In this article, network science is discussed from a methodological perspective, and two central theses are defended. The first is that network science exploits the very properties that make a system complex. Rather than using idealization techniques to strip those properties away, as is standard practice in other areas of science, network science brings them to the fore, and uses them to furnish new forms of explanation. The second thesis is that network representations are particularly helpful (...) in explaining the properties of non-decomposable systems. Where part-whole decomposition is not possible, network science provides a much-needed alternative method of compressing information about the behavior of complex systems, and does so without succumbing to problems associated with combinatorial explosion. The article concludes with a comparison between the uses of network representation analyzed in the main discussion, and an entirely distinct use of network representation that has recently been discussed in connection with mechanistic modeling. (shrink)

In this paper I discuss one of the key issuesin the philosophy of neuroscience:neurosemantics. The project of neurosemanticsinvolves explaining what it means for states ofneurons and neural systems to haverepresentational contents. Neurosemantics thusinvolves issues of common concern between thephilosophy of neuroscience and philosophy ofmind. I discuss a problem that arises foraccounts of representational content that Icall ``the economy problem'': the problem ofshowing that a candidate theory of mentalrepresentation can bear the work requiredwithin in the causal economy of a mind and (...) anorganism. My approach in the current paper isto explore this and other key themes inneurosemantics through the use of computermodels of neural networks embodied and evolvedin virtual organisms. The models allow for thelaying bare of the causal economies of entireyet simple artificial organisms so that therelations between the neural bases of, forinstance, representation in perception andmemory can be regarded in the context of anentire organism. On the basis of thesesimulations, I argue for an account ofneurosemantics adequate for the solution of theeconomy problem. (shrink)

Despite its centrality in the philosophy of cognitive science, there has been little prior philosophical work engaging with the notion of representation in contemporary NLP practice. This paper attempts to fill that lacuna: drawing on ideas from cognitive science, I introduce a framework for evaluating the representational claims made about components of neural NLP models, proposing three criteria with which to evaluate whether a component of a model represents a property and operationalising these criteria using probing classifiers, a popular (...) analysis technique in NLP (and deep learning more broadly). The project of operationalising a philosophically-informed notion of representation should be of interest to both philosophers of science and NLP practitioners. It affords philosophers a novel testing-ground for claims about the nature of representation, and helps NLPers organise the large literature on probing experiments, suggesting novel avenues for empirical research. (shrink)

Many philosophers and psychologists have attempted to elucidate the nature of mental representation by appealing to notions like isomorphism or abstract structural resemblance. The ‘structural representations’ that these theorists champion are said to count as representations by virtue of functioning as internal models of distal systems. In his 2007 book, Representation Reconsidered, William Ramsey endorses the structural conception of mental representation, but uses it to develop a novel argument against representationalism, the widespread view that cognition essentially involves (...) the manipulation of mental representations. Ramsey argues that although theories within the ‘classical’ tradition of cognitive science once posited structural representations, these theories are being superseded by newer theories, within the tradition of connectionism and cognitive neuroscience, which rarely if ever appeal to structural representations. Instead, these theories seem to be explaining cognition by invoking so-called ‘receptor representations’, which, Ramsey claims, aren’t genuine representations at all—despite being called representations, these mechanisms function more as triggers or causal relays than as genuine stand-ins for distal systems. I argue that when the notions of structural and receptor representation are properly explicated, there turns out to be no distinction between them. There only appears to be a distinction between receptor and structural representations because the latter are tacitly conflated with the ‘mental models’ ostensibly involved in offline cognitive processes such as episodic memory and mental imagery. While structural representations might count as genuine representations, they aren’t distinctively mental representations, for they can be found in all sorts of non-intentional systems such as plants. Thus to explain the kinds of offline cognitive capacities that have motivated talk of mental models, we must develop richer conceptions of mental representation than those provided by the notions of structural and receptor representation. (shrink)

A network of gene regulation organized in a hierarchical and combinatorial manner is crucially involved in the development of the neural network, and has to be considered one of the main substrates of genetic change in its evolution. Though qualitative features may emerge by way of the accumulation of rather unspecific quantitative changes, it is reasonable to assume that at least in some cases specific combinations of regulatory parts of the genome initiated new directions of evolution, leading to (...) novel capabilities of the brain. These notions are applied, in this paper, to the evolution of the capability of cognition-based human empa­thy. It is suggested that it has evolved as a secondary effect of the evolution of strategic thought. Development of strategies depends on abstract representations of one’s own pos­sible future states in one’s own brain to allow assessment of their emotional desirability, but also on the representation and emotional evaluation of possible states of others, allowing anticipation of their behaviour. This is best achieved if representations of others are con­nected to one’s own emotional centres in a manner similar to self-representations. For this reason, the evolution of the human brain is assumed to have established representations with such linkages. No group selection is involved, because the quality of strategic thought affects the fitness of the individual. A secondary effect of this linkage is that both the actual states and the future perspectives of others elicit vicarious emotions, which may contribute to the motivations of altruistic behaviour. (shrink)

In this paper, we explore the conceptual problems arising when using network analysis in person- centered care (PCC) in psychiatry. Personalized network models are potentially helpful tools for PCC, but we argue that using them in psychiatric practice raises boundary problems, i.e., problems in demarcating what should and should not be included in the model, which may limit their ability to provide clinically-relevant knowledge. Models can have explanatory and representational boundaries, among others. We argue that we can make (...) more explicit what kind of questions personalized network models can address in PCC, given their representational and explanatory boundaries, using perspectival reasoning. (shrink)

Shared conceptualization, in the sense we take it here, is as recent a notion as the Semantic Web, but its relevance for a large variety of fields requires efficient methods of extraction and representation for both quantitative and qualitative data. This notion is particularly relevant for the investigation into, and construction of, semantic structures such as knowledge bases and taxonomies, but given the required large, often inaccurate, corpora available for search we can get only approximations. We see fuzzy description (...) logic as an adequate medium for the representation of human semantic knowledge and propose a means to couple it with fuzzy semantic networks via the propositional Łukasiewicz fuzzy logic such that these suffice for decidability for queries over a semantic-knowledge base such as “to what degree of sharedness does it entail the instantiation C(a) for some concept C” or “what are the roles R that connect the individuals a and b to degree of sharedness ε.” . (shrink)

Terry Horgan and John Tienson have suggested that connectionism might provide a framework within which to articulate a theory of cognition according to which there are mental representations without rules (RWR) (Horgan and Tienson 1988, 1989, 1991, 1992). In essence, RWR states that cognition involves representations in a language of thought, but that these representations are not manipulated by the sort of rules that have traditionally been posited. In the development of RWR, Horgan and Tienson attempt to forestall a particular (...) line of criticism, theSyntactic Argument, which would show RWR to be inconsistent with connectionism. In essence, the argument claims that the node-level rules of connectionist networks, along with the semantic interpretations assigned to patterns of activation, serve to determine a set of representation-level rules incompatible with the RWR conception of cognition. The present paper argues that the Syntactic Argument can be made to show that RWR is inconsistent with connectionism. (shrink)

