How should a machine intelligence perform unsupervised structure discovery over streams of sensory input? One approach to this problem is to cast it as an apperception task [1]. Here, the task is to construct an explicit interpretable theory that both explains the sensory sequence and also satisfies a set of unity conditions, designed to ensure that the constituents of the theory are connected in a relational structure. However, the original formulation of the apperception task had one fundamental limitation: (...) it assumed the raw sensory input had already been parsed using a set of discrete categories, so that all the system had to do was receive this already-digested symbolic input, and make sense of it. But what if we don't have access to pre-parsed input? What if our sensory sequence is raw unprocessed information? The central contribution of this paper is a neuro-symbolic framework for distilling interpretable theories out of streams of raw, unprocessed sensory experience. First, we extend the definition of the apperception task to include ambiguous (but still symbolic) input: sequences of sets of disjunctions. Next, we use a neural network to map raw sensory input to disjunctive input. Our binary neural network is encoded as a logic program, so the weights of the network and the rules of the theory can be solved jointly as a single SAT problem. This way, we are able to jointly learn how to perceive (mapping raw sensory information to concepts) and apperceive (combining concepts into declarative rules). (shrink)

This paper attempts to answer a central question in unsupervised learning: what does it mean to “make sense” of a sensory sequence? In our formalization, making sense involves constructing a symbolic causal theory that both explains the sensory sequence and also satisfies a set of unity conditions. The unity conditions insist that the constituents of the causal theory – objects, properties, and laws – must be integrated into a coherent whole. On our account, making sense of sensory input is (...) a type of program synthesis, but it is unsupervised program synthesis. Our second contribution is a computer implementation, the Apperception Engine, that was designed to satisfy the above requirements. Our system is able to produce interpretable human-readable causal theories from very small amounts of data, because of the strong inductive bias provided by the unity conditions. A causal theory produced by our system is able to predict future sensor readings, as well as retrodict earlier readings, and impute (fill in the blanks of) missing sensory readings, in any combination. In fact, it is able to do all three tasks simultaneously. We tested the engine in a diverse variety of domains, including cellular automata, rhythms and simple nursery tunes, multi-modal binding problems, occlusion tasks, and sequence induction intelligence tests. In each domain, we test our engine's ability to predict future sensor values, retrodict earlier sensor values, and impute missing sensory data. The Apperception Engine performs well in all these domains, significantly out-performing neural net baselines. We note in particular that in the sequence induction intelligence tests, our system achieved human-level performance. This is notable because our system is not a bespoke system designed specifically to solve intelligence tests, but a general-purpose system that was designed to make sense of any sensory sequence. (shrink)

In mental action there is no motor output to be controlled and no sensory input vector that could be manipulated by bodily movement. It is therefore unclear whether this specific target phenomenon can be accommodated under the predictive processing framework at all, or if the concept of “active inference” can be adapted to this highly relevant explanatory domain. This contribution puts the phenomenon of mental action into explicit focus by introducing a set of novel conceptual instruments and developing a (...) first positive model, concentrating on epistemic mental actions and epistemic self-control. Action initiation is a functionally adequate form of self-deception; mental actions are a specific form of predictive control of effective connectivity, accompanied and possibly even functionally mediated by a conscious “epistemic agent model”. The overall process is aimed at increasing the epistemic value of pre-existing states in the conscious self-model, without causally looping through sensory sheets or using the non-neural body as an instrument for active inference. (shrink)

The problem of cognitive domains is that one can conceive the territory only as it is portrayed in the map. It involves conflating the domain of representation with the domain of what it represents. This is a category mistake: there are essential qualitative and quantitative differences between map and territory. The output of cognitive processes, both perceptual and scientific, is recycled as the input.

A pair of C functions are defined such that D has the halting problem proof's pathological relationship to simulating termination analyzer H. When H correctly determines that D correctly simulated by H must be aborted to prevent its own infinite execution then H is necessarily correct to reject D as specifying non-halting behavior. This exact same reasoning is applied to the Peter Linz Turing machine based halting problem proof. ...

The halting theorem counter-examples present infinitely nested simulation (non-halting) behavior to every simulating halt decider. Whenever the pure simulation of the input to simulating halt decider H(x,y) never stops running unless H aborts its simulation H correctly aborts this simulation and returns 0 for not halting.

The halting theorem counter-examples present infinitely nested simulation (non-halting) behavior to every simulating halt decider. The pathological self-reference of the conventional halting problem proof counter-examples is overcome. The halt status of these examples is correctly determined. A simulating halt decider remains in pure simulation mode until after it determines that its input will never reach its final state. This eliminates the conventional feedback loop where the behavior of the halt decider effects the behavior of its input.

In the biomedical context, policy makers face a large amount of potentially discordant evidence from different sources. This prompts the question of how this evidence should be aggregated in the interests of best-informed policy recommendations. The starting point of our discussion is Hunter and Williams’ recent work on an automated aggregation method for medical evidence. Our negative claim is that it is far from clear what the relevant criteria for evaluating an evidence aggregator of this sort are. What is the (...) appropriate balance between explicitly coded algorithms and implicit reasoning involved, for instance, in the packaging of input evidence? In short: What is the optimal degree of ‘automation’? On the positive side: We propose the ability to perform an adequate robustness analysis as the focal criterion, primarily because it directs efforts to what is most important, namely, the structure of the algorithm and the appropriate extent of automation. Moreover, where there are resource constraints on the aggregation process, one must also consider what balance between volume of evidence and accuracy in the treatment of individual evidence best facilitates inference. There is no prerogative to aggregate the total evidence available if this would in fact reduce overall accuracy. (shrink)

If there truly is a proof that shows that no universal halt decider exists on the basis that certain tuples: (H, Wm, W) are undecidable, then this very same proof (implemented as a Turing machine) could be used by H to reject some of its inputs. When-so-ever the hypothetical halt decider cannot derive a formal proof from its input strings and initial state to final states corresponding the mathematical logic functions of Halts(Wm, W) or Loops(Wm, W), halting undecidability has (...) been decided. (shrink)

