Philosophy of Computing and Information

Issues of reliability are claiming center-stage in the epistemology of machine learning. This paper unifies different branches in the literature and points to promising research directions, whilst also providing an accessible introduction to key concepts in statistics and machine learning---as far as they are concerned with reliability.

Mới đây, chương trình máy tính ‘bayesvl’ chạy trên môi trường R do TS. Vương Quân Hoàng và kĩ sư Lã Việt Phương (Trung tâm Nghiên cứu Xã hội Liên ngành ISR, Đại học Phenikaa) thiết kế và phát triển chính thức được ra mắt trên CRAN - hệ thống thư viện chuẩn của R...

Trong bối cảnh các hoạt động kinh tế và xã hội ngày càng được kết nối tốt hơn thông qua IoT, và sắp tới đây là tiềm năng tích hợp AI vào hầu như mọi mặt của đời sống ở cả thế giới thực và thế giới ảo, thì không chỉ cá nhân, mà cả doanh nghiệp và quốc gia cũng sẽ phải đối mặt với các thách thức chưa từng có tiền lệ đối với rủi ro về an ninh (...) thông tin... (shrink)

This paper examines a looming reproducibility crisis in the core of the hard sciences. Namely, it concentrates on molecular modeling and simulation (MMS), a family of methods that predict properties of substances through computing interactions on a molecular level and that is widely popular in physics, chemistry, materials science, and engineering. The paper argues that in order to make quantitative predictions, sophisticated models are needed which have to be evaluated with complex simulation procedures that amalgamate theoretical, technological, and social factors (...) – leading to problems with reproducibility. Thus, for methodological reasons, the predictive success causes a reproducibility problem. (shrink)

In contemporary healthcare, medical image categorization is essential for illness prediction, diagnosis, and therapy planning. The emergence of digital imaging technology has led to a significant increase in research into the use of machine learning (ML) techniques for the categorization of images in medical data. We provide a thorough summary of recent developments in this area in this review, using knowledge from the most recent research and cutting-edge methods.We begin by discussing the unique challenges and opportunities associated with medical image (...) classification, including the complexity of anatomical structures, variability in imaging modalities, and the need for interpretability and reliability in clinical settings. Subsequently, we survey a wide range of ML algorithms and techniques employed for medical image classification, including traditional methods such as support vector machines and k-nearest neighbors, as well as deep learning approaches like convolutional neural networks (CNNs) and recurrent neural networks (RNNs). Furthermore, we examine key considerations in dataset preparation, feature extraction, and model evaluation specific to medical image classification tasks. We highlight the importance of large, annotated datasets, transfer learning, and data augmentation techniques in enhancing model performance and generalization. (shrink)

Integrating medical imaging with computing technologies, such as Artificial Intelligence (AI) and its subsets: Machine learning (ML) and Deep Learning (DL) has advanced into an essential facet of present-day medicine, signaling a pivotal role in diagnostic decision-making and treatment plans (Huang et al., 2023). The significance of medical imaging is escalated by its sustained growth within the realm of modern healthcare (Varoquaux and Cheplygina, 2022). Nevertheless, the ever-increasing volume of medical images compared to the availability of imaging experts. Biomedical experts (...) and radiologists have resulted in a widening disparity, causing an excess and overwhelming workload on these healthcare professionals (Chen et al., 2021). Several studies indicate that the present-day biomedical radiologist is now saddled with the daunting task of interpreting an image almost every 10 seconds to keep pace with the burgeoning clinical demands (McDonald et al., 2015; Hosny et al., 2018; Lantsman et al., 2022). This cognitive drain has invariably led to inevitable consequences such as delays in diagnosis and an amplified risk of diagnostic errors – thus, the biomedical imaging aspect is in dire need of methods that would aid accurate diagnostics and analytics for improved decision making. In this review, the importance of AI-related technologies such as ML and/or DL methods are reviewed in relation to the processing of medical or biomedical images along with their potentials, challenges, and possible suggestions for future studies in the health landscape. The focus will be on machine learning methods associated with the medical field of image classification. (shrink)