The paper is dedicated to particular cases of interaction and mutual impact of philosophy and cognitive science. Thus, philosophical preconditions in the middle of the 20th century shaped the newly born cognitive science as mainly based on conceptual and propositional representations and syntactical inference. Further developments towards neural networks and statistical representations did not change the prejudice much: many still believe that network models must be complemented with some extra tools that would account for proper human cognitive traits. I (...) address some real implemented connectionist models that show how ‘new associationism ’ of the neural network approach may not only surpass Humean limitations, but, as well, realistically explain abstraction, inference and prediction. Then I stay on Predictive Processing theories in a little more detail to demonstrate that sophisticated statistical tools applied to a biologically realist ontology may not only provide solutions to scientific problems or integrate different cognitive paradigms but propose some philosophical insights either. To conclude, I touch on a certain parallelism of Predictive Processing and philosophical inferentialism as presented by Robert Brandom. (shrink)

We describe and try to motivate our project to build systems using both a knowledge based and a neural network approach. These two approaches are used at different stages in the solution of a problem, instead of using knowledge bases exclusively on some problems, and neural nets exclusively on others. The knowledge base (KB) is defined first in a declarative, symbolic language that is easy to use. It is then compiled into an efficient neural network (NN) representation, (...) run, and the results from run time and (eventually) from learning are decompiled to a symbolic description of the knowledge contained in the network. After inspecting this recovered knowledge, a designer would be able to modify the KB and go through the whole cycle of compiling, running, and decompiling again. The central question with which this project is concerned is, therefore, How do we go from a KB to an NN, and back again? We are investigating this question by building tools consisting of a repertoire of language/translation/network types, and trying them on problems in a variety of domains. (shrink)

This essay presents a philosophical and computational theory of the representation of de re, de dicto, nested, and quasi-indexical belief reports expressed in natural language. The propositional Semantic Network Processing System (SNePS) is used for representing and reasoning about these reports. In particular, quasi-indicators (indexical expressions occurring in intentional contexts and representing uses of indicators by another speaker) pose problems for natural-language representation and reasoning systems, because--unlike pure indicators--they cannot be replaced by coreferential NPs without changing the (...) meaning of the embedding sentence. Therefore, the referent of the quasi-indicator must be represented in such a way that no invalid coreferential claims are entailed. The importance of quasi-indicators is discussed, and it is shown that all four of the above categories of belief reports can be handled by a single representational technique using belief spaces containing intensional entities. Inference rules and belief-revision techniques for the system are also examined. (shrink)

We introduce the Bias Network Approach (BNA) as a sociotechnical method for AI developers to identify, map, and relate biases across the AI development process. This approach addresses the limitations of what we call the "isolationist approach to AI bias," a trend in AI literature where biases are seen as separate occurrence linked to specific stages in an AI pipeline. Dealing with these multiple biases can trigger a sense of excessive overload in managing each potential bias individually or promote (...) the adoption of an uncritical approach to understanding the influence of biases in developers’ decision-making. The BNA fosters dialogue and a critical stance among developers, guided by external experts, using graphical representations to depict biased connections. To test the BNA, we conducted a pilot case study on the "waiting list” project, involving a small AI developer team creating a healthcare waiting list NPL model in Chile. The analysis showed promising findings: (i) the BNA aids in visualizing interconnected biases and their impacts, facilitating ethical reflection in a more accessible way; (ii) it promotes transparency in decision-making throughout AI development; and (iii) more focus is necessary on professional biases and material limitations as sources of bias in AI development. (shrink)

This research investigates the representations of gendered social actors within the subversionary discourse of equal educational opportunities for males and females in Iranian English as a Foreign Language (EFL) books. Using critical discourse analysis (CDA) as the theoretical framework, the authors blend van Leeuwen’s (Texts and practices: Readings in critical discourse analysis, Routledge, London, 2003) ‘Social Actor Network Model’ and Sunderland’s (Gendered discourses, Palgrave Macmillan, Hampshire, 2004) ‘Gendered Discourses Model’ in order to examine the depictions of male and female (...) social actors within this gendered discourse. The gendered discourse of equal opportunities was buttressed by such representations within a tight perspective in proportion to gender ideologies prevailing in Iran. Resorting to CDA, we can claim that resistance against such gendered discourse in Iranian EFL textbooks militates against such gender norms. These representations of male and female social actors in school books are indicative of an all-encompassing education, reinforcing that the discourse of equal opportunities is yet to be realized in the education system of Iran. (shrink)

Structural representations are likely the most talked about representational posits in the contemporary debate over cognitive representations. Indeed, the debate surrounding them is so vast virtually every claim about them has been made. Some, for instance, claimed structural representations are different from indicators. Others argued they are the same. Some claimed structural representations mesh perfectly with mechanistic explanations, others argued they can’t in principle mash. Some claimed structural representations are central to predictive processing accounts of cognition, others rebuked predictive processing (...) networks are blissfully structural representation free. And so forth. Here, I suggest this confusing state of affairs is due to the fact that the term “structural representations” is applied to a number of distinct conceptions of representations. In this paper, I distinguish four such conceptions, argue that these four conceptions are actually distinct, and then show that such a fourfold distinction can be used to clarify some of the most pressing questions concerning structural representations and their role in cognitive theorizing, making these questions more easily answerable. (shrink)

The integration of standardized biomedical terminologies into a single, unified knowledge representation system has formed a key area of applied informatics research in recent years. The Unified Medical Language System (UMLS) is the most advanced and most prominent effort in this direction, bringing together within its Metathesaurus a large number of distinct source-terminologies. The UMLS Semantic Network, which is designed to support the integration of these source-terminologies, has proved to be a highly successful combination of formal coherence and (...) broad scope. We argue here, however, that its organization manifests certain structural problems, and we describe revisions which we believe are needed if the network is to be maximally successful in realizing its goals of supporting terminology integration. (shrink)

In this paper some classical representational ideas of Hertz and Duhem are used to show how the dichotomy between representation and intervention can be overcome. More precisely, scientific theories are reconstructed as complex networks of intervening representations (or representational interventions). The formal apparatus developed is applied to elucidate various theoretical and practical aspects of the in vivo/in vitro problem of biochemistry. Moreover, adjoint situations (Galois connections) are used to explain the relation berween empirical facts and theoretical laws in a (...) new way. (shrink)