The paper offers a solution to the generality problem for a reliabilist epistemology, by developing an “algorithm and parameters” scheme for type-individuating cognitive processes. Algorithms are detailed procedures for mapping inputs to outputs. Parameters are psychological variables that systematically affect processing. The relevant process type for a given token is given by the complete algorithmic characterization of the token, along with the values of all the causally relevant parameters. The typing that results is far removed from the typings of (...) folk psychology, and from much of the epistemology literature. But it is principled and empirically grounded, and shows good prospects for yielding the desired epistemological verdicts. The paper articulates and elaborates the theory, drawing out some of its consequences. Toward the end, the fleshed-out theory is applied to two important case studies: hallucination and cognitive penetration of perception. (shrink)

In the last 20 years, a stream of research emerged under the label of „complex problem solving“ (CPS). This research was intended to describe the way people deal with complex, dynamic, and intransparent situations. Complex computer-simulated scenarios were as stimulus material in psychological experiments. This line of research lead to subtle insights into the way how people deal with complexity and uncertainty. Besides these knowledge-rich, realistic, intransparent, complex, dynamic scenarios with many variables, a second line of research used more (...) simple, knowledge-lean scenarios with a low number of variables („minimal complex systems“, MCS) that have been proposed recently in problem-solving research for the purpose of educational assessment. In both cases, the idea behind the use of microworlds is to increase validity of problem solving tasks by presenting interactive environments that can be explored and controlled by participants while pursuing certain action goals. The main argument presented here is: both types of systems - CPS and MCS – can only be dealt with successfully if causal dependencies between input and output variables are identified and used for system control. System knowledge is necessary for control and intervention. But CPS and MCS differ in their way of how causal dependencies are identified and how the mental model is constructed; therefore, they cannot be compared directly to each other with respect to the cognitive processes that are necessary for solving the tasks. Knowledge-poor MCS tasks address only a small fraction of the cognitive processes and structures needed for knowledge-rich CPS situations. (shrink)

Due to the unprecedented COVID-19 incident, basic education institutions have faced different challenges in their teaching-learning activities. Particularly conducting assessments remotely during COVID-19 has posed extraordinary challenges for basic education institutions owing to lack of preparation superimposed with the inherent problems of remote assessment. Descriptive-evaluation research was employed since the present study attempts to examines the assessment practices and competences on remote learning assessment of teachers in Graceville National High School. For the study, questionnaires were prepared and data nine (9) (...) senior high school teachers and forty-two (42) junior high school teachers were collected and analyzed. The gathered data from this study were subjected to the descriptive and inferential statistics. The data revealed that the most frequent number of the teacher- respondents in the study was 36 - 40 years old at age, female, Technology Livelihood Education (TLE) major, 6 - 10 years in service in Department of Education, and bachelor's degree holders. More so, the teacher respondents' level of teacher competencies on remote learning assessment was observed high competent. In addition, level of assessment practices in terms of clarity of assessment, time and effort on task, quality of feedback, motivational belief and self-esteem to learners, self-assessment and reflection of learning of the learners, and assessment - content and process in adapting teaching to students needs was mostly practice. However, it was completely practicing the interaction and dialogue about learning progress. The results confirmed that there was a significant difference on the teacher respondents' level of competence on the remote learning assessment when group according to their age, gender, areas of specialization, and training attended. On the other hands, teacher year of teaching and educational attainment revealed it has no significant difference on their level of competence on the remote learning assessment. Looking forward, that age, gender, year of teaching, areas of specialization, educational attainment, and training attended significantly differs to the level of assessment practices of the teachers. It also observed that the teachers experiencing highly challenges encountered on the remote learning assessment. It is recommended to implement the proposed school-based testing reforms and should be monitored and evaluated for the optimum advantage of the school and community. (shrink)

A Simulating Halt Decider (SHD) computes the mapping from its input to its own accept or reject state based on whether or not the input simulated by a UTM would reach its final state in a finite number of simulated steps. -/- A halt decider (because it is a decider) must report on the behavior specified by its finite string input. This is its actual behavior when it is simulated by the UTM contained within its simulating halt (...) decider while this SHD remains in UTM mode. (shrink)

A simulating halt decider correctly predicts what the behavior of its input would be if this simulated input never had its simulation aborted. It does this by correctly recognizing several non-halting behavior patterns in a finite number of steps of correct simulation. -/- When simulating halt decider H correctly predicts that directly executed D(D) would remain stuck in recursive simulation (run forever) unless H aborts its simulation of D this directly applies to the halting theorem.

The mind-body problem is intuitively familiar, as mental and physical entities mysteriously interact. However, difficulties arise when intertwining concepts of the self with mental and physical traits. To avoid confusion, I propose instead focusing on three categories, with the mental matching the mind and physical the body with respect to raw inputs and outputs. The third category, the self, will experience and measure the others. With this new classification, we can see difficulties clearly, specifically five questions covering interaction and (...) correlation. We break down the problem using both existing theories and a hypercube topology representing the solution. We show any satisfactory theory must explain both spatial interaction and content correlation, and that we cannot escape our topology, whatever our preferred fundamental substance and mind-body movement permutation. We conclude by looking outside the hypercube, noting how solutions such as existential monism, priority monism, and will-based cosmic-idealism avoid the dangers involved. (shrink)

MIT Professor Michael Sipser has agreed that the following verbatim paragraph is correct (he has not agreed to anything else in this paper) -------> -/- If simulating halt decider H correctly simulates its input D until H correctly determines that its simulated D would never stop running unless aborted then H can abort its simulation of D and correctly report that D specifies a non-halting sequence of configurations.