Text classification is one of the most important task in natural language processing, In this research, we carried out several experimental research on three (3) of the most popular Text classification NLP classifier in Convolutional Neural Network (CNN), Multinomial Naive Bayes (MNB), and Support Vector Machine (SVN). In the presence of enough training data, Deep Learning CNN work best in all parameters for evaluation with 77% accuracy, followed by SVM with accuracy of 76%, and multinomial Bayes with least performance of (...) 69% accuracy. CNN has the best performance in the presence of large enough training dataset because of the presence of filter/ kernels which help to indentify patterns in text data regardless of their position in the sentence. We repeated the training again with just one-third of our data, at this point SVM comes with the best performance, the performance of CNN noticeably drops but still better than Multinomial Naive Bayes, the reason why SVM works best when we reduce the training data was because of its ability to look for a hyper-plane that creates a boundary between different classes of data so as to properly classify them, so we believed that getting the hyper-plane was more efficient when we reduce the dataset, hence reason for the good performance. Multinomial Naive Bayes have the least performance which we attributed to its assumption of independency between the features which sometimes does not hold true. We concluded that availability of data should be an important factor when choosing classifier for Natural Language Processing Text Classification task. CNN should be use in the presence of enough dataset, and SVM should be use when data is not enough. Multinomial Naive Bayes must not be trusted with state of the art NLP task due to its assumption of independency between the features . (shrink)

This work demonstrates the implementation and use of an encoder-decoder model to perform a many-to-many mapping of video data to text captions. The many-to-many mapping occurs via an input temporal sequence of video frames to an output sequence of words to form a caption sentence. Data preprocessing, model construction, and model training are discussed. Caption correctness is evaluated using 2-gram BLEU scores across the different splits of the dataset. Specific examples of output captions were shown to demonstrate model generality over (...) the video temporal dimension. Predicted captions were shown to generalize over video action, even in instances where the video scene changed dramatically. Model architecture changes are discussed to improve sentence grammar and correctness. (shrink)

We often hear how scientists have discovered that a certain human trait – or a trait of another type of organism – is innate, genetic, heritable, inherited, naturally selected etc. All these claims have something in common: they all declare a trait to have significant organism internal (for instance genetic) causes that are present in the organism at its birth. I call claims like these “nativist claims”. Nativist claims are important. They shape our overall understanding of what we are, what (...) we can do and what is good for us. They also impact our practical decision making. For instance, we know that certain types of genes increase the risk of certain diseases, and increasingly this has an effect on healthcare practices. Many people believe that in the matter of children’s education we should consider the fact that many cognitive abilities are highly heritable. Associations between genetics and aggressive behavior have been taken to be alleviating circumstances in criminal sentencing. It is therefore important to correctly understand the content and implications of nativist claims. In this doctoral thesis I analyze, by philosophical means, the content of different kinds of nativist claim. What does it mean that some psychological tendency is innate? What does it mean that intelligence is 50% heritable? How (if at all) do genes come to carry information about the traits of an organism – they are nothing like the information carriers familiar to us, such as newspapers, e-mails, secret codes, and audio files. This thesis emphasizes two things. Firstly, I demonstrate, that in different scientific contexts these words ('innate', 'genetically caused' etc.) can mean very different things. Therefore, the implications of a given nativist claim in one scientific context can be remarkably different from another nativist claim’s implications in some other scientific context. Secondly, I explain how the development of innate, heritable and genetically caused traits can be determined in different and often unexpected ways by our environment and experiences. I demonstrate that the content of a concept that is considered innate by cognitive scientists can be determined by what the organism that possesses the concept has experienced. I also show that even a trait which is 100% heritable can at the same time be a socially constructed trait. (shrink)