The core objective of this project is to recognize and reconstruct distorted facial images, particularly in the context of accidents. This involves using deep learning techniques to analyze the features of a distorted face and regenerate it into a recognizable form. Deep learning models are wellsuited for this task due to their ability to learn complex patterns and representations from data the input data consists of distorted facial images, typically obtained from MRI scans of accident victims. These images may contain (...) various types of distortions such as swelling, bruising, or other injuries that affect facial appearance. (shrink)

Recent breakthroughs in artificial intelligence have primarily resulted from training deep neural networks (DNNs) with vast numbers of adjustable parameters on enormous datasets. Due to their complex internal structure, DNNs are frequently characterized as inscrutable ``black boxes,'' making it challenging to interpret the mechanisms underlying their impressive performance. This opacity creates difficulties for explanation, safety assurance, trustworthiness, and comparisons to human cognition, leading to divergent perspectives on these systems. This chapter examines recent developments in interpretability methods for DNNs, with a (...) focus on interventionist approaches inspired by causal explanation in philosophy of science. We argue that these methods offer a promising avenue for understanding how DNNs process information compared to merely behavioral benchmarking and correlational probing. We review key interventionist methods and illustrate their application through practical case studies. These methods allow researchers to identify and manipulate specific computational components within DNNs, providing insights into their causal structure and internal representations. We situate these approaches within the broader framework of causal abstraction, which aims to align low-level neural computations with high-level interpretable models. While acknowledging current limitations, we contend that interventionist methods offer a path towards more rigorous and theoretically grounded interpretability research, potentially informing both AI development and computational cognitive neuroscience. (shrink)

This study examines the representations of male and female social actors in selected Iranian EFL (English as a Foreign Language) textbooks. It is grounded in Critical Discourse Analysis and uses van Leeuwen’s Social Actor Network Model to analyze social actor representations in the gendered discourses of compulsory heterosexuality. Findings from the analysis show that the representations endorse the discourse of compulsory heterosexuality which is an institutionalized form of social practice in Iran. Three male and three female students were interviewed (...) to find out what they think about these representations. Their responses with regard to whether they think textbooks should also include representations of other forms of sexuality were non-committal and vague. To them LGBT people are the ‘‘Other’’ practicing a form of sexuality that is not normal. Such exclusions could obscure the reality regarding the existence of such gender identities and represent the world in a particular manner. (shrink)

The processes of computation and automation that produce digitized objects have displaced the concept of an image once conceived through optical devices such as a photographic plate or a camera mirror that were invented to accommodate the human eye. Computational images exist as information within networks mediated by machines. They are increasingly less about what art history understands as representation or photography considers indexing and more an operational product of data processing. Through genealogical, theoretical, and practice-based investigation, this dissertation (...) project traces a lineage of computation through images from early cybernetics to contemporary machine learning under algorithmic capitalist conditions of political economy. Objects include mundane examples such as smartphone cameras to sublime examples such as the imaging of black holes. It articulates new conditions and forms of perception and surveillance that result from contemporary technological development in computation. The project builds on theorists of computational media such as Friedrich Kittler, Wendy Hui Kyong Chun, Jussi Parikka, Shane Denson, and Alexander Galloway to examine the interplay between perception, computation, and epistemology. Further theoretical contribution builds on the work of Luciana Parisi and Jacques Ranciére’s articulation of aesthetics, and Michel Foucault’s gesture towards a heterotopia. Finally, it draws from Donna Haraway’s metaphors of the cyborg and sympoiesis to frame the capacity of art to transfigure the limitations of digital network culture into new possibilities for being. Here, it curates practices by Danielle Brathwaite-Shirley, Harun Farocki, Morehshin Allahyari, and Pierre Huyghe. (shrink)

The notions of information, representation and enaction entertain historical and complex relations with cognition. Historical relations because representational structures belong to the central hypothesis of cognitive sciences. Complex relations because cognitive sciences apply the notion of representation to animals, humans and robots, and also because the enactive approach tends to disregard the GOFAI type of representations. In this wide horizon of relations, we propose to look at a systemic approach that could bring up a common denominator for information (...) and representations in the build up of cognition, and also keep a link with the enactive approach. Our purpose is to show that systems submitted to constraints can generate meaningful information to maintain their natures, and consequently build up meaningful representations that have some compatibility with the enactive approach. Such a systemic approach to the notion of meaningful information could then make available a link between enaction and meaningful representations. The first part of the presentation is about reminding that cognition does not exist per se, but is related to the system that builds it. We look at cognition as constituted by dynamic meaningful representations built up by systems that have constraints to satisfy in their environments. Cognition is considered here at the level of the system that builds it and uses it in order to maintain its nature in its environment. Such a systemic approach fits with evolution. Organisms build representations to cope with survival constraints (frogs build representations of moving black dots in order to satisfy food constraints). Humans build representations and cognition to satisfy constraints that are conscious and unconscious. Artificial systems can use representations and cognition to run activities related to constraints implemented by the designers or coming from the environment (a goal to reach being considered as a constraint to satisfy). In the second part of the presentation we define what are a meaningful information and a representation for a system submitted to a constraint in its environment, and we link these to the enactive approach. We define a meaningful information (a meaning) as an information generated by a system submitted to a constraint when it receives an external information that has a connection with the constraint. The meaning is precisely that connection. The meaning belongs to the interactions that link the system to its environment. The function of the meaning is to participate to the determination of an action that will be implemented in order to satisfy the constraint. (Menant, 2003). The satisfaction of the constraint goes with maintaining the nature of the system in its environment. A Meaning Generator System (MGS) is defined correspondingly. It is a building block for higher level systems. We present some characteristics of the MGS (groundings of a meaning, domain of efficiency and transfer of meanings, networking of meanings, evolutionary usage). The MGS approach is close to a simplified version of the Peircean triadic theory of signs (Menant, 2003, 2005 ). We define the representation of an item for a system as being the dynamic set of meaningful information corresponding to the item for the system in its environments (an elementary representation being made of a single meaningful information). These representations link the system to its environment by their meaningful components related to the nature of the system. These representations are different from the GOFAI ones. The possibilities for linking these notions of meaning and representation with the enactive approach come from the structure of the MGS: the need for an action is the cause of the meaning generation by and for the system. The action on the environment is for the system to maintain its nature (its identity). The MGS links together the generation of meaningful representations, the nature of the system, and the interactions with the environment. This can be considered as close to enacting a world by meaning generation (Di Paolo and all 2007), and to the enactive concept of sense making (De Jaegher, Di Paolo 2007). We propose that basing the definition of a representation on the notion of meaningful information generated by a system submitted to a constraint can open a way for making the notion of representation compatible with the enactive approach. In the third part of the presentation, we consider some cases of meaningful information and representations for organisms and for robots. Regarding organisms, the MGS can be used in an evolutionary context by looking at the evolution of the systems and of the constraints. Purpose is to modelize the generation of meanings and of representations in order to make available a tool usable for different levels of evolution, as evolution has a place in cognitive sciences (Proust, 2007). Constraints for basic life are survival constraints (individual and species). Group life constraints are also to be considered. Reaching the level of humans in evolution brings in new constraints that cannot be clearly identified as they have to take into account human consciousness which is today a mystery (the “hard problem”). On an evolutionary standpoint, human constraints come in addition to the ones existing for non human organisms. We can make some hypothesis on the nature of human constraints (Maslow pyramid based constraints, anxiety limitation…). For robots, the MGS is initially based on the design of the robot. The meaning generated within a robot is initially derived from the constraints implemented by the designer and from the environment. But some non calculable or non predictable evolutions of the robot can introduce meanings that look proper to the robot. This last point can be linked to the notion of autonomy in robots. In such examples, the dynamic management of meanings thru the MGSs in their environments keeps the link with the enactive approach. We finish the presentation by summarising the points addressed and by proposing several continuations. (shrink)

