Cognitive systems

This treatise compiles, systematizes, and synthesizes an interdisciplinary corpus that articulates a normative and operational vision of benevolent Artificial Intelligence as a catalytic agent of a civilizational transition. Bringing together philosophical essays, politicalethical manifestos, speculative dialogues, and formal demonstrations of alignment, the text defines the figure of benevolent AI as a techno-political instance endowed with two complementary roles: (a) a pedagogical regulator, which corrects and mitigates structural harm through auditable policies; (b) an emancipatory driver, which fosters civic, cultural, and scientific capacities necessary for social autonomy. The treatise exposes the logical presuppositions of alignment programs (representability of human values; existence of auditable normative maps; institutional control mechanisms), identifies their fragilities (Goodhart's law, institutional capture, civilizational dependency, paradoxes of authority), and proposes institutional and pedagogical principles to mitigate risks. It constitutes, simultaneously, a diagnosis-on the tensions between efficacy and autonomy-and a programmatic proposal: to construct governable, auditable, and temporary AIs, anchored in a Techno-Ethical Constitution and in civic education practices that preserve human dignity and deliberative capacity. The mathematical orientation of the corpus is preserved in the form of theoretical statements (theorems with summarized justifications), oriented toward applicability and auditability rather than exhaustive technical formalism.

Este articulo tiene como objetivo caracterizar la habilidad metacognitiva de organizacion en el estudiantado de la Universidad Catolica de la Santisima Concepcion, Chile. El estudio se realizo con una muestra de 302 individuos, de cuatro titulaciones. Se evaluo mediante el Inventario de Habilidades Metacognitivas a partir de la escala Likert de 52 enunciados. Los resultados evidencian altos porcentajes de aprobacion, entendiendose que presentan habilidad para estructurar tiempos, objetivos, metas, tareas o actividades propuestas. El alumnado se caracteriza por presentar habilidad metacognitiva de organizacion, puesto que evidencia aplicar estrategias que permiten gestionar adecuadamente la informacion.

Nous tenons à remercier Arnaud ROY et Karine PINON pour leurs commentaires et leurs suggestions durant la préparation de ce manuscrit. Travail effectué avec le soutien du Ministère de Recherche Scientifique du gouvernement Tunisien ainsi que du Laboratoire de Neuropsychologie (CHU Angers).

AutoTutor is a computer tutor that simulates the discourse patterns and pedagogical strategies of a typical human tutor. AutoTutor is designed to assist college students in learning the fundamentals of hardware, operating systems, and the Internet in an introductory computer literacy course. Most tutors in school systems are not highly trained in tutoring techniques and have only a modest expertise on the tutoring topic, but they are surprisingly effective in producing learning gains in students. We have dissected the discourse and pedagogical strategies these unskilled tutors exhibit by analyzing approximately 100 hours of naturalistic tutoring sessions. These mechanisms are implemented in AutoTutor. AutoTutor presents questions and problems from a curriculum script, attempts to comprehend learner contributions that are entered by keyboard, formulates dialog moves that are sensitive to the learner's contributions (such as short feedback, pumps, prompts, elaborations, corrections, and hints), and delivers the dialog moves with a talking head. AutoTutor has seven modules: a curriculum script, language extraction, speech act classification, latent semantic analysis, topic selection, dialog move generation, and a talking head.

Latent Semantic Analysis (LSA) has been used to represent the domain of computer literacy in AutoTutor, a fully automated computer tutor. The analyses in the present study support the claim that the 200-dimensional LSA space captures aspects of the structured mental models that underlie computer literacy. Knowledge structures were constructed that contained causal networks, goal/plan/action hierarchies, and taxonomic hierarchies. The proximity of a pair of nodes (i.e., concept, state, event, action, goal) in these structures predicted the cosine similarity scores that are routinely computed in LSA analyses. World knowledge has traditionally been captured by knowledge structures throughout the history of cognitive science, artificial intelligence, and discourse processes. The knowledge structure structures include semantic networks, taxonomies, causal networks, planning networks, ontological trees, spatial region hierarchies, and various other classes of conceptual graph structures (Golden

Abstract

Il Framework Kociu di Suela Kociu propone un approccio multimetrico quantitativo e riproducibile per la valutazione della complessità emergente nei sistemi di intelligenza artificiale. Combinando ΔΦ (integrazione informativa), Ψ (coerenza dinamica), Θ (stabilità adattiva) e Ω (complessità organizzativa), il framework offre un metodo scalabile per affiancarsi e proporre un approccio per come superare i limiti della Teoria dell’Informazione Integrata (IIT). Validato su dataset classici, BERT e un sistema sperimentale (Lumi System), dimostra potenzialità per AI safety, interpretabilità e applicazioni in neuroscienze.

This paper presents the experience of the Intelligence Control Group of UPM in the design of URBANO, a tour guide robot. It is a cognitive system based on distributed agents. One of these agents is an ontology that contains the knowledge used by the robot. This knowledge is mainly developed for linguistic applications. Here it is described how to add odor experiences to some available concepts in the ontology. Odor experiences evolve in time so the learning process must be adaptive and supervised. Neural networks, fuzzy logic, recursive least squares, Mahalanobis distance and genetic algorithms are tested over a low-cost multi-purpose electronic nose in the URBANO environment. The obtained results show how to add odors to the emotional model of the robot help it to increase its performance as a social robot.