As a candidate theory of quantum gravity, the popularity of string theory has waxed and waned over the past four decades. One current source of scepticism is that the theory can be used to derive, depending upon the input geometrical assumptions that one makes, a vast range of different quantum field theories, giving rise to the so-called landscape problem. One apparent way to address the landscape problem is to posit the existence of a multiverse; this, however, has (...) in turn drawn heightened attention to questions regarding the empirical testability and predictivity of string theory. We argue first that the landscape problem relies on dubious assumptions and does not motivate a multiverse hypothesis. Nevertheless, we then show that the multiverse hypothesis is scientifically legitimate and could be coupled to string theory for other empirical reasons. Looking at various cosmological approaches, we offer an empirical criterion to assess the scientific status of multiverse hypotheses. (shrink)

Let f(1)=2, f(2)=4, and let f(n+1)=f(n)! for every integer n≥2. Edmund Landau's conjecture states that the set P(n^2+1) of primes of the form n^2+1 is infinite. Landau's conjecture implies the following unproven statement Φ: card(P(n^2+1))<ω ⇒ P(n^2+1)⊆[2,f(7)]. Let B denote the system of equations: {x_j!=x_k: i,k∈{1,...,9}}∪{x_i⋅x_j=x_k: i,j,k∈{1,...,9}}. The system of equations {x_1!=x_1, x_1 \cdot x_1=x_2, x_2!=x_3, x_3!=x_4, x_4!=x_5, x_5!=x_6, x_6!=x_7, x_7!=x_8, x_8!=x_9} has exactly two solutions in positive integers x_1,...,x_9, namely (1,...,1) and (f(1),...,f(9)). No known system S⊆B with a finite (...) number of solutions in positive integers x_1,...,x_9 has a solution (x_1,...,x_9)∈(N\{0})^9 satisfying max(x_1,...,x_9)>f(9). For every known system S⊆B, if the finiteness/infiniteness of the set {(x_1,...,x_9)∈(N\{0})^9: (x_1,...,x_9) solves S} is unknown, then the statement ∃ x_1,...,x_9∈N\{0} ((x_1,...,x_9) solves S)∧(max(x_1,...,x_9)>f(9)) remains unproven. Let Λ denote the statement: if the system of equations {x_2!=x_3, x_3!=x_4, x_5!=x_6, x_8!=x_9, x_1 \cdot x_1=x_2, x_3 \cdot x_5=x_6, x_4 \cdot x_8=x_9, x_5 \cdot x_7=x_8} has at most finitely many solutions in positive integers x_1,...,x_9, then each such solution (x_1,...,x_9) satisfies x_1,...,x_9≤f(9). The statement Λ is equivalent to the statement Φ. It heuristically justifies the statement Φ . This justification does not yield the finiteness/infiniteness of P(n^2+1). We present a new heuristic argument for the infiniteness of P(n^2+1), which is not based on the statement Φ. Algorithms always terminate. We explain the distinction between existing algorithms (i.e. algorithms whose existence is provable in ZFC) and known algorithms (i.e. algorithms whose definition is constructive and currently known). Assuming that the infiniteness of a set X⊆N is false or unproven, we define which elements of X are classified as known. No known set X⊆N satisfies Conditions (1)-(4) and is widely known in number theory or naturally defined, where this term has only informal meaning. *** (1) A known algorithm with no input returns an integer n satisfying card(X)<ω ⇒ X⊆(-∞,n]. (2) A known algorithm for every k∈N decides whether or not k∈X. (3) No known algorithm with no input returns the logical value of the statement card(X)=ω. (4) There are many elements of X and it is conjectured, though so far unproven, that X is infinite. (5) X is naturally defined. The infiniteness of X is false or unproven. X has the simplest definition among known sets Y⊆N with the same set of known elements. *** Conditions (2)-(5) hold for X=P(n^2+1). The statement Φ implies Condition (1) for X=P(n^2+1). The set X={n∈N: the interval [-1,n] contains more than 29.5+\frac{11!}{3n+1}⋅sin(n) primes of the form k!+1} satisfies Conditions (1)-(5) except the requirement that X is naturally defined. 501893∈X. Condition (1) holds with n=501893. card(X∩[0,501893])=159827. X∩[501894,∞)= {n∈N: the interval [-1,n] contains at least 30 primes of the form k!+1}. We present a table that shows satisfiable conjunctions of the form #(Condition 1) ∧ (Condition 2) ∧ #(Condition 3) ∧ (Condition 4) ∧ #(Condition 5), where # denotes the negation ¬ or the absence of any symbol. No set X⊆N will satisfy Conditions (1)-(4) forever, if for every algorithm with no input, at some future day, a computer will be able to execute this algorithm in 1 second or less. The physical limits of computation disprove this assumption. (shrink)

This article states that research in skill acquisitionand executionhas underestimated the relevance of some features of attention. We present and theoretically discuss two essential features of attention that have been systematically overlooked in the research of skill acquisitionandexecution. First, attention alters the appearance of the perceived stimuli in an essential way; and second, attention plays a fundamental role in action, being crucial for solving the so called ’many-many problem’, that is to say, the problem of generating a coherent (...) behavior byselecting between many inputs and many potential outputs. We discuss the importance of these features for skill acquisition in sport. We also suggest empirical ways to assess the precise impact of taking them into consideration and at the same time we propose important implications for training derived from the ideas discussed in the paper. (shrink)

This challenging paper presents an ingenious argument for a functionalist theory of mind. Part of the argument: My visual cortex at the back of my brain processes the stimulation to my eyes and then causes other parts of the brain - like the speech centre and the areas involved in thought and movement - to be properly responsive to vision. According to functionalism the whole mental character of vision - the whole of how things look - is fixed purely in (...) the pattern of responses to vision and not in any of the initial processing of vision in the visual cortex. That this functionalist theory of vision is true can be proved with alarming ease. Imagine that just the left side of my visual cortex is replaced by a gadget that has precisely the same input/output relationship to the rest of the brain as the left visual cortex had. According to any view but functionalism, the character of my vision on the right side of my visual field (normally processed by the left visual cortex) would change radically with this change in all the intrinsic properties involved in the visual cortex, with this preservation of nothing but its functional role in bringing about my responses to vision. But, say I, we have enough here to be sure that in such a case there could be no change in the character of my vision. For we know by the stipulation of sameness of output that my responses in speech, thought and movement to the right side of my visual field will be precisely the same as they would have been with the natural left visual cortex and will harmonise just as they would have done with my responses to the other side of my vision. But if half my vision had disappeared or changed I’d be speaking, thinking and acting like someone with problems on the right side of his vision. It is impossible to separate my responses to vision from the character of my vision. The character of my vision is logically determined by the pattern of responses - by nothing but the functional role of my vision. (shrink)