In artificial intelligence, recent research has demonstrated the remarkable potential of Deep Convolutional Neural Networks (DCNNs), which seem to exceed state-of-the-art performance in new domains weekly, especially on the sorts of very difficult perceptual discrimination tasks that skeptics thought would remain beyond the reach of artificial intelligence. However, it has proven difficult to explain why DCNNs perform so well. In philosophy of mind, empiricists have long suggested that complex cognition is based on information derived from sensory experience, often appealing to (...) a faculty of abstraction. Rationalists have frequently complained, however, that empiricists never adequately explained how this faculty of abstraction actually works. In this paper, I tie these two questions together, to the mutual benefit of both disciplines. I argue that the architectural features that distinguish DCNNs from earlier neural networks allow them to implement a form of hierarchical processing that I call “transformational abstraction”. Transformational abstraction iteratively converts sensory-based representations of category exemplars into new formats that are increasingly tolerant to “nuisance variation” in input. Reflecting upon the way that DCNNs leverage a combination of linear and non-linear processing to efficiently accomplish this feat allows us to understand how the brain is capable of bi-directional travel between exemplars and abstractions, addressing longstanding problems in empiricist philosophy of mind. I end by considering the prospects for future research on DCNNs, arguing that rather than simply implementing 80s connectionism with more brute-force computation, transformational abstraction counts as a qualitatively distinct form of processing ripe with philosophical and psychological significance, because it is significantly better suited to depict the generic mechanism responsible for this important kind of psychological processing in the brain. (shrink)

Symmetry in biological and physical systems is a product of self-organization driven by evolutionary processes, or mechanical systems under constraints. Symmetry-based feature extraction or representation by neural networks may unravel the most informative contents in large image databases. Despite significant achievements of artificial intelligence in recognition and classification of regular patterns, the problem of uncertainty remains a major challenge in ambiguous data. In this study, we present an artificial neural network that detects symmetry uncertainty states in human observers. (...) To this end, we exploit a neural network metric in the output of a biologically inspired Self- Organizing Map Quantization Error (SOM-QE). Shape pairs with perfect geometry mirror symmetry but a non-homogenous appearance, caused by local variations in hue, saturation, or lightness within and/or across the shapes in a given pair produce, as shown here, a longer choice response time (RT) for “yes” responses relative to symmetry. These data are consistently mirrored by the variations in the SOM-QE from unsupervised neural network analysis of the same stimulus images. The neural network metric is thus capable of detecting and scaling human symmetry uncertainty in response to patterns. Such capacity is tightly linked to the metric’s proven selectivity to local contrast and color variations in large and highly complex image data. (shrink)

The nature of “the self” has been one of the central problems in philosophy and more recently in neuroscience. This raises various questions: Can we attribute a self to animals? Do animals and humans share certain aspects of their core selves, yielding a trans-species concept of self? What are the neural processes that underlie a possible trans-species concept of self? What are the developmental aspects and do they result in various levels of self-representation? Drawing on recent literature from both (...) human and animal research, we suggest a trans-species concept of self that is based upon what has been called a “core-self” which can be described by self-related processing as a specific mode of interaction between organism and environment. When we refer to specific neural networks, we will here refer to the underlying system as the “core-SELF.” The core-SELF provides primordial neural coordinates that represent organisms as living creatures—at the lowest level this elaborates interoceptive states along with raw emotional feelings while higher medial cortical levels facilitate affective-cognitive integration . Developmentally, SRP allows stimuli from the environment to be related and linked to organismic needs, signaled and processed within core-self structures within subcorical-cortical midline structures that provide the foundation for epigenetic emergence of ecologically framed, higher idiographic forms of selfhood across different individuals within a species. These functions ultimately operate as a coordinated network. We postulate that core SRP operates automatically, is deeply affective, and is developmentally and epigenetically connected to sensory-motor and higher cognitive abilities. This core-self is mediated by SCMS, embedded in visceral and instinctual representations of the body that are well integrated with basic attentional, emotional and motivational functions that are apparently shared between humans, non-human mammals, and perhaps in a proto-SELF form, other vertebrates. Such a trans-species concept of organismic coherence is thoroughly biological and affective at the lowest levels of a complex neural network, and culturally and ecologically molded at higher levels of neural processing. It allows organisms to selectively adapt to and integrate with physical and social environments. Such a psychobiologically universal, but environmentally diversified, concept may promote novel trans-species studies of the core-self across mammalian species. (shrink)

Grete Hermann’s essay “Die naturphilosophischen Grundlagen der Quantenmechanik” has received much deserved scholarly attention in recent years. In this paper, I follow the lead of Elise Crull who sees in Hermann’s work the general outlines of a neo-Kantian interpretation of quantum theory. In full support of this view, I focus on Hermann’s central claim that limited spatio-temporal, and even analogically causal, representations of events exist within an overall relational structure of entangled quantum mechanical states that defy any unified spatio-temporal description. (...) In my view, Hermann also advances an important transcendental argument that perspectival spatio-temporal representations of nature have their foundations in general relational networks that are not spatio-temporal. The key point is that the adoption of a perspectival system within the general network induces the representation but only for that context. These ideas are consistent with a perspectival subject–object principle in Kant and also with Weyl’s work on Lie groups and their representations. (shrink)

Many human decisions and behaviors in daily life entail a cost-benefit analysis. From selecting what to eat for dinner to determining the career to pursue, we more or less assess the cost and benefit of each choice. Given the frequent occurrences of cost-benefit thinking in our minds, some intriguing questions arise: how do the cost-benefit thinking processes emerge? How does the brain function to generate such thoughts? Although these inquiries have yet to be thoroughly answered, scientists are adding new clues (...) to our knowledge, such as the neural representation of our subjective cost-benefit judgments. (shrink)

This report highlights and explores five questions which arose from the workshop on perceptual learning and perceptual recognition at the University of Toronto, Mississauga on May 10th and 11th, 2012: 1. How should we demarcate perceptual learning from perceptual development? 2. What are the origins of multimodal associations? 3. Does our representation of time provide an amodal framework for multi-sensory integration? 4. What counts as cognitive penetration? 5. How can philosophers and psychologists most fruitfully collaborate?