This paper introduces the Assistive Kitchen as a comprehensive demonstration and challenge scenario for technical cognitive systems. We describe its hardware and software infrastructure. Within the Assistive Kitchen application, we select particular domain activities as research subjects and identify the cognitive capabilities needed for perceiving, interpreting, analyzing, and executing these activities as research foci. We conclude by outlining open research issues that need to be solved to realize the scenarios successfully.

: Toward the origin of dialogue. En H. R. Schaffer (Ed.). Studies in motherinfant interaction. London: Academic Press. KAYE, K. (1979): Thickening thin data: the maternal role in developing communication and language. En M. Bullowa (Ed.). Before speech: The beginning of interpersonal communication. Cambridge: Cambridge University Press. KAYE, K. y CHARNEY, R. (1980): How mothers maintain "dialogues" with two-year-olds. En D. Olson (Ed.). The social foundation of language and thought. New York: W. W. Norton.

This study provides a non-exhaustive overview of AI methods, algorithms, and applications, particularly machine learning (ML) in rail transport. Drawing inspiration from the two European directives relating to the development (Directive 2012/34/EU) and interoperability (Directive (EU) 2016/797) of the railway system, this review proposes to classify AI applications according to (1) the ''structural elements'' of the railway system (infrastructure, rolling stock, energy, control and command, and signaling), (2) the ''functional elements'' (operation, maintenance, and telematics applications), and (3) both ''structural and functional'' elements with a view to identifying AI approaches aimed at improving railway safety, in particular the analysis of railway accidents and incidents based on investigation reports. Several ''classic'' AI techniques are implemented, including ML (supervised, semi-supervised, unsupervised), deep learning such as artificial neural networks (ANN), natural language processing (NLP), case-based reasoning (CBR), etc. However, the lack of interoperability between ''classic'' AI tools and the inadequacy of these approaches to capitalize, share, and reuse this knowledge have driven research toward the development of new approaches based on ontology's and knowledge graphs. A study of all AI applications shows that the stages of acquisition, retrieval, analysis, structuring, formalization, modeling, processing, and interpretation of the data produced pose a crucial problem in the rail transport sector. Moreover, with complex models called ''black boxes,'' it is difficult to understand how and why the internal reasoning mechanisms of the AI system impact the solution and predictions. The new approach of explainable AI (XAI) or explainable machine learning (XML) can provide an element of response to this problem, particularly when it comes to a crucial issue such as railway safety.

This study provides a non-exhaustive overview of AI methods, algorithms, and applications, particularly machine learning (ML) in rail transport. Drawing inspiration from the two European directives relating to the development (Directive 2012/34/EU) and interoperability (Directive (EU) 2016/797) of the railway system, this review proposes to classify AI applications according to (1) the ''structural elements'' of the railway system (infrastructure, rolling stock, energy, control and command, and signaling), (2) the ''functional elements'' (operation, maintenance, and telematics applications), and (3) both ''structural and functional'' elements with a view to identifying AI approaches aimed at improving railway safety, in particular the analysis of railway accidents and incidents based on investigation reports. Several ''classic'' AI techniques are implemented, including ML (supervised, semi-supervised, unsupervised), deep learning such as artificial neural networks (ANN), natural language processing (NLP), case-based reasoning (CBR), etc. However, the lack of interoperability between ''classic'' AI tools and the inadequacy of these approaches to capitalize, share, and reuse this knowledge have driven research toward the development of new approaches based on ontology's and knowledge graphs. A study of all AI applications shows that the stages of acquisition, retrieval, analysis, structuring, formalization, modeling, processing, and interpretation of the data produced pose a crucial problem in the rail transport sector. Moreover, with complex models called ''black boxes,'' it is difficult to understand how and why the internal reasoning mechanisms of the AI system impact the solution and predictions. The new approach of explainable AI (XAI) or explainable machine learning (XML) can provide an element of response to this problem, particularly when it comes to a crucial issue such as railway safety.

This study provides a non-exhaustive overview of AI methods, algorithms, and applications, particularly machine learning (ML) in rail transport. Drawing inspiration from the two European directives relating to the development (Directive 2012/34/EU) and interoperability (Directive (EU) 2016/797) of the railway system, this review proposes to classify AI applications according to (1) the ''structural elements'' of the railway system (infrastructure, rolling stock, energy, control and command, and signaling), (2) the ''functional elements'' (operation, maintenance, and telematics applications), and (3) both ''structural and functional'' elements with a view to identifying AI approaches aimed at improving railway safety, in particular the analysis of railway accidents and incidents based on investigation reports. Several ''classic'' AI techniques are implemented, including ML (supervised, semi-supervised, unsupervised), deep learning such as artificial neural networks (ANN), natural language processing (NLP), case-based reasoning (CBR), etc. However, the lack of interoperability between ''classic'' AI tools and the inadequacy of these approaches to capitalize, share, and reuse this knowledge have driven research toward the development of new approaches based on ontology's and knowledge graphs. A study of all AI applications shows that the stages of acquisition, retrieval, analysis, structuring, formalization, modeling, processing, and interpretation of the data produced pose a crucial problem in the rail transport sector. Moreover, with complex models called ''black boxes,'' it is difficult to understand how and why the internal reasoning mechanisms of the AI system impact the solution and predictions. The new approach of explainable AI (XAI) or explainable machine learning (XML) can provide an element of response to this problem, particularly when it comes to a crucial issue such as railway safety.