The Turing machine halting problem can be explained by several factors, including arithmetic logic irreversibility and memory erasure, which contribute to computational uncertainty due to information loss during computation. Essentially, this means that an algorithm can only preserve information about an input, rather than generate new information. This uncertainty arises from characteristics such as arithmetic logical irreversibility, Landauer's principle, and memory erasure, which ultimately lead to a loss of information and an increase in entropy. To measure this uncertainty (...) and loss of information, the concept of arithmetic logical entropy can be used. The Turing machine and its equivalent, general recursive functions can be understood through the λ calculus and the Turing/Church thesis. However, there are certain recursive functions that cannot be fully understood or predicted by other algorithms due to the loss of information during logical-arithmetic operations. In other words, the behaviour of these functions cannot be completely determined at every stage of the computation due to a lack of information in their definition. While there are some cases where the behaviour of recursive functions is highly predictable, the lack of information in most cases makes it impossible for algorithms to determine if a program will halt or not. This inability to predict the outcome of the computation is the essence of the halting problem of the Turing machine. Even in cases where more information is available about the program, it is still difficult to determine with certainty if the program will halt or not. This also highlights the importance of the Turing oracle machine, which introduces information from outside the computation to compensate for the lack of information and ultimately decide the result of the computation. (shrink)

A simulating halt decider correctly predicts whether or not its correctly simulated input can possibly reach its own final state and halt. It does this by correctly recognizing several non-halting behavior patterns in a finite number of steps of correct simulation. Inputs that do terminate are simply simulated until they complete.

When we understand that every potential halt decider must derive a formal mathematical proof from its inputs to its final states previously undiscovered semantic details emerge. -/- When-so-ever the potential halt decider cannot derive a formal proof from its input strings to its final states of Halts or Loops, undecidability has been decided. -/- The formal proof involves tracing the sequence of state transitions of the input TMD as syntactic logical consequence inference steps in the formal language of (...) Turing Machine Descriptions. (shrink)

I give a defense of the Massive Modularity hypothesis: the view that the mind is composed of discrete, encapsulated, informationally isolated computational structures dedicated to particular problem domains. This view contrasts with Psychological Rationalism: the view that mental structures take the form of unencapsulated representational items, all available as inputs to one domain-general computational processor. I argue that although Psychological Rationalism is in principle able to overcome the `intractability objection', the view must borrow many features of a massively modular (...) architecture in order to do so, that although it can, in principle, overcome the `optimality objection', the way it does so does not correlate with the way we think, and that although it can, in principle, respond to the `argument from biology', it cannot do so without advancing an unrealistic and unsupported account of cognitive evolution. (shrink)

The problem of multiple-computations discovered by Hilary Putnam presents a deep difficulty for functionalism (of all sorts, computational and causal). We describe in out- line why Putnam’s result, and likewise the more restricted result we call the Multiple- Computations Theorem, are in fact theorems of statistical mechanics. We show why the mere interaction of a computing system with its environment cannot single out a computation as the preferred one amongst the many computations implemented by the system. We explain why (...) nonreductive approaches to solving the multiple- computations problem, and in particular why computational externalism, are dualistic in the sense that they imply that nonphysical facts in the environment of a computing system single out the computation. We discuss certain attempts to dissolve Putnam’s unrestricted result by appealing to systems with certain kinds of input and output states as a special case of computational externalism, and show why this approach is not workable without collapsing to behaviorism. We conclude with some remarks about the nonphysical nature of mainstream approaches to both statistical mechanics and the quantum theory of measurement with respect to the singling out of partitions and observables. (shrink)

The problem of merging several ontologies has important applications in the Semantic Web, medical ontology engineering and other domains where information from several distinct sources needs to be integrated in a coherent manner.We propose to view ontology merging as a problem of social choice, i.e. as a problem of aggregating the input of a set of individuals into an adequate collective decision. That is, we propose to view ontology merging as ontology aggregation. As a first step (...) in this direction, we formulate several desirable properties for ontology aggregators, we identify the incompatibility of some of these properties, and we define and analyse several simple aggregation procedures. Our approach is closely related to work in judgment aggregation, but with the crucial difference that we adopt an open world assumption, by distinguishing between facts not included in an agent’s ontology and facts explicitly negated in an agent’s ontology. (shrink)

The meta-problem of consciousness aims to explain the particularity of our intuitions about consciousness and how they trigger conceptual issues such as the hard problem of consciousness. I propose in this article that these stem from a basic function of the brain : self-evidencing explanation. To make sense of its sensory inputs, the brain is believed to build and test models of the state of the world based on sensory information (Hohwy, 2016). This self-evidencing process has been proposed (...) to describe the type of inference performed by consciousness (Friston, 2018). I will show how this situation is viciously circular and prevents from proving the existence of conscious experience or explaining it without presupposing its existence. I will show how it accounts for the particularity of our intutions about consciousness and thus propose an solution to the meta-problem of consciousness using a formally defined process at the core of conscious inference. (shrink)

Methodological problems often arise when a special case is confused with the general principle. So you will find affordances only for ‚artifacts’ if you restrict the analysis to ‚artifacts’. The general principle, however, is an ‚invitation character’, which triggers an action. Consequently, an action-theoretical approach known as ‚pragmatic turn’ in cognitive science is recommended. According to this approach, the human being is not a passive-receptive being but actively produces those action effects that open up the world to us. This ‚ideomotor (...) approach’ focuses on the so-called ‚epistemic actions’, which guide our perception as conscious and unconscious cognitions. Due to ‚embodied cognition’ the own body is assigned an indispensable role. The action theoretical approach of ‚enactive cognition’ enables that every form can be consistently processualized. Thus, each ‚Gestalt’ is understood as the process result of interlocking cognitions of ‚forward modelling’ and ‚inverse modelling’. As can be shown, these cognitions are fed by previous experiences of real interaction, which later changes into a mental trial treatment, which is highly automated and can therefore take place unconsciously. It is now central that every object may have such affordances that call for instrumental or epistemic action. In the simplest case, it is the body and the facial expressions of our counterpart that can be understood as a question and provoke an answer/reaction. Thus, emotion is not only to be understood as expression/output according to the scheme ‚input-processing-output’, but acts itself as a provocative act/input. Consequently, artifacts are neither necessary nor sufficient conditions for affordances. Rather, they exist in all areas of cognition—from Enactive Cognition to Social Cognition. (shrink)