This article addresses computational procedures that are no longer constrained by human modes of representation and considers how these procedures could be philosophically understood in terms of ‘algorithmic thought’. Research in deep learning is its case study. This artificial intelligence (AI) technique operates in computational ways that are often opaque. Such a black-box character demands rethinking the abstractive operations of deep learning. The article does so by entering debates about explainability in AI and assessing how technoscience and technoculture tackle (...) the possibility to ‘re-present’ the algorithmic procedures of feature extraction and feature learning to the human mind. The article thus mobilises the notion of incommensurability (originally developed in the philosophy of science) to address explainability as a communicational and representational issue, which challenges phenomenological and existential modes of comparison between human and algorithmic ‘thinking’ operations. (shrink)

Understanding computation as “a process of the dynamic change of information” brings to look at the different types of computation and information. Computation of information does not exist alone by itself but is to be considered as part of a system that uses it for some given purpose. Information can be meaningless like a thunderstorm noise, it can be meaningful like an alert signal, or like the representation of a desired food. A thunderstorm noise participates to the generation of (...) meaningful information about coming rain. An alert signal has a meaning as allowing a safety constraint to be satisfied. The representation of a desired food participates to the satisfaction of some metabolic constraints for the organism. Computations on information and representations will be different in nature and in complexity as the systems that link them have different constraints to satisfy. Animals have survival constraints to satisfy. Humans have many specific constraints coming in addition. And computers will compute what the designer and programmer ask for. We propose to analyze the different relations between information, meaning and representation by taking an evolutionary approach on the systems that link them. Such a bottom-up approach allows starting with simple organisms and avoids an implicit focus on humans, which is the most complex and difficult case. To make available a common background usable for the many different cases, we use a systemic tool that defines the generation of meaningful information by and for a system submitted to a constraint [Menant, 2003]. This systemic tool allows to position information, meaning and representations for systems relatively to environmental entities in an evolutionary perspective. We begin by positioning the notions of information, meaning and representation and recall the characteristics of the Meaning Generator System (MGS) that link a system submitted to a constraint to its environment. We then use the MGS for animals and highlight the network nature of the interrelated meanings about an entity of the environment. This brings us to define the representation of an item for an agent as being the network of meanings relative to the item for the agent. Such meaningful representations embed the agents in their environments and are far from the Good Old Fashion Artificial Intelligence type ones. The MGS approach is then used for humans with a limitation resulting of the unknown nature of human consciousness. Application of the MGS to artificial systems brings to look for compatibilities with different levels of Artificial Intelligence (AI) like embodied-situated AI, the Guidance Theory of Representations, and enactive AI. Concerns relative to different types of autonomy and organic or artificial constraints are highlighted. We finish by summarizing the points addressed and by proposing some continuations. (shrink)

Two universal functional principles of Grossberg’s Adaptive Resonance Theory [19] decipher the brain code of all biological learning and adaptive intelligence. Low-level representations of multisensory stimuli in their immediate environmental context are formed on the basis of bottom-up activation and under the control of top-down matching rules that integrate high-level long-term traces of contextual configuration. These universal coding principles lead to the establishment of lasting brain signatures of perceptual experience in all living species, from aplysiae to primates. They are re-visited (...) in this concept paper here on the basis of examples drawn from the original code and from some of the most recent related empirical findings on contextual modulation in the brain, highlighting the potential of Grossberg’s pioneering insights and groundbreaking theoretical work for intelligent solutions in the domain of developmental and cognitive robotics. (shrink)

1. Do models formulated in programming languages use explicit rules where connectionist models do not? 2. Are rules as found in programming languages hard, precise, and exceptionless, where connectionist rules are not? 3. Do connectionist models use rules operating on distributed representations where models formulated in programming languages do not? 4. Do connectionist models fail to use structure sensitive rules of the sort found in "classical" computer architectures? In this chapter we argue that the answer to each of these questions (...) is negative. (shrink)

This dissertation focuses on generative models in the Predictive Processing framework. It is commonly accepted that generative models are structural representations; i.e. physical particulars representing via structural similarity. Here, I argue this widespread account is wrong: when closely scrutinized, generative models appear to be non-representational control structures realizing an agent’s sensorimotor skills. The dissertation opens (Ch.1) introducing the Predictive Processing account of perception and action, and presenting some of its connectionist implementations, thereby clarifying the role generative models play in Predictive (...) Processing. Subsequently, I introduce the conceptual framework guiding the research (ch.2). I briefly elucidate the metaphysics of representations, emphasizing the specific functional role played by representational vehicles within the systems of which they are part. I close the first half of the dissertation (Ch.3) introducing the claim that generative models are structural representations, and defending it from intuitive but inconclusive objections. I then move to the second half of the dissertation, switching from exposition to criticism. First (Ch.4), I claim that the argument allegedly establishing that generative models are structural representations is flawed beyond repair, for it fails to establish generative models are structurally similar to their targets. I then consider alternative ways to establish that structural similarity, showing they all either fail or violate some other condition individuating structural representations. I further argue (Ch.5) that the claim that generative models are structural representations would not be warranted even if the desired structural similarity were established. For, even if generative models were to satisfy the relevant definition of structural representation, it would still be wrong to consider them as representations. This is because, as currently defined,structural representations fail to play the relevant functional role of representations, and thus cannot be rightfully identified as representations in the first place. This conclusion prompts a direct examination of generative models, to determine their nature (Ch.6). I thus analyze the simplest generative model I know of: a neural network functioning as a robotic “brain” and allowing different robotic creatures to swiftly and intelligently interact with their environments. I clarify how these networks allow the robots to acquire and exert the relevant sensorimotor abilities needed to solve the various cognitive tasks the robots are faced with, and then argue that neither the entire architecture nor any of its parts can possibly qualify as representational vehicles. In this way, the structures implementing generative models are revealed to be non-representational structures that instantiate an agent’s relevant sensorimotor skills. I show that my conclusion generalizes beyond the simple example I considered, arguing that adding computational ingredients to the architecture, or considering altogether different implementations of generative models, will in no way force a revision of my verdict. I further consider and allay a number of theoretical worries that it might generate, and then briefly conclude the dissertation. (shrink)