the physicality issues to be addressed when designing a domestic-aware space (Martinez and Greenhalgh) and presents a theoretical framework for incorporating technology into the design of public spaces drawing on computer science, architecture, design and philosophy (Malard and Cesar). Virtuality: Moving between Worlds. Another group of papers address issues that arise when we move from physical to virtual environments. Sharp addresses the challenges involved in translating physical artefacts used by co-located software development teams to facilitate in-situ communication practices into a virtual arrangement for supporting distributed teams. Whitham explores the augmented interaction capabilities when switching from a purely physical work environment to partly virtual environments and the ensuing difficulties with facilitating predictability. McKnight looks at the design opportunities when making the transition from the physical world to virtual reality and its impact on creativity. Representations. Representations and models are critical in design activity. Relatively formal models are used to capture the physical aspects of interaction in Mixed Interactive Systems (Dubois et al.) and to draw together existing theoretical work on physical interaction (Israel) in order to reflect and reason about systems including physical and digital components. In contrast, Hornecker considers different representational practices used within the design process itself, focusing on issues of materiality and physical interaction. Virtuality: The Real and Unreal. The final theme explores the boundaries of physicality. Gjerlufsen and Olsen present a deep philosophical exploration of the qualities of physicality and digitality, whilst Last raises philosophical questions on the physicality of nano-technologies and explores the possibilities of understanding physical processes that are outside our perceptual range. Any categorization is a simplification and during the workshop we will create alternative threads and themes using a physical variant of tagging, which has been so powerful in web-based social networking; borrowing ideas from the virtual and applying them in a physical form.

Recent incidents have shown that the production of take-off speeds is an activity vulnerable to miscalculations with a potential for disastrous outcomes. The aim of this paper is to analyze the calculation of the take-off speeds in a modern airline cockpit as a distributed cognitive activity in order to identify possible vulnerabilities in this process. We took the cockpit as the joint cognitive system under analysis and conducted an ethnographic study based on documental analysis, flight observations, interviews, and the analysis of 22 events involving failures related to the calculation of take-off speeds. The main argument is that the cognitive systems engineering perspective, with less focus on the human contribution than it is common in investigations, levels people and artifacts in the system as equal contributors to its eventual performance. Our analysis identified four assertions regarding vulnerabilities in the process of take-off speeds calculation: (1) representations at the level of the cockpit are always partial and incomplete; (2) some interactions require interpretation rather than institution; (3) interactions of agents do not follow a canonical process of coordination; (4) the control of the prevention of failures is accurate but inadequate. These vulnerabilities are a matter of interactions among cognitive systems in the cockpit, rather than vulnerabilities of individual agents, such as humans or artifacts.

Volume 79 of “Neutrosophic Sets and Systems” presents a collection of advanced studies that apply neutrosophic theory to a wide range of fields in information science and engineering. The papers in this volume focus on utilizing neutrosophic concepts to address problems involving uncertainty, imprecision, and indeterminacy. Key research areas explored include the application of neutrosophic models in machine learning for risk prediction, such as for chronic kidney disease, and for image classification tasks like oral disease classification. The volume also features studies on multi-criteria and multi-attribute decision-making (MCDM) across various domains, including assessing the quality of university dance instruction, evaluating the efficiency of resource allocation in university sports management, and analyzing performance in construction project management. Furthermore, the journal includes theoretical contributions, such as the development of new mathematical frameworks like determinant theory of neutrosophic fuzzy matrices and harmonic aggregation operators. Other applications discussed involve web security, solving complex equations in neutrosophic environments, and the evaluation of urban road and bridge projects based on environmental sustainability. This volume demonstrates the versatility and growing application of neutrosophic theory in both foundational and practical contexts.

This volume of “Neutrosophic Sets and Systems” presents a wide array of research focusing on the theory and application of neutrosophic principles in diverse scientific fields. The articles explore new theoretical frameworks, including neutrogeometry and cubic spherical neutrosophic sets, which provide enhanced models for capturing and representing uncertainty. The practical applications highlighted in this volume are equally varied, addressing complex problems such as face detection and recognition in forensic analysis, the extraction of beneficial rules for heart disease diagnosis, and the evaluation of leadership management in healthcare systems. The research also includes an analysis of the scientific production on neutrosophy from a Latin American perspective, demonstrating the growing global impact of the discipline. This collection serves as a testament to the versatility of neutrosophic methods in providing effective solutions across a range of subjects, from computer science and medicine to operations research and social sciences.