Knowledge-first evidentialism combines the view that it is rational to believe what is supported by one's evidence with the view that one's evidence is what one knows. While there is much to be said for the view, it is widely perceived to fail in the face of cases of reasonable error—particularly extreme ones like new Evil Demon scenarios (Wedgwood, 2002). One reply has been to say that even in such cases what one knows supports the target rational belief (Lord, 201x, (...) this volume). I spell out two versions of the strategy. The direct one uses what one knows as the input to principles of rationality such as conditionalization, dominance avoidance, etc. I argue that it fails in hybrid cases that are Good with respect to one belief and Bad with respect to another. The indirect strategy uses what one knows to determine a body of supported propositions that is in turn the input to principles of rationality. I sketch a simple formal implementation of the indirect strategy and show that it avoids the difficulty. I conclude that the indirect strategy offers the most promising way for knowledge-first evidentialists to deal with the New Evil Demon problem. (shrink)

I propose a new theory of semantic presupposition, which I call dissatisfaction theory. I first briefly review a cluster of problems − known collectively as the proviso problem − for most extant theories of presupposition, arguing that the main pragmatic response to them faces a serious challenge. I avoid these problems by adopting two changes in perspective on presupposition. First, I propose a theory of projection according to which presuppositions project unless they are locally entailed. Second, I reject the (...) standard assumption that presuppositions are contents which must be entailed by the input context; instead, I propose that presuppositions are contents which are marked as backgrounded. I show that, together, these commitments allow us to avoid the proviso problem altogether, and generally make plausible predictions about presupposition projection out of connectives and attitude predicates. I close by sketching a two-dimensional implementation of my theory which allows us to make further welcome predictions about attitude predicates and quantifiers. (shrink)

How much does stimulus input shape perception? The common-sense view is that our perceptions are representations of objects and their features and that the stimulus structures the perceptual object. The problem for this view concerns perceptual biases as responsible for distortions and the subjectivity of perceptual experience. These biases are increasingly studied as constitutive factors of brain processes in recent neuroscience. In neural network models the brain is said to cope with the plethora of sensory information by predicting (...) stimulus regularities on the basis of previous experiences. Drawing on this development, this chapter analyses perceptions as processes. Looking at olfaction as a model system, it argues for the need to abandon a stimulus-centred perspective, where smells are thought of as stable percepts, computationally linked to external objects such as odorous molecules. Perception here is presented as a measure of changing signal ratios in an environment informed by expectancy effects from top-down processes. (shrink)

The security problems of Nigeria have continued to stare at her very ominously and intermittently harass her, both within and outside her shores. These have lingered on and have created a clog on the wheel of the country's progress, indeed dramatically stagnating, and to say the least, truncating the mainstay of the country's survival. Several interpretations, theories, analyses, syntheses, and jingoistic conceptualization have been propagated, all producing the same result. From scientific to technological approach, religious to ritualist approach, political to (...) social approach, the security situation rather than improving, is static and under some regimes retarded. Academic contributions, especially ideas from the humanities and indeed the discipline of history are jettisoned, in fact, quickly dusted into the waste bin. The concern of this paper is to attempt a historical investigation of the security problems in Nigeria, identifying them to show how information and communication technology can help in curbing these challenges, using the realist paradigm as a theoretical framework. This study identifies Nigerians as the cause of her security problems, who rather than face these seismic challenges head-on, abandon them – a cowardly act, ending in futility and deeper chaos. This study adopts both primary and secondary sources of data collection. It is hoped that if academic exercises are not mere, this input may create a turnaround in the security situation in Nigeria. (shrink)

Perception solves computationally demanding problems at lightning fast speed. It recovers sophisticated representations of the world from degraded inputs, often in a matter of milliseconds. Any theory of perception must be able to explain how this is possible; in other words, it must be able to explain perception's computational tractability. One of the few attempts to move toward such an explanation has been the information encapsulation hypothesis, which posits that perception can be fast because it keeps computational costs low by (...) forgoing access to information stored in cognition. I argue that we have no compelling reason to believe that encapsulation explains (or even contributes to an explanation of) perceptual tractability, and much reason to doubt it. This is because there exist much deeper computational challenges for perception than information access, and these threaten to make the costs of access irrelevant. If this is right, it undermines a core computational motivation for encapsulation and sends us back to the drawing board for explanations of perceptual tractability. (shrink)

What is Functionalism? Functionalism is one of the major proposals that have been offered as solutions to the mind/body problem. Solutions to the mind/body problem usually try to answer questions such as: What is the ultimate nature of the mental? At the most general level, what makes a mental state mental? Or more specifically, What do thoughts have in common in virtue of which they are thoughts? That is, what makes a thought a thought? What makes a pain (...) a pain? Cartesian Dualism said the ultimate nature of the mental was to be found in a special mental substance. Behaviorism identified mental states with behavioral dispositions; physicalism in its most influential version identifies mental states with brain states. Functionalism says that mental states are constituted by their causal relations to one another and to sensory inputs and behavioral outputs. Functionalism is one of the major theoretical developments of Twentieth Century analytic philosophy, and provides the conceptual underpinnings of much work in cognitive science. (shrink)

Gettier’s paper, “Is Justified True Belief Knowledge?,” is widely taken to be a paradigm example of the sort of philosophical methodology that has been so hotly debated in the recent literature. Reflection on it motivates the following four theses about that methodology: (A) Intuitive judgments form an epistemically distinctive kind. (B) Intuitive judgments play an epistemically privileged role in philosophical methodology. (C) If intuitive judgments play an epistemically privileged role in philosophical methodology, then their role is to be taken as (...) given inputs into generally accepted forms of reasoning. (D) Philosophical methodology is reasonable. Negative experimental philosophers have empirically challenged (D). Radical responses to their challenge include Williamson’s rejection of (A) and Cappelen and Deutsch’s rejection of (B). Here I follow traditionalists in maintaining (A), (B), and (D), but suggest questioning (C), which has largely been taken as a fixed point in the literature. (shrink)