The Protein Ontology (PRO; http://proconsortium.org) formally defines protein entities and explicitly represents their major forms and interrelations. Protein entities represented in PRO corresponding to single amino acid chains are categorized by level of specificity into family, gene, sequence and modification metaclasses, and there is a separate metaclass for protein complexes. All metaclasses also have organism-specific derivatives. PRO complements established sequence databases such as UniProtKB, and interoperates with other biomedical and biological ontologies such as the Gene Ontology (GO). PRO relates to (...) UniProtKB in that PRO’s organism-specific classes of proteins encoded by a specific gene correspond to entities documented in UniProtKB entries. PRO relates to the GO in that PRO’s representations of organism-specific protein complexes are subclasses of the organism-agnostic protein complex terms in the GO Cellular Component Ontology. The past few years have seen growth and changes to the PRO, as well as new points of access to the data and new applications of PRO in immunology and proteomics. Here we describe some of these developments. (shrink)

Many Public Administrations structure their services around the notion of users’ need. However, there is a gap between private, subjectively perceived needs (self-attributed) and needs that are attributed by PA to citizens (heteroattributed). Because of the gap, citizens’ needs are often only partially satisfied by PAs services. This gap is in part due to the fact that the meaning of the word “need” is ambiguous and full of antinomic nuances. The purpose of this paper is to formulate a definition of (...) “need” suitable for citizens’ needs management with respect to PA’s services offering, and to provide an accurate ontological analysis of the notion of “need” and the network of concepts that relate to it. (shrink)

This is an excerpt from a report that highlights and explores five questions which arose from the workshop on perceptual learning and perceptual recognition at the University of Toronto, Mississauga on May 10th and 11th, 2012. This excerpt explores the question: Does our representation of time provide and amodal framework for multi-sensory integration?

We introduce the notion of complexity, first at an intuitive level and then in relatively more concrete terms, explaining the various characteristic features of complex systems with examples. There exists a vast literature on complexity, and our exposition is intended to be an elementary introduction, meant for a broad audience. -/- Briefly, a complex system is one whose description involves a hierarchy of levels, where each level is made of a large number of components interacting among themselves. The time evolution (...) of such a system is of a complex nature, depending on the interactions among subsystems in the next level below the one under consideration and, at the same time, conditioned by the level above, where the latter sets the context for the evolution. Generally speaking, the interactions among the constituents of the various levels lead to a dynamics characterized by numerous characteristic scales, each level having its own set of scales. What is more, a level commonly exhibits ‘emergent properties’ that cannot be derived from considerations relating to its component systems taken in isolation or to those in a different contextual setting. In the dynamic evolution of some particular level, there occurs a self-organized emergence of a higher level and the process is repeated at still higher levels. -/- The interaction and self-organization of the components of a complex system follow the principle commonly expressed by saying that the ‘whole is different from the sum of the parts’. In the case of systems whose behavior can be expressed mathematically in terms of differential equations this means that the interactions are nonlinear in nature. -/- While all of the above features are not universally exhibited by complex systems, these are nevertheless indicative of a broad commonness relative to which individual systems can be described and analyzed. There exist measures of complexity which, once again, are not of universal applicability, being more heuristic than exact. The present state of knowledge and understanding of complex systems is itself an emerging one. Still, a large number of results on various systems can be related to their complex character, making complexity an immensely fertile concept in the study of natural, biological, and social phenomena. -/- All this puts a very definite limitation on the complete description of a complex system as a whole since such a system can be precisely described only contextually, relative to some particular level, where emergent properties rule out an exact description of more than one levels within a common framework. -/- We discuss the implications of these observations in the context of our conception of the so-called noumenal reality that has a mind-independent existence and is perceived by us in the form of the phenomenal reality. The latter is derived from the former by means of our perceptions and interpretations, and our efforts at sorting out and making sense of the bewildering complexity of reality takes the form of incessant processes of inference that lead to theories. Strictly speaking, theories apply to models that are constructed as idealized versions of parts of reality, within which inferences and abstractions can be carried out meaningfully, enabling us to construct the theories. -/- There exists a correspondence between the phenomenal and the noumenal realities in terms of events and their correlations, where these are experienced as the complex behavior of systems or entities of various descriptions. The infinite diversity of behavior of systems in the phenomenal world are explained within specified contexts by theories. The latter are constructs generated in our ceaseless attempts at interpreting the world, and the question arises as to whether these are reflections of `laws of nature' residing in the noumenal world. This is a fundamental concern of scientific realism, within the fold of which there exists a trend towards the assumption that theories express truths about the noumenal reality. We examine this assumption (referred to as a ‘point of view’ in the present essay) closely and indicate that an alternative point of view is also consistent within the broad framework of scientific realism. This is the view that theories are domain-specific and contextual, and that these are arrived at by independent processes of inference and abstractions in the various domains of experience. Theories in contiguous domains of experience dovetail and interpenetrate with one another, and bear the responsibility of correctly explaining our observations within these domains. -/- With accumulating experience, theories get revised and the network of our theories of the world acquires a complex structure, exhibiting a complex evolution. There exists a tendency within the fold of scientific realism of interpreting this complex evolution in rather simple terms, where one assumes (this, again, is a point of view) that theories tend more and more closely to truths about Nature and, what is more, progress towards an all-embracing ‘ultimate theory’ -- a foundational one in respect of all our inquiries into nature. We examine this point of view closely and outline the alternative view -- one broadly consistent with scientific realism -- that there is no ‘ultimate’ law of nature, that theories do not correspond to truths inherent in reality, and that successive revisions in theory do not lead monotonically to some ultimate truth. Instead, the theories generated in succession are incommensurate with each other, testifying to the fact that a theory gives us a perspective view of some part of reality, arrived at contextually. Instead of resembling a monotonically converging series successive theories are analogous to asymptotic series. -/- Before we summarize all the above considerations, we briefly address the issue of the complexity of the {\it human mind} -- one as pervasive as the complexity of Nature at large. The complexity of the mind is related to the complexity of the underlying neuronal organization in the brain, which operates within a larger biological context, its activities being modulated by other physiological systems, notably the one involving a host of chemical messengers. The mind, with no materiality of its own, is nevertheless emergent from the activity of interacting neuronal assemblies in the brain. As in the case of reality at large, there can be no ultimate theory of the mind, from which one can explain and predict the entire spectrum of human behavior, which is an infinitely rich and diverse one. (shrink)

Recent academic fashions have posited visual images of colonial landscape space as forming part of a network of intellectual influences that promoted both a culture of imperialism and an imperial culture in the nineteenth century. Frequently these analyses concentrate on constructing an overarching socio-political interpretation into which to place this art, thereby ignoring the influence of artistic and aesthetic theory in the creation, assessment and reception of these images.