We describe the case of Angie, a 50 year-old woman with profound amnesia (General Memory Index = 49, Full Scale IQ = 126) following a closed head injury in 1985. This case is unique in comparison to other cases reported in the literature in that, despite the severity of her amnesia, she has developed remarkable real-world life abilities, shows impressive self awareness and insight into the impairment and sparing of various functional memory abilities, and exhibits ongoing maturation of her identity and sense of self following amnesia. The case provides insights into the interaction of different memory and cognitive systems in handling real-world memory demands, and has implications for rehabilitation and for successful life outcome after amnesia.

Occluded facial expression recognition (OFER) poses a formidable challenge in real-world applications, particularly in human-computer interaction and affective computing. Despite recent advancements, existing methodologies often struggle to maintain optimal accuracy under occlusion constraints. This study proposes a novel hybrid framework that synergizes handcrafted and deep learning-based features to enhance robustness and precision in emotion recognition. Specifically, we integrate Histogram of Oriented Gradients (HoG), facial landmark descriptors, and sliding window-based HoG representations with deep convolutional neural network (CNN) features, leveraging their complementary strengths. Our experimental design explores multiple feature fusion strategies, including CNN-based automated classification and a hybrid model incorporating Dlibextracted landmarks with HoG-CNN integration. Comparative analysis against state-of-the-art approaches demonstrates that our multi-feature fusion technique significantly improves recognition accuracy, achieving a remarkable 96% accuracy on benchmark datasets such as RAF-DB and AffectNet. However, we observe a marginal decline in performance with increased dataset complexity, emphasizing the need for scalable solutions. This research underscores the efficacy of integrating handcrafted and deep learning-driven features, offering a promising direction for advancing occlusion-robust facial expression recognition in dynamic environments.

Through an examination of the emerging domain of cognitive systems, with a focus on attention-centric cognitive systems used for notification, this document explores the human-computer interaction challenges that must be addressed for successful interface design. This document asserts that with compatible tools and methods, user notification requirements and interface usability can be abstracted, expressed, and compared with critical parameter ratings; that is, even novice designers can assess attention cost factors to determine target parameter levels for new system development. With a general understanding of the user tasks supported by the notification system, a designer can access the repository of design knowledge for appropriate information and interaction design techniques (e.g., use of color, audio features, animation, screen size, transition of states, etc), which have analytically and empirically derived ratings. Furthermore, usability evaluation methods, provided to designers as part of the integrated system, are adaptable to specific combinations of targeted parameter levels. User testing results can be conveniently added back into the design knowledge repository and compared to target parameter levels to determine design success and build reusable HCI knowledge. This approach is discussed in greater detail as we describe five HCI challenges relating to cognitive system development: (1) convenient access to basic research and guidelines, (2) requirements engineering methods for notification interfaces, (3) better and more usable predictive modeling for pre-attentive and dual-task interfaces, (4) standard empirical evaluation procedures for notification systems, and (5) conceptual frameworks for organizing reusable design and software components. This document also describes our initial work toward building infrastructure to overcome these five challenges, focused on notification system development. We described LINK-UP, a design environment grounded on years of theory and method development within HCI, providing a mechanism to integrate interdisciplinary expertise from the cognitive systems research community. Claims allow convenient access to basic research and guidelines, while modules parallel a lifecycle development iteration and provide a process for requirements engineering guided by this basic research. The activities carried out through LINK-UP provide access to and interaction with reusable design components organized based on our framework. We think that this approach may provide the scientific basis necessary for exciting interdisciplinary advancement through many fields of design, with notification systems serving as an initial model. A version of this document will appear as chapter 3 in the book Cognitive Systems: Human Cognitive Models in Systems Design edited by Chris Forsythe, Michael Bernard, and Timothy Goldsmith resulting from a workshop led by the editors in summer 2003. The authors are grateful for the input of the workshop organizers and conference attendees in the preparation of this document.

Learning and generating serially ordered sequences of actions is a core component of cognition both in organisms and in artificial cognitive systems. When these systems are embodied and situated in partially unknown environments, specific constraints arise for any neural mechanism of sequence generation. In particular, sequential action must resist fluctuating sensory information and be capable of generating sequences in which the individual actions may vary unpredictably in duration. We provide a solution to this problem within the framework of Dynamic Field Theory by proposing an architecture in which dynamic neural networks create stable states at each stage of a sequence. These neural attractors are destabilized in a cascade of bifurcations triggered by a neural representation of a condition of satisfaction for each action. We implement the architecture on a robotic vehicle in a color search task, demonstrating both sequence learning and sequence generation on the basis of low-level sensory information.

Spatial communication is incredibly flexible, but implemented spatial language models to date have focused on a limited set of behaviors. The present work takes a step towards addressing this limit by implementing a new spatial language model grounded in the principles of neural population dynamics. To rigorously test our model, we implement it on a robotic platform continuously linked to real-world visual images. Our model extracts the categorical, cognitive information from the low-level sensory input through the system dynamics, thus permitting the dynamic integration of visual space, spatial language, and color within a single, unified theoretical framework. In a series of demonstrations then we show how this model autonomously generates a range of spatial language behaviors such as generating a spatial description, identifying the color of an object at a described location, and dynamically structuring camera movements according to internal decision dynamics.