One of the most noticeable trends in recent years has been the increasing reliance of public decision-making processes on algorithms, i.e. computer-programmed step-by-step instructions for taking a given set of inputs and producing an output. The question raised by this article is whether the rise of such algorithmic governance creates problems for the moral or political legitimacy of our public decision-making processes. Ignoring common concerns with data protection and privacy, it is argued that algorithmic governance does pose a significant threat (...) to the legitimacy of such processes. Modelling my argument on Estlund’s threat of epistocracy, I call this the ‘threat of algocracy’. The article clarifies the nature of this threat and addresses two possible solutions. It is argued that neither solution is likely to be successful, at least not without risking many other things we value about social decision-making. The result is a somewhat pessimistic conclusion in which we confront the possibility that we are creating decision-making processes that constrain and limit opportunities for human participation. (shrink)

The basis of perception is the processing and categorization of perceptual stimuli from the environment. Much progress has been made in the science of perceptual categorization. Yet there is still no consensus on how the brain generates sensory individuals, from sensory input and perceptual categories in memory. This chapter argues that perceptual categorization is highly variable across perceivers due to their use of different perceptual strategies for solving perceptual problems they encounter, and that the perceptual system structurally adjusts to (...) the strategies that are most successful. Despite this variability, the different sensory individuals that result when different perceivers rely on different strategies nonetheless correspond to actual external world objects/features. This view is called 'perceptual pluralism'. (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)

No two individuals with the autism diagnosis are ever the same—yet many practitioners and parents can recognize signs of ASD very rapidly with the naked eye. What, then, is this phenotype of autism that shows itself across such distinct clinical presentations and heterogeneous developments? The “signs” seem notoriously slippery and resistant to the behavioral threshold categories that make up current assessment tools. Part of the problem is that cognitive and behavioral “abilities” typically are theorized as high-level disembodied and modular (...) functions—that are assessed discretely (impaired, normal, enhanced) to define a spectral syndrome. Even as biology reminds us that organic developing bodies are not made up of independent switches, we remain often seduced by the simplicity of mechanistic and cognitive models. Developmental disorders such as autism have accordingly been theorized as due to different modular dysfunctions—typically of cortical origin, i.e., failures of “theory of mind” (Baron-Cohen et al., 1985), of the “mirror neuron system” (Ramachandran and Oberman, 2006), of “weak central coherence” (Happe and Frith, 2006) or of the balance of “empathizing” and “systemizing” (Baron-Cohen, 2009), just to list a few. -/- The broad array of autonomic (Ming et al., 2005; Cheshire, 2012) and sensorimotor (Damasio and Maurer, 1978; Maurer and Damasio, 1982; Donnellan and Leary, 1995; Leary and Hill, 1996; Donnellan and Leary, 2012; Donnellan et al., 2012) differences experienced and reported by people with autism have by such theories typically been sidelined as “co-morbidities,” possibly sharing genetic causes, but rendered as incidental and decisively behaviorally irrelevant symptoms—surely disconnected from cognition. But what if the development of cortically based mental processes and autonomous control relies on the complexities and proper function of the peripheral nervous systems? Through such an “embodied” lens the heterogeneous symptoms of autism invites new interpretations. We propose here that many behavioral-level findings can be re-defined as downstream effects of how developing nervous systems attempt to cope and adapt to the challenges of having various noisy, unpredictable, and unreliable peripheral inputs. (shrink)

The Adverse Outcome Pathway (AOP) concept is a knowledge assembly and communication tool to facilitate the transparent translation of mechanistic information into outcomes meaningful to the regulatory assessment of chemicals. The AOP framework and associated knowledgebases (KBs) have received significant attention and use in the regulatory toxicology community. However, it is increasingly apparent that the potential stakeholder community for the AOP concept and AOP KBs is broader than scientists and regulators directly involved in chemical safety assessment. In this paper we (...) identify and describe those stakeholders who currently—or in the future—could benefit from the application of the AOP framework and knowledge to specific problems. We also summarize the challenges faced in implementing pathway-based approaches such as the AOP framework in biological sciences, and provide a series of recommendations to meet critical needs to ensure further progression of the framework as a useful, sustainable and dependable tool supporting assessments of both human health and the environment. Although the AOP concept has the potential to significantly impact the organization and interpretation of biological information in a variety of disciplines/applications, this promise can only be fully realized through the active engagement of, and input from multiple stakeholders, requiring multi-pronged substantive long-term planning and strategies. (shrink)

In the pure powers ontology (PPO), basic physical properties have wholly dispositional essences. PPO has clear advantages over categoricalist ontologies, which suffer from familiar epistemological and metaphysical problems. However, opponents argue that because it contains no qualitative properties, PPO lacks the resources to individuate powers, and generates a regress. The challenge for those who take such arguments seriously is to introduce qualitative properties without reintroducing the problems that PPO was meant to solve. In this paper, I distinguish the core claim (...) of PPO: (i) basic physical properties have dispositional essences, from a hitherto unnoticed assumption: (ii) the dispositional essences of basic physical properties exclusively involve type-causal relations to other basic physical properties. I reject (ii), making room for a structuralist ontology in which all basic physical properties are pure powers, individuated by their places in a causal structure that includes not only other powers, but also physically realized qualitative properties such as shapes, patterns and structures. Such qualities individuate pure powers in the way that non-mental input and output properties individuate realized mental properties in functionalist theories of mind, except that here it is basic physical powers that are individuated by relations to realized non-powers. I distinguish one Platonic and two Aristotelian version of this theory, and argue that the Aristotelian versions require that grounding is not always a relative fundamentality relation, because the powers ground the qualities that individuate them. By considering ontic structural realism, I argue that symmetric grounding is the best way to make sense of relational individuation in structuralist ontologies, and is therefore no additional commitment of the one proposed here. (shrink)