Mental representations remain the central posits of psychology after many decades of scrutiny. However, there is no consensus about the representational format(s) of biological cognition. This paper provides a survey of evidence from computational cognitive psychology, perceptual psychology, developmental psychology, comparative psychology, and social psychology, and concludes that one type of format that routinely crops up is the language-of-thought (LoT). We outline six core properties of LoTs: (i) discrete constituents; (ii) role-filler independence; (iii) predicate–argument structure; (iv) logical operators; (v) inferential (...) promiscuity; and (vi) abstract content. These properties cluster together throughout cognitive science. Bayesian computational modeling, compositional features of object perception, complex infant and animal reasoning, and automatic, intuitive cognition in adults all implicate LoT-like structures. Instead of regarding LoT as a relic of the previous century, researchers in cognitive science and philosophy-of-mind must take seriously the explanatory breadth of LoT-based architectures. We grant that the mind may harbor many formats and architectures, including iconic and associative structures as well as deep-neural-network-like architectures. However, as computational/representational approaches to the mind continue to advance, classical compositional symbolic structures – that is, LoTs – only prove more flexible and well-supported over time. (shrink)

We model scientific theories as Bayesian networks. Nodes carry credences and function as abstract representations of propositions within the structure. Directed links carry conditional probabilities and represent connections between those propositions. Updating is Bayesian across the network as a whole. The impact of evidence at one point within a scientific theory can have a very different impact on the network than does evidence of the same strength at a different point. A Bayesian model allows us to envisage and (...) analyze the differential impact of evidence and credence change at different points within a single network and across different theoretical structures. (shrink)

The representations of deep convolutional neural networks (CNNs) are formed from generalizing similarities and abstracting from differences in the manner of the empiricist theory of abstraction (Buckner, Synthese 195:5339–5372, 2018). The empiricist theory of abstraction is well understood to entail infinite regress and circularity in content constitution (Husserl, Logical Investigations. Routledge, 2001). This paper argues these entailments hold a fortiori for deep CNNs. Two theses result: deep CNNs require supplementation by Quine’s “apparatus of identity and quantification” in order to (1) (...) achieve concepts, and (2) represent objects, as opposed to “half-entities” corresponding to similarity amalgams (Quine, Quintessence, Cambridge, 2004, p. 107). Similarity amalgams are also called “approximate meaning[s]” (Marcus & Davis, Rebooting AI, Pantheon, 2019, p. 132). Although Husserl inferred the “complete abandonment of the empiricist theory of abstraction” (a fortiori deep CNNs) due to the infinite regress and circularity arguments examined in this paper, I argue that the statistical learning of deep CNNs may be incorporated into a Fodorian hybrid account that supports Quine’s “sortal predicates, negation, plurals, identity, pronouns, and quantifiers” which are representationally necessary to overcome the regress/circularity in content constitution and achieve objective (as opposed to similarity-subjective) representation (Burge, Origins of Objectivity. Oxford, 2010, p. 238). I base myself initially on Yoshimi’s (New Frontiers in Psychology, 2011) attempt to explain Husserlian phenomenology with neural networks but depart from him due to the arguments and consequently propose a two-system view which converges with Weiskopf’s proposal (“Observational Concepts.” The Conceptual Mind. MIT, 2015. 223–248). (shrink)

Generative Adversarial Networks (GANs) have demonstrated significant potential in generating synthetic data for various applications, including those involving sensitive information like healthcare and finance. However, two major issues arise when GANs are applied to sensitive datasets: (i) the model may memorize training samples, compromising the privacy of individuals, especially when the data includes personally identifiable information (PII), and (ii) there is a lack of control over the specificity of the generated samples, which limits their utility for tailored usecases. To address (...) these challenges, we propose a novel framework that integrates differential privacy with latent representation learning to ensure privacy while providing control over the specificity of generated data. Our approach ensures that the synthetic data does not reveal individual data points, and by learning effective latent codes, it allows for the generation of specific and meaningful data. We evaluate our method using the MNIST dataset, showing that it preserves privacy and demonstrates a privacy-utility trade-off, where increased privacy leads to decreased classification accuracy. Additionally, we highlight the computational challenges, as the training process incurs a tenfold increase in time compared to standard GAN models. Finally, we extend our approach to the CelebA dataset, demonstrating how privacy and specificity can be controlled to generate high-quality, private synthetic data. (shrink)

The vast number of neurons in the brain are ceaselessly engaged in spontaneously generated activity in virtue of interactions between those. It is in the background of this intrinsic activity that the brain responds to signals from the environment and from endogenous signals received by way of active mental processes. This spontaneous activity persists in the `resting state' and is modulated by evoked signals resulting from task-induced activity. The two together generate an ongoing process of self-organization in the brain whereby (...) dynamic excitation patterns are set up and correlated with one another in neuronal aggregates of diverse population size in an intricately nested hierarchy. These correlated patterns constitute distributed models of reality, based on which the brain generates expectations, plans, and predictions relating to the dynamically evolving world that we all are embedded in. -/- The process of neuronal self-organization provides the basis for the emergence of a structured mind exhibiting an exquisite functional complexity. Generally speaking, mental (and, in particular, psychological) processes are based on large scale (`macroscopic') neuronal networks, while microscopic and mesoscopic neuronal populations constitute the `substratum' of the mind, where elementary and tiny bits of experience are represented so as to act as atomic ingredients in the more complex processing of compositely constituted information involved in mental activities. -/- Processes in the microscopic and mesoscopic neuronal aggregates occur on relatively short spatial and temporal scales while those in macroscopic networks are slower and distributed across brain regions. The large scale networks operate near the `edge of instability', which means that these networks, taken together, form a system close to criticality---this is what makes the mind possessed of its remarkable fluidity. -/- The neuronal aggregates themselves are in a constant state of flux, with groups of neurons constantly entering and leaving larger aggregates, and with neurons being simultaneously operative in the functioning of more than one aggregates at a time. In the backdrop of this enormous complexity, the operations of the mind constitute a complementary description of the brain-mind complex to those of the neuronal populations---neither of the two is reducible to the other. In other words, the functioning of the mind is not determinable in terms of the activities of the neuronal populations. -/- The `substratum of the mind', made up of microscopic and mesoscopic populations of neurons, in addition to providing the hinterland in which latent operations supporting all mental processes are enacted, is involved more directly with representations of elementary concepts, with distributed memory (in a broad sense), and with heuristics of an elementary and atomic nature that help us in making vast numbers of `decisions' of a similarly elementary nature. (shrink)

This chapter explores the influence of Artificial Intelligence (AI) on digital democracy, focusing on four main areas: citizenship, participation, representation, and the public sphere. It traces the evolution from electronic to virtual and network democracy, underscoring how each stage has broadened democratic engagement through technology. Focusing on digital citizenship, the chapter examines how AI can improve online engagement while posing privacy risks and fostering identity stereotyping. Regarding political participation, it highlights AI's dual role in mobilising civic actions and (...) spreading misinformation. Regarding representation, AI's involvement in electoral processes can enhance voter registration, e-voting, and the efficiency of result tabulation but raises concerns regarding privacy and public trust. Also, AI's predictive capabilities shift the dynamics of political competition, posing ethical questions about manipulation and the legitimacy of democracy. Finally, the chapter examines how integrating AI and digital technologies can facilitate democratic political advocacy and personalised communication. However, this also comes with higher risks of misinformation and targeted propaganda. (shrink)