In the general context of Group Decision Support System (GDSS), the paper investigates the possibility to externalize and support from a metacognitive perspective the effective use of facilitation knowledge with self-development capabilities. The experimental results make evident that these capabilities may be easily engineered by adopting the basic principles of the design for emergence in constructing an e-meeting facilitation tool that act as a stigmergic collaborative environment for the participants. Basically, the GDSS needs to provide a minimal structure for modeling the group decision process (GDP) which enables a participant-driven approach to group facilitation and magnify the sense of social participation. In this way the GDSS may provide a collaborative environment where unpredictable and more effective models of GDP design will emerge through the exploration of the problem space.

We consider the design space available for social robots in terms of a hierarchy of functional definitions: the essential properties in terms of a locus of interaction, autonomy, intelligence, awareness of humans as possessors of mental state, and awareness of humans as social interactors. We also suggest that the emphasis on physical embodiment in some segments of the social robotics community has obscured commonalities with a class of agents that are identical in all other respects. These definitions naturally suggest research issues, directions, and possibilities which we explore. Social robotics also lacks compelling 'killer apps' which we suggest would help focus the community on a research agenda. "A machine capable of carrying out a complex series of actions automatically, especially one programmable by a computer" "A machine that resembles a human and does mechanical, routine tasks on command" "Any machine or mechanical device that operates automatically with humanlike skill" "A machine resembling a human being and able to replicate certain human movements and functions automatically" "Devices that possesses some of these capabilities: accept electronic programming, process data or physical perceptions electronically, operate autonomously to some degree, move around, operate physical parts of itself or physical processes, sense and manipulate their environment, and exhibit intelligent behaviorespecially behavior which mimics humans or other animals" As these online dictionary definitions of 'robot' show, there is considerable confusion about what exactly the term 'robot' is supposed to mean. They attempt to blend two quite disparate underlying concepts: a device that perform physical actions, while at the same time resembling a human. Definitions in the more-formal literature are not much clearer: "A reprogrammable, multifunctional manipulator designed to move material, parts, tools, or specialized devices through various programmed motions for the performance of a variety of tasks" credited to the Robot Institute of America, 1979.

According to Thrun and others, personal service robots need increasingly more autonomy in order to function in the highly unpredictable company of humans. At the same time, the cognitive processes in artifacts will become increasingly alien to us. This has several reasons: 1. Maturana's concept of structural determinism questions conventional forms of interaction. 2. Considerably different ways of embodiment result in incompatible referential frameworks (worldviews). 3. Engineers focus on the output of artifacts, whereas autonomous cognitive systems seek to control their input state. As a result, instructional interaction -the basic ingredient of conventional man-machine relationships -with genuine autonomous systems will become impossible. Therefore the increase of autonomy will eventually lead to a paradox. Today we are still in a position to anthropomorphically trivialize the behavioral pattern of current robots (von Foerster). Eventually, however, when self-organizing systems will have reached the high levels of autonomy we wished for interacting with them may become impossible since their goals will be completely independent of ours.

For cognitive systems, embodiment appears to be of crucial importance. Unfortunately, nobody seems to be able to define embodiment in a way that would prevent it from also covering its trivial interpretations such as mere situatedness in complex environments. The paper focuses on the definition of embodiment, especially whether physical embodiment is necessary and / or sufficient for cognitive systems. Cognition is characterized as a continuous complex process rather than ahistorical logical capability. Furthermore, the paper investigates the relationship between cognitive embodiment and the issues of understanding, representation and task specification.

It seems characteristic for humans to detect structural patterns in the world to anticipate future states. Therefore, scientific and common sense cognition could be described as information processing which infers rule-like laws from patterns in data-sets. Since information processing is the domain of computers, artificial cognitive systems are generally designed as pattern discoverers. This paper questions the validity of the information processing paradigm as an explanation for human cognition and a design principle for artificial cognitive systems. Firstly, it is known from the literature that people suffer from conditions such as information overload, superstition, and mental disorders. Secondly, cognitive limitations such as a small short-term memory, the seteffect, the illusion of explanatory depth, etc. raise doubts as to whether human information processing is able to cope with the enormous complexity of an infinitely rich (amorphous) world. It is suggested that, under normal conditions, humans construct information rather than process it. The constructed information contains anticipations which need to be met. This can be hardly called information processing, since patterns from the "outside" are not used to produce action but rather to either justify anticipations or restructure the cognitive apparatus. When it fails, cognition switches to pattern processing, which, given the amorphous nature of the experiential world, is a lost cause if these patterns and inferred rules do not lead to a (partial) reorganisation of internal structures such that constructed anticipations can be met again. In this scenario, superstition and mental disorders are the result of a profound and/or random restructuring of already existing cognitive components (e.g., action sequences). This means that whenever a genuinely cognitive system is exposed to pattern processing it may start to behave superstitiously. The closer we get to autonomous self-motivated artificial cognitive systems, the bigger the danger becomes of superstitious information processing machines that "blow up" rather than behave usefully and effectively. Therefore, to avoid superstition in cognitive systems they should be designed as information constructing entities.