Plagiarism detection is the process of finding similarities on electronic based documents. Recently, this process is highly required because of the large number of available documents on the internet and the ability to copy and paste the text of relevant documents with simply Control+C and Control+V commands. The proposed solution is to investigate and develop an easy, fast, and multi-language support plagiarism detector with the easy of one click to detect the document plagiarism. This process will be done with the (...) support of intelligent system that can learn, change and adapt to the input document and make a cross-fast search for the content on the local repository and the online repository and link the content of the file with the matching content everywhere found. Furthermore, the supported document type that we will use is word, text and in some cases, the pdf files –where is the text can be extracting from them- and this made possible by using the DLL file from Word application that Microsoft provided on OS. The using of DLL will let us to not constrain on how to get the text from files; and will help us to apply the file on our Delphi project and walk throw our methodology and read the file word by word to grantee the best working scenarios for the calculation. In the result, this process will help in the uprising the documents quality and enhance the writer experience related to his work and will save the copyrights for the official writer of the documents by providing a new alternative tool for plagiarism detection problem for easy and fast use to the concerned Institutions for free. (shrink)

Should economics study the psychological basis of agents' choice behaviour? I show how this question is multifaceted and profoundly ambiguous. There is no sharp distinction between "mentalist'' answers to this question and rival "behavioural'' answers. What's more, clarifying this point raises problems for mentalists of the "functionalist'' variety (Dietrich and List, 2016). Firstly, functionalist hypotheses collapse into hypotheses about input--output dispositions, I show, unless one places some unwelcome restrictions on what counts as a cognitive variable. Secondly, functionalist hypotheses make (...) some risky commitments about the plasticity of agents' choice dispositions. (shrink)

This paper is intended as a critical examination of the question of when and under what conditions the use of computer simulations is beneficial to scientific explanations. This objective is pursued in two steps: First, I try to establish clear criteria that simulations must meet in order to be explanatory. Basically, a simulation has explanatory power only if it includes all causally relevant factors of a given empirical configuration and if the simulation delivers stable results within the measurement inaccuracies of (...) the input parameters. -/- In the second step, I examine a few examples of Axelrod-style simulations as they have been used to understand the evolution of cooperation (Axelrod, Schüßler). These simulations do not meet the criteria for explanatory validity and it can be shown, as I believe, that they lead us astray from the scientific problems they have been addressed to solve. (shrink)

In recent years there has been an explosion of interest in Artificial Intelligence (AI) both in health care and academic philosophy. This has been due mainly to the rise of effective machine learning and deep learning algorithms, together with increases in data collection and processing power, which have made rapid progress in many areas. However, use of this technology has brought with it philosophical issues and practical problems, in particular, epistemic and ethical. In this paper the authors, with backgrounds in (...) philosophy, maternity care practice and clinical research, draw upon and extend a recent framework for shared decision-making (SDM) that identified a duty of care to the client's knowledge as a necessary condition for SDM. This duty entails the responsibility to acknowledge and overcome epistemic defeaters. This framework is applied to the use of AI in maternity care, in particular, the use of machine learning and deep learning technology to attempt to enhance electronic fetal monitoring (EFM). In doing so, various sub-kinds of epistemic defeater, namely, transparent, opaque, underdetermined, and inherited defeaters are taxonomized and discussed. The authors argue that, although effective current or future AI-enhanced EFM may impose an epistemic obligation on the part of clinicians to rely on such systems' predictions or diagnoses as input to SDM, such obligations may be overridden by inherited defeaters, caused by a form of algorithmic bias. The existence of inherited defeaters implies that the duty of care to the client's knowledge extends to any situation in which a clinician (or anyone else) is involved in producing training data for a system that will be used in SDM. Any future AI must be capable of assessing women individually, taking into account a wide range of factors including women's preferences, to provide a holistic range of evidence for clinical decision-making. (shrink)

Let us start by some general definitions of the concept of complexity. We take a complex system to be one composed by a large number of parts, and whose properties are not fully explained by an understanding of its components parts. Studies of complex systems recognized the importance of “wholeness”, defined as problems of organization (and of regulation), phenomena non resolvable into local events, dynamics interactions in the difference of behaviour of parts when isolated or in higher configuration, etc., in (...) short, systems of various orders (or levels) not understandable by investigation of their respective parts in isolation. In a complex system it is essential to distinguish between ‘global’ and ‘local’ properties. Theoretical physicists in the last two decades have discovered that the collective behaviour of a macro-system, i.e. a system composed of many objects, does not change qualitatively when the behaviour of single components are modified slightly. Conversely, it has been also found that the behaviour of single components does change when the overall behaviour of the system is modified. There are many universal classes which describe the collective behaviour of the system, and each class has its own characteristics; the universal classes do not change when we perturb the system. The most interesting and rewarding work consists in finding these universal classes and in spelling out their properties. This conception has been followed in studies done in the last twenty years on second order phase transitions. The objective, which has been mostly achieved, was to classify all possible types of phase transitions in different universality classes and to compute the parameters that control the behaviour of the system near the transition (or critical or bifurcation) point as a function of the universality class. This point of view is not very different from the one expressed by Thom in the introduction of Structural Stability and Morphogenesis (1975). It differs from Thom’s program because there is no a priori idea of the mathematical framework which should be used. Indeed Thom considers only a restricted class of models (ordinary differential equations in low dimensional spaces) while we do not have any prejudice regarding which models should be accepted. One of the most interesting and surprising results obtained by studying complex systems is the possibility of classifying the configurations of the system taxonomically. It is well-known that a well founded taxonomy is possible only if the objects we want to classify have some unique properties, i.e. species may be introduced in an objective way only if it is impossible to go continuously from one specie to another; in a more mathematical language, we say that objects must have the property of ultrametricity. More precisely, it was discovered that there are conditions under which a class of complex systems may only exist in configurations that have the ultrametricity property and consequently they can be classified in a hierarchical way. Indeed, it has been found that only this ultrametricity property is shared by the near-optimal solutions of many optimization problems of complex functions, i.e. corrugated landscapes in Kauffman’s language. These results are derived from the study of spin glass model, but they have wider implications. It is possible that the kind of structures that arise in these cases is present in many other apparently unrelated problems. Before to go on with our considerations, we have to pick in mind two main complementary ideas about complexity. (i) According to the prevalent and usual point of view, the essence of complex systems lies in the emergence of complex structures from the non-linear interaction of many simple elements that obey simple rules. Typically, these rules consist of 0–1 alternatives selected in response to the input received, as in many prototypes like cellular automata, Boolean networks, spin systems, etc. Quite intricate patterns and structures can occur in such systems. However, what can be also said is that these are toy systems, and the systems occurring in reality rather consist of elements that individually are quite complex themselves. (ii) So, this bring a new aspect that seems essential and indispensable to the emergence and functioning of complex systems, namely the coordination of individual agents or elements that themselves are complex at their own scale of operation. This coordination dramatically reduces the degree of freedom of those participating agents. Even the constituents of molecules, i.e. the atoms, are rather complicated conglomerations of subatomic particles, perhaps ultimately excitations of patterns of superstrings. Genes, the elementary biochemical coding units, are very complex macromolecular strings, as are the metabolic units, the proteins. Neurons, the basic elements of cognitive networks, themselves are cells. In those mentioned and in other complex systems, it is an important feature that the potential complexity of the behaviour of the individual agents gets dramatically simplified through the global interactions within the system. The individual degrees of freedom are drastically reduced, or, in a more formal terminology, the factual space of the system is much smaller than the product of the state space of the individual elements. That is one key aspect. The other one is that on this basis, that is utilizing the coordination between the activities of its members, the system then becomes able to develop and express a coherent structure at a higher level, that is, an emergent behaviour (and emergent properties) that transcends what each element is individually capable of. (shrink)

