptation are specific for the relationship and which are more general for multiple relationships. We focus on this differentiation between relationship-specific and relationship-independent adaptation in social interactions. We analysed computationally how an interplay of adaptive relation-specific and relation-independent mechanisms occurs within the causal pathways for social interaction. As part of this, we cover also the context-sensitive control of these types of adaptation (adaptive speeds and strengths of adaptation), which is sometimes termed higher-order adaptation or metaplasticity. The model was evaluated by a number of explored runs where within a group of four agents each agent randomly has episodes of interaction with one of the three other agents. The outcomes of the analysis of the (stochastic) simulation results show a strong dependence of adaptation on the extent of social interaction: more social interaction leads to more adaptation of the interaction behaviour. This holds both for the short-term and long-term first-order adaptation and for the second-order adaptation, which is long-term.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"84 ","pages":"Article 101182"},"PeriodicalIF":3.9,"publicationDate":"2023-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S138904172300116X/pdfft?md5=08091b812cf5784f11b2664d81cfcecb&pid=1-s2.0-S138904172300116X-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"135763353","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":"Semantic configuration model with natural transformations","authors":"Viacheslav Wolfengagen ,&nbsp;Larisa Ismailova ,&nbsp;Sergey Kosikov ,&nbsp;Igor Slieptsov ,&nbsp;Sebastian Dohrn ,&nbsp;Alexander Marenkov ,&nbsp;Vladislav Zaytsev","doi":"10.1016/j.cogsys.2023.101185","DOIUrl":"https://doi.org/10.1016/j.cogsys.2023.101185","url":null,"abstract":"<div><p>In the present work, efforts have been made to create a configuration-based approach to knowledge extraction. The notion of granularity is developed, which allows fine-tuning the expressive possibilities of the semantic network. As known, the central issues for knowledge-based systems are what’s-in-a-node and what’s-in-a-link. As shown, the answer can be obtained from the functor-as-object representation. Then the nodes are functors, and the main links are natural transformations. Such a model is applicable to represent morphing, and the object is considered as a process, which is in a harmony with current ideas on computing. It is possible to represent information channels that carry out the transformations of processes. The possibility of generating displaced concepts and the generation of families of their morphs is shown, the evolvent of stages of knowledge and properties of the process serve as parameters.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"83 ","pages":"Article 101185"},"PeriodicalIF":3.9,"publicationDate":"2023-11-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"92066729","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":"Building a cognitive system based on process interaction","authors":"Viacheslav E. Wolfengagen ,&nbsp;Larisa Ismailova ,&nbsp;Sergey Kosikov","doi":"10.1016/j.cogsys.2023.101183","DOIUrl":"https://doi.org/10.1016/j.cogsys.2023.101183","url":null,"abstract":"<div><p>According to modern notions, computing is not separable from cognitive modeling and activity. This paper continues the tradition of the uniform approach and proposes a small number of general mechanisms that cope with the main known effects of computing as a science — the interaction of objects-as-processes, the interaction of processes with the environment, generalized interaction. As shown, the applicative prestructure (objects-as-processes, application) generates an applicative structure (processes, application, values), which ensures the generation of the result — the value of interactions, enabling the process of evaluation. The theory of combinators is used as the main (meta)mathematical means. A diagram mechanism has been developed that implements the emerging applicative computational system of object interaction and reflects the arity of accompanying the induced information processes. The processes are bidirectional in nature, both with a decrease in arity – reduction, and with an increase in arity – expansion.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"83 ","pages":"Article 101183"},"PeriodicalIF":3.9,"publicationDate":"2023-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"92046585","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":"Mind surfing: Attention in musical absorption","authors":"Simon Høffding ,&nbsp;Nanette Nielsen ,&nbsp;Bruno Laeng","doi":"10.1016/j.cogsys.2023.101180","DOIUrl":"https://doi.org/10.1016/j.cogsys.2023.101180","url":null,"abstract":"<div><p>Literature in the psychology of music and in cognitive psychology claims – paradoxically – that musical absorption includes processes of both focused attention and mind wandering. We examine this paradox and aim to resolve it by integrating accounts from cognitive psychology on attention and mind wandering with qualitative phenomenological research on some of the world’s most skilled musicians. We claim that a mode of experience that involves intense attention and what superficially seems like mind wandering is possible. We propose to grasp this different mode of experience with a new concept: “mind surfing”. We suggest that a conjoined consideration of attention’s intensive and selective capacities can partially explain how one can be both focused and freely “surfing” on a “musical wave” at the same time. Finally, we couple this novel and foundational work on attention with a 4E cognition account to show how music acts as an affective and cognitive scaffold, thereby enabling the surfing.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"83 ","pages":"Article 101180"},"PeriodicalIF":3.9,"publicationDate":"2023-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S1389041723001146/pdfft?md5=ea568a4f3adf4702b3d7ff4de196c295&pid=1-s2.0-S1389041723001146-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"109145914","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":"Transformability, generalizability, but limited diffusibility: Comparing global vs. task-specific language representations in deep neural networks","authors":"Yanru Jiang ,&nbsp;Rick Dale ,&nbsp;Hongjing Lu","doi":"10.1016/j.cogsys.2023.101184","DOIUrl":"https://doi.org/10.1016/j.cogsys.2023.101184","url":null,"abstract":"<div><p>This study investigates the integration of two prominent neural network representations into a hybrid cognitive model for solving a natural language task, where pre-trained large-language models serve as global learners and recurrent neural networks offer more “local” task-specific representations in the neural network. To explore the fusion of these two types of representations, we employ an autoencoder to transform them between each other or fuse them into a single model. Our exploration identifies a computational constraint, which we term <em>limited diffusibility</em>, highlighting the limitations of hybrid systems that operate with distinct types of representation. The findings from our hybrid system confirm the crucial role of global knowledge in adapting to a new learning task, as having only local knowledge greatly reduces the system’s transferability.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"83 ","pages":"Article 101184"},"PeriodicalIF":3.9,"publicationDate":"2023-11-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91964484","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":"Human-inspired goal reasoning implementations: A survey","authors":"Ursula Addison","doi":"10.1016/j.cogsys.2023.101181","DOIUrl":"https://doi.org/10.1016/j.cogsys.2023.101181","url":null,"abstract":"<div><p>Goal reasoning is the ability of an artificial system to reason over its goals; it can identify, manage, plan, and execute its goals. In complex environments where requirements could change often, goal reasoning functionality is essential. Goal reasoning agents may rely on a motivation system to guide the goal reasoning process; we refer to such agents as motivated agents. Motivated agents can be explicitly or implicitly motivated by external or internal motivations. While the bulk of goal reasoning work has focused on agents that have implicit external motivations, internal motivations may offer some unique benefits to goal reasoning. As artificial internal motivations have a natural analogue to the human motivation system, this work investigates recent advances in motivated agents, where motivations are modeled on the human integrated-self. In this survey, we review those goal reasoning systems whose meta-reasoning and other goal reasoning subprocesses are at least in part intrinsic or identified, i.e., arising from idiosyncratic factors such as identity, a value system, emotions, experiences and so forth. For each system surveyed we evaluate its goal reasoning processes according to an analysis framework. We use our findings to draw conclusions about the potential benefits the three self-system categories: motives, mental simulation, and emotion bring to the goal reasoning paradigm.</p></div>","PeriodicalId":55242,"journal":{"name":"Cognitive Systems Research","volume":"83 ","pages":"Article 101181"},"PeriodicalIF":3.9,"publicationDate":"2023-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134657004","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}