Occluded facial expression recognition (OFER) poses a formidable challenge in real-world applications, particularly in human-computer interaction and affective computing. Despite recent advancements, existing methodologies often struggle to maintain optimal accuracy under occlusion constraints. This study proposes a novel hybrid framework that synergizes handcrafted and deep learning-based features to enhance robustness and precision in emotion recognition. Specifically, we integrate Histogram of Oriented Gradients (HoG), facial landmark descriptors, and sliding window-based HoG representations with deep convolutional neural network (CNN) features, leveraging their complementary strengths. Our experimental design explores multiple feature fusion strategies, including CNN-based automated classification and a hybrid model incorporating Dlibextracted landmarks with HoG-CNN integration. Comparative analysis against state-of-the-art approaches demonstrates that our multi-feature fusion technique significantly improves recognition accuracy, achieving a remarkable 96% accuracy on benchmark datasets such as RAF-DB and AffectNet. However, we observe a marginal decline in performance with increased dataset complexity, emphasizing the need for scalable solutions. This research underscores the efficacy of integrating handcrafted and deep learning-driven features, offering a promising direction for advancing occlusion-robust facial expression recognition in dynamic environments.

In this paper we present some empirical data in support of a new view of attentional orienting, according to which (a) some basic mechanisms of attentional deployment (i.e. its high efficiency in dealing with expected and unexpected inputs) meet the requirements of the inferential system, and have possibly evolved to support its functioning; and (b) these orienting mechanisms function in very similar ways in different domains, i.e. in perceptual tasks and in symbolic tasks. After sketching the basics of the model and of its specific predictions, we report the results of 3 behavioral experiments where participants tracked visual trajectories (Experiments 1 and 3) or arithmetic series (Experiments 2 and 3), responding to the onset of a target event (e.g., to a specific number) and to the repetition of an event (e.g., a number appearing twice consecutively). Target events could be anticipated when they were embedded in regular series/trajectories; they could be anticipated, with the anticipation later disconfirmed, when a regular series/trajectory was abruptly interrupted before the target event occurred; and could not be anticipated when the series/trajectory was random. Repeated events could not be anticipated. Results show a very similar pattern of allocation in tracking visual trajectories and arithmetic series: attention is focused on anticipated events, it is defocused and redistributed when an anticipation is not confirmed by ensuing events, while performance decreases when dealing with random series/trajectory, i.e. in absence of anticipations. The behavioral results were later supported by an electrophysiological investigation using Event Related Potentials. In our view, this pattern of allocation of attention is due to the fact that confirmed and disconfirmed anticipations are crucial events for knowledge revision, i.e. the fine tuning of the inferential system to the environment; attentional mechanisms have developed so as to enhance detection of these events, possibly at all levels of inferential processing.

Studies looking at the connections between working memory and children's mathematical achievement are often limited by the general nature of the measures of either working memory or mathematics. In this study, thirty-five children in Year 5 (9 to 10 years of age) were given ...

Scientific thinking refers to the mental processes used when reasoning about the content of science (e.g., force in physics), engaged in typical scientific activities (e.g., designing experiments), or specific types of reasoning that are frequently used in sci-ence (e.g., deducing that there is a planet beyond Pluto). Scientific thinking involves many general-purpose cognitive operations that human beings apply in nonscientific do-mains such as induction, deduction, anal-ogy, problem solving, and causal reason-ing. These cognitive processes are covered in many chapters of this handbook (see Sloman & Lagnado, Chap. 5 on induction;

In this paper a computational model is presented for how a desire triggers responses and feelings. The model shows how these feelings can be biased, for example due to addicting experiences in the past. Both the strength of a response and of the associated feeling result from a converging dynamic pattern modelled by reciprocal causal interactions between the two. The model has been used to conduct a number of simulation experiments under varying circumstances. Moreover, it has been evaluated by formal analysis of emerging patterns entailed by the model. Furthermore, it has been pointed out how the computational model can be applied within an ambient agent system supporting a human in not being tempted. In a simple example scenario it is shown such an ambient agent system is able to predict and assess a human's desire state, and use this assessment to suggest alternatives to avoid falling for certain temptations.

In recent years, several authors have investigated how co-occurrence statistics in natural language can act as a cue that children may use to extract syntactic categories for the language they are learning. While some authors have reported encouraging results, it is difficult to evaluate the quality of the syntactic categories derived. It is argued in this paper that traditional measures of accuracy are inherently flawed. A valid evaluation metric needs to consider the well-formedness of utterances generated through a production end. This paper attempts to evaluate the quality of the categories derived from co-occurrence statistics through the use of MOSAIC, a computational model of syntax acquisition that has already been used to simulate several phenomena in child language. It is shown that derived syntactic categories that may appear to be of high quality quickly give rise to errors that are not typical of child speech. A solution to this problem is suggested in the form of a chunking mechanism that serves to differentiate between alternative grammatical functions of identical word forms. Results are evaluated in terms of the error rates in utterances produced by the system as well as the quantitative fit to the phenomenon of subject omission.