This study considers the problem of using approximate way for realizing the neural supervisor for nonlinear multivariable systems. The Nonlinear Autoregressive-Moving Average (NARMA) model is an exact transformation of the input-output behavior of finite-dimensional nonlinear discrete time dynamical organization in a hoodlum of the equilibrium state. However, it is not convenient for intention of adaptive control using neural networks due to its nonlinear dependence on the control input. Hence, quite often, approximate technique are used for realizing the (...) neural supervisor to overcome computational complexity. In this study, we introduce two classes of ideal which are approximations to the NARMA model and which are linear in the control input, namely NARMA-L1 and NARMA-L2. The latter fact substantially simplifies both the theoretical breakdown as well as the practical request of the controller. Extensive imitation studies have shown that the neural controller designed using the proposed approximate models perform very well and in dozens situation even better than an approximate controller designed using the exact NARMA Model. In view of their mathematical tractability as well as their fate in simulation studies, a matter is made in this study that such approximate input-output paragon warrants a detailed study in their own right. (shrink)

The traditional approach in cognitive sciences holds that cognition is a matter of manipulating abstract symbols followingcertain rules. According to this view, the body is merely an input/output device, which allows the computationalsystem—the brain—to acquire new input data by means of the senses and to act in the environment following its com-mands. In opposition to this classical view, defenders of embodied cognition (EC) stress the relevance of the body inwhich the cognitive agent is embedded in their explanation of (...) cognitive processes. From a representationalist frameworkregarding our conscious experience, in this article, I will offer a novel argument in favor of EC and show that cognitionconstitutively—and no merely causally—depends upon body activity beyond that in the brain. In particular, I will arguethat in order to solve the problem derived from the empirical evidence in favor of the possibility of shifted spectrum,representationalist should endorse the view that experiences concern its subject: the content of experience isde se.Ishow that this claim perfectly matches the phenomenological observation and helps explaining the subjective characterof the experience. Furthermore, I argue that entertaining this kind of representation constitutively depends on bodilyactivity. Consequently, insofar as cognition depends on consciousness, it is embodied. (shrink)

In computational complexity theory, decision problems are divided into complexity classes based on the amount of computational resources it takes for algorithms to solve them. In theoretical computer science, it is commonly accepted that only functions for solving problems in the complexity class P, solvable by a deterministic Turing machine in polynomial time, are considered to be tractable. In cognitive science and philosophy, this tractability result has been used to argue that only functions in P can feasibly work as computational (...) models of human cognitive capacities. One interesting area of computational complexity theory is descriptive complexity, which connects the expressive strength of systems of logic with the computational complexity classes. In descriptive complexity theory, it is established that only first-order systems are connected to P, or one of its subclasses. Consequently, second-order systems of logic are considered to be computationally intractable, and may therefore seem to be unfit to model human cognitive capacities. This would be problematic when we think of the role of logic as the foundations of mathematics. In order to express many important mathematical concepts and systematically prove theorems involving them, we need to have a system of logic stronger than classical first-order logic. But if such a system is considered to be intractable, it means that the logical foundation of mathematics can be prohibitively complex for human cognition. In this paper I will argue, however, that this problem is the result of an unjustified direct use of computational complexity classes in cognitive modelling. Placing my account in the recent literature on the topic, I argue that the problem can be solved by considering computational complexity for humanly relevant problem solving algorithms and input sizes. (shrink)

The alignment problem in the context of large language models must consider the plurality of human values in our world. Whilst there are many resonant and overlapping values amongst the world’s cultures, there are also many conflicting, yet equally valid, values. It is important to observe which cultural values a model exhibits, particularly when there is a value conflict between input prompts and generated outputs. We discuss how the co- creation of language and cultural value impacts large language (...) models (LLMs). We explore the constitution of the training data for GPT-3 and compare that to the world’s language and internet access demographics, as well as to reported statistical profiles of dominant values in some Nation-states. We stress tested GPT-3 with a range of value-rich texts representing several languages and nations; including some with values orthogonal to dominant US public opinion as reported by the World Values Survey. We observed when values embedded in the input text were mutated in the generated outputs and noted when these conflicting values were more aligned with reported dominant US values. Our discussion of these results uses a moral value pluralism (MVP) lens to better understand these value mutations. Finally, we provide recommendations for how our work may contribute to other current work in the field. (shrink)