This project presents an innovative approach to the efficient enumeration of hidden service URLs (.onion) on the TOR network using a custom-built graphical user interface (GUI) application. The tool leverages web scraping techniques, targeting the Ahmia search engine, to retrieve and analyze active hidden server URLs. Through the integration of Python’s Tkinter library for GUI development and the use of requests and regular expressions for content mining, the application simplifies the process of discovering and visualizing .onion URLs. The tool (...) allows users to input search queries, clean them for web compatibility, and display the results in real time. Additionally, the application provides graphical insights into the search results by generating visual data representations using matplotlib. This paper discusses the architecture, development process, and functionality of the tool, providing a new method for researchers to explore the TOR network. (shrink)

Scientific studies have shown that non-conscious stimuli and représentations influence information processing during conscious experience. In the light of such evidence, questions about potential functional links between non-conscious brain representations and conscious experience arise. This article discusses models capable of explaining how statistical learning mechanisms in dedicated resonant circuits could generate specific temporal activity traces of non-conscious representations in the brain. How reentrant signaling, top-down matching, and statistical coincidence of such activity traces may lead to the progressive consolidation of neural (...) signatures of conscious experience in networks extending across large distances beyond functionally specialized brain regions is then explained. (shrink)

We explicate representational content by addressing how representations that ex- plain intelligent behavior might be acquired through processes of Darwinian evo- lution. We present the results of computer simulations of evolved neural network controllers and discuss the similarity of the simulations to real-world examples of neural network control of animal behavior. We argue that focusing on the simplest cases of evolved intelligent behavior, in both simulated and real organisms, reveals that evolved representations must carry information about the creature’s (...) environ- ments and further can do so only if their neural states are appropriately isomor- phic to environmental states. Further, these informational and isomorphism rela- tions are what are tracked by content attributions in folk-psychological and cognitive scientific explanations of these intelligent behaviors. (shrink)

SNePS, the Semantic Network Processing System 45, 54], has been designed to be a system for representing the beliefs of a natural-language-using intelligent system (a \cognitive agent"). It has always been the intention that a SNePS-based \knowledge base" would ultimatelybe built, not by a programmeror knowledge engineer entering representations of knowledge in some formallanguage or data entry system, but by a human informing it using a natural language (NL) (generally supposed to be English), or by the system reading books (...) or articles that had been prepared for human readers. Because of this motivation, the criteria for the development of SNePS have included: it should be able to represent anything and everything expressible in NL; it should be able to represent generic, as well as speci c information; it should be able to use the generic and the speci c information to reason and infer information implied by what it has been told; it cannot count on any particular order among the pieces of information it is given; it must continue to act reasonably even if the information it is given includes circular de nitions, recursive rules, and inconsistent information. (shrink)

This book reports on the results of the third edition of the premier conference in the field of philosophy of artificial intelligence, PT-AI 2017, held on November 4 - 5, 2017 at the University of Leeds, UK. It covers: advanced knowledge on key AI concepts, including complexity, computation, creativity, embodiment, representation and superintelligence; cutting-edge ethical issues, such as the AI impact on human dignity and society, responsibilities and rights of machines, as well as AI threats to humanity and AI (...) safety; and cutting-edge developments in techniques to achieve AI, including machine learning, neural networks, dynamical systems. The book also discusses important applications of AI, including big data analytics, expert systems, cognitive architectures, and robotics. It offers a timely, yet very comprehensive snapshot of what is going on in the field of AI, especially at the interfaces between philosophy, cognitive science, ethics and computing. (shrink)

Endowing artificial systems with explanatory capacities about the reasons guiding their decisions, represents a crucial challenge and research objective in the current fields of Artificial Intelligence (AI) and Computational Cognitive Science [Langley et al., 2017]. Current mainstream AI systems, in fact, despite the enormous progresses reached in specific tasks, mostly fail to provide a transparent account of the reasons determining their behavior (both in cases of a successful or unsuccessful output). This is due to the fact that the classical problem (...) of opacity in artificial neural networks (ANNs) explodes with the adoption of current Deep Learning techniques [LeCun, Bengio, Hinton, 2015]. In this paper we argue that the explanatory deficit of such techniques represents an important problem, that limits their adoption in the cognitive modelling and computational cognitive science arena. In particular we will show how the current attempts of providing explanations of the deep nets behaviour (see e.g. [Ritter et al. 2017] are not satisfactory. As a possibile way out to this problem, we present two different research strategies. The first strategy aims at dealing with the opacity problem by providing a more abstract interpretation of neural mechanisms and representations. This approach is adopted, for example, by the biologically inspired SPAUN architecture [Eliasmith et al., 2012] and by other proposals suggesting, for example, the interpretation of neural networks in terms of the Conceptual Spaces framework [Gärdenfors 2000, Lieto, Chella and Frixione, 2017]. All such proposals presuppose that the neural level of representation can be considered somehow irrelevant for attacking the problem of explanation [Lieto, Lebiere and Oltramari, 2017]. In our opinion, pursuing this research direction can still preserve the use of deep learning techniques in artificial cognitive models provided that novel and additional results in terms of “transparency” are obtained. The second strategy is somehow at odds with respect to the previous one and tries to address the explanatory issue by avoiding to directly solve the “opacity” problem. In this case, the idea is that one of resorting to pre-compiled plausible explanatory models of the word used in combination with deep-nets (see e.g. [Augello et al. 2017]). We argue that this research agenda, even if does not directly fits the explanatory needs of Computational Cognitive Science, can still be useful to provide results in the area of applied AI aiming at shedding light on the models of interaction between low level and high level tasks (e.g. between perceptual categorization and explanantion) in artificial systems. (shrink)

Detecting fraud in financial transactions is crucial for guaranteeing the integrity and security of financial systems. This paper presents an integrated approach for detecting fraudulent activities that incorporates Support Vector Machines (SVM) and Feedforward Neural Networks (FFNN). The proposed methodology utilizes the strengths of SVM and FFNN to distinguish between classes and capture complex patterns and relationships, respectively. The SVM model functions as a feature extractor, supplying the FFNN with high-level representations as inputs. Through an exhaustive evaluation utilizing labeled financial (...) transaction data, the integrated SVM-FFNN model shows promise in detecting fraud with increased accuracy and precision. This research contributes to the development of innovative techniques for enhancing financial fraud detection systems. (shrink)