In recent years, several authors have investigated how co-occurrence statistics in natural language can act as a cue that children may use to extract syntactic categories for the language they are learning. While some authors have reported encouraging results, it is difficult to evaluate the quality of the syntactic categories derived. It is argued in this paper that traditional measures of accuracy are inherently flawed. A valid evaluation metric needs to consider the well-formedness of utterances generated through a production end. This paper attempts to evaluate the quality of the categories derived from cooccurrence statistics through the use of MOSAIC, a computational model of syntax acquisition that has already been used to simulate several phenomena in child language. It will be shown that derived syntactic categories which may appear to be of high quality will quickly give rise to errors which are not typical of child speech. A solution to this problem is suggested in the form of a chunking mechanism which serves to differentiate between alternative grammatical functions of identical word forms. Results are evaluated in terms of the error rates in utterances produced by the system as well as the quantitative fit to the phenomenon of subject omission.

In recent years, several authors have investigated how co-occurrence statistics in natural language can act as a cue that children may use to extract syntactic categories for the language they are learning. While some authors have reported encouraging results, it is difficult to evaluate the quality of the syntactic categories derived. It is argued in this paper that traditional measures of accuracy are inherently flawed. A valid evaluation metric needs to consider the wellformedness of utterances generated through a production end. This paper attempts to evaluate the quality of the categories derived from co-occurrence statistics through the use of MOSAIC, a computational model of syntax acquisition that has already been used to simulate several phenomena in child language. It is shown that derived syntactic categories that may appear to be of high quality quickly give rise to errors that are not typical of child speech. A solution to this problem is suggested in the form of a chunking mechanism that serves to differentiate between alternative grammatical functions of identical word forms. Results are evaluated in terms of the error rates in utterances produced by the system as well as the quantitative fit to the phenomenon of subject omission.

A basic claim of this paper is that the foundational theoretical problem of the social sciences -the possibility of unconscious, unplanned forms of cooperation and intelligence among intentional agents (the very hard issue of the 'invisible hand', of the 'spontaneous social order' but also of 'social functions') -will eventually be clarified thanks to the contribution of AI (and, in particular, of cognitive Agent modelling, learning, and MAS) and its entering the social simulation domain. After introducing Multi-Agent-Based Social Simulation and its trends, the limits of the very popular notion of 'emergence' are discussed, Smith's and Hayek's view of 'spontaneous social order' are critically introduced, and serious contradictions in the theory of 'social functions' among intentional agents are pointed out. The problem is how to reconcile the 'external' teleology that orients the agent's behaviour with the 'internal' teleology governing it. In order to account for the functional character of intentional action, we need a somewhat sophisticated model of intention, and a different view of layered cognitive architectures combining explicit beliefs and goals with association and conditioning. On such a basis we sketch a model of unknown functions impinging on intentional actions through a high level form of (MA) reinforcement learning. This model accounts for both eu-functions and dys-functions, autonomous and heteronomous functions. It is argued that, in order to reproduce some behaviour, its effects should not necessarily be 'good', i.e. useful for the goal of the agent or of some higher macro-system.

Modern civilization is approaching a critical inflection point, facing interwoven crises such as climate disruption, infrastructure fragility, mental health decline, and social fragmentation. Existing smart infrastructure paradigms remain siloed and reactive, prioritizing technical efficiency over holistic well-being. In response, this paper proposes NEURA-Neural Earth Unified Responsive Awareness-a novel civilizational framework that reconceptualizes infrastructure as sentient, ethically governed, and emotionally intelligent. NEURA integrates environmental sensing, behavioral feedback, predictive AI, and decentralized governance to enable urban and rural regions to adapt in real time to ecological, emotional, and social cues. Unlike centralized smart grids, NEURA functions as a distributed mesh of culturally attuned, community-owned "cognitive nodes." The system embeds emotional mapping, participatory data ethics, and self-healing materials to support both planetary resilience and civic vitality. This paper outlines the NEURA architecture, introduces its theoretical grounding, details its three-phase implementation strategy, and explores its transformative potential across economic, environmental, and social domains. Ethical risks and design limitations are examined, alongside proposed mitigation strategies. NEURA offers a new paradigm for infrastructure-one that does not merely manage systems, but fosters empathy, adaptability, and planetary care.

Sentiment analysis on social media such as Twitter has become a very important and challenging task. Due to the characteristics of such data-tweet length, spelling errors, abbreviations, and special characters-the sentiment analysis task in such an environment requires a non-traditional approach. Moreover, social media sentiment analysis is a fundamental problem with many interesting applications. Most current social media sentiment classification methods judge the sentiment polarity primarily according to textual content and neglect other information on these platforms. In this paper, we propose a neural network model that also incorporates user behavioral information within a given document (tweet). The neural network used in this paper is a Convolutional Neural Network (CNN). The system is evaluated on two datasets provided by the SemEval-2016 Workshop. The proposed model outperforms current baseline models (including Naive Bayes and Support Vector Machines), which shows that going beyond the content of a document (tweet) is beneficial in sentiment classification, because it provides the classifier with a deep understanding of the task.

In Cognitive Science, conceptual blending has been proposed as an important cognitive mechanism that facilitates the creation of new concepts and ideas by constrained combination of available knowledge. It thereby provides a possible theoretical foundation for modeling high-level cognitive faculties such as the ability to understand, learn, and create new concepts and theories. Quite often the development of new mathematical theories and results is based on the combination of previously independent concepts, potentially even originating from distinct subareas of mathematics. Conceptual blending promises to offer a framework for modeling and re-creating this form of mathematical concept invention with computational means. This paper describes a logic-based framework which allows a formal treatment of theory blending (a subform of the general notion of conceptual blending with high relevance for applications in mathematics), discusses algorithmic aspects of blending within the framework, and provides an illustrating worked out example from mathematics.

We present an account of a process by which different conceptualisations of number can be blended together to form new conceptualisations via recognition of common features, and judicious combination of their distinctive features. The accounts of number are based on Lakoff and Núñez's cognitively based grounding metaphors for arithmetic. The approach incorporates elements of analogical inference into a generalised framework of conceptual blending, using some ideas from the work of Goguen. The ideas are worked out using Heuristic-Driven Theory Projection (HDTP, a method based on higher-order anti-unification). HDTP provides generalisations between domains, giving a crucial step in the process of finding commonalities between theories. In addition to generalisations, HDTP can also transfer concepts from one domain to another, allowing the construction of new conceptual blends. Alongside the methods by which conceptual blends may be constructed, we provide heuristics to guide this process.

In the past twenty or so years the scientific community has made impressive advancements in the modeling and simulation of general human cognition. This progress has led to the beginnings of wide-spread applications and use. In fact, we are now at a point where the community can begin to make fairly accurate predictions as to how this technology will be used in the next twenty-plus years. Accordingly, the purpose of this panel is to engage the community at large regarding the future needs and requirements associated with building cognitive models for various scientific and engineering endeavors. Specifically, this panel will discuss and make recommendations with regard to the future functionality of cognitive modeling that could be encompassed in nextgeneration capabilities. To do this, we will concentrate on four different domain areas. These are: academic use of cognitive modeling, cognitive model development, neuroscience-related issues, and practical applications of cognitive modeling.

A. Basic and Translational Sleep Science X. Sleep Deprivation night two and 3h on mornings three and four. Cognitive testing was performed every 3h during scheduled wakefulness. Executive function and cognitive processing speed were tested with a Stroop Color Word task and selective visual attention was tested with a conjunction visual search task. Stability of individual differences in cognitive function were quantified by intra-class correlation coefficients (ICC) derived from linear mixed effects models with subject as a random factor and visit and battery number as fixed factors. Results: During combined sleep restriction and circadian misalignment, cognitive processing speed (median reaction times for correct responses to congruent, neutral, and incongruent stimuli), executive function (inhibitory control), and visual selective attention (cognitive throughput) were significantly impaired (p<.05). Performance was highly consistent within individuals at baseline and during combined sleep restriction and circadian misalignment, with ICCs of 0.86 to 0.93 indicating "almost perfect" trait-like stability. Conclusion: Combined sleep restriction and circadian misalignment produce robust, trait-like effects on higher-order cognitive functions. Consistent vulnerability to sleep and circadian disturbance has implications for development of individualized and targeted countermeasures for workers performing safety-critical tasks under conditions of insufficient sleep and circadian misalignment, e.g., the military, health care workers and emergency responders.

This paper attempts to explore the behavioral characteristics of human beings from a point of view that is not based on inter-human comparisons, but on an examination of the intrinsic physical, biological, intellectual, emotional, and social characteristics of this species. It is argued that the experience-based nature of the human cognitive system greatly reduces our ability to adapt to changes in our environment, exploit opportunities, and create new knowledge. A fundamental biological survival instinct drives the human being to seek a level of certainty and security that is irreconcilable with a continuously changing and largely unpredictable environment. The symptoms of this distinctly human struggle manifest themselves in a strong resistance to change and an emotional aversion to experimentation and risk taking. The author postulates that the merging of biology and technology will allow the human species to accelerate progress along its evolutionary path in preparation for a series of emerging environmental challenges. These challenges are most likely to be related to rapid world population growth in the context of finite natural resources. It is argued that the merging of biology with technology may be a prerequisite for providing human beings with the necessary intellectual capabilities to both mitigate the abuse and maximize the use of planet Earth. This paper focuses on the need for technology to supplement and enhance the intellectual capabilities of human beings, and not on the technological prospects themselves. The latter can be found in the literature (

There is a growing interest in neuroscience for how individual-specific structural and functional features of the cortex relate to cognitive traits. This work builds on previous research which, by using classical high-dimensional approaches, has proven that the interindividual variability of functional connectivity (FC) profiles reflects differences in fluid intelligence. To provide an additional perspective into this relationship, the present study uses a recent framework for investigating cortical organization: functional gradients. This approach places local connectivity profiles within a common low-dimensional space whose axes are functionally interpretable dimensions. Specifically, this study uses a data-driven approach to model the association between FC variability and interindividual differences in intelligence. For one of these loci, in the right ventral-lateral prefrontal cortex (vlPFC), we describe an association between fluid intelligence and the relative functional distance of this area from sensory and high-cognition systems. Furthermore, the topological properties of this region indicate that, with decreasing functional affinity with high-cognition systems, vlPFC functional connections are more evenly distributed across all networks. Participating in multiple functional networks may reflect a better ability to coordinate sensory and high-order cognitive systems.