The common use of Shannon’s information, specified for the needs of telecommunications, gives rise to many misunderstandings outside of this context. (e.g. in conceptions of such well-known theorists as C.F. Weizsäcker and J. A. Wheeler). This article shows that the terms of the general definition of information meet the structural information, and Shannon’s information is a special case of it. Similarly, complexity is misunderstood today as exemplified by the concepts of reputable computer scientists, such as S. Lloyd, S. Wolfram and (...) J. Schmidhuber. These theorists use an algorithmic definition of complexity and the so-called logical depth, neither of which meets the intuitive criterion of complexity. Hence, their misconceptions of beauty, art and the universe. It will be shown that the intuitive criterion is met by Abstract Complexity Definition. This definition also fulfils the criterion for a general complexity definition, as it defines complexity of the most general/abstract structure of our reality, that is, binary structure. It also explains such fundamental issues of information theory as the information–energy relationship and the value of information, which are still discussed and need to be clarified. (shrink)

In this chapter, we propose a non-traditional RCR training in data science that is grounded in a virtue theory framework. First, we delineate the approach in more theoretical detail by discussing how the goal of RCR training is to foster the cultivation of certain moral abilities. We specify the nature of these ‘abilities’: while the ideal is the cultivation of virtues, the limited space allowed by RCR modules can only facilitate the cultivation of superficial abilities or proto-virtues, which help students (...) to familiarize themselves with moral and political issues in the data science environment. Third, we operationalize our approach by stressing that (proto-)virtue acquisition (like skill acquisition) occurs through the technical and social tasks of daily data science activities, where these repetitive tasks provide the opportunities to develop (proto-)virtue capacity and to support the development of ethically robust data systems. Finally, we discuss a concrete example of implementing this approach. In particular, we describe how this method is applied to teach data ethics to students participating in the CODATA-RDA Data Science Summer Schools. (shrink)

Enterprise Risk Management involves the process of identification, evaluation, treatment, and communication regarding risks throughout the enterprise. To support the tasks associated with this process, several frameworks and modeling languages have been proposed, such as the Risk and Security Overlay (RSO) of ArchiMate. An ontological investigation of this artifact would reveal its adequacy, capabilities, and limitations w.r.t. the domain of risk and security. Based on that, a language redesign can be proposed as a refinement. Such analysis and redesign have been (...) executed for the risk elements of the RSO grounded in the Common Ontology of Value and Risk. The next step along this line of research is to address the following research problems: What would be the outcome of an ontological analysis of security-related elements of the RSO? That is, can we identify other semantic deficiencies in the RSO through an ontological analysis? Once such an analysis is provided, can we redesign the security elements of the RSO accordingly, in order to produce an improved artifact? Here, with the aid of the Reference Ontology for Security Engineering (ROSE) and the ontological theory of prevention behind it, we address the remaining gap by proceeding with an ontological analysis of the security-related constructs of the RSO. The outcome of this assessment is an ontology-based redesign of the ArchiMate language regarding security modeling. In a nutshell, we report the following contributions: (1) an ontological analysis of the RSO that identifies six limitations concerning security modeling; (2) because of the key role of the notion of prevention in security modeling, the introduction of the ontological theory of prevention in ArchiMate; (3) a well-founded redesign of security elements of ArchiMate; and (4) ontology-based security modeling patterns that are logical consequences of our proposal of redesign due to its underlying ontology of security. As a form of evaluation, we show that our proposal can describe risk treatment options, according to ISO 31000. Finally, besides presenting multiple examples, we proceed with a real-world illustrative application taken from the cybersecurity domain. (shrink)

This study examines the concept of “description” and its theoretical foundations. The literature about it is surprisingly limited, and its usage is vague, sometimes even conflicting. Description should be considered in relation to other processes, such as representation, data capturing, and categorizing, which raises the question about what it means to describe something. Description is often used for any type of predication but may better be limited to predications based on observations. Research aims to establish criteria for making optimal descriptions; (...) however, the problems involved in describing something have seldom been addressed. Specific ideals are often followed without examine their fruitfulness. This study provides evidence that description cannot be a neutral, objective activity; rather, it is a theory-laden and interest-based activity. In information science, description occurs in processes such as document description, descriptive metadata assignment, and information resource description. In this field, description has equally been used in conflicting ways that mostly do not evince a recognition of the value- and theory-laden nature of descriptions. It is argued that descriptive activities in information science should always be based on consciously explicit considerations of the goals that descriptions are meant to serve. (shrink)

The mechanistic account of computation offers one promising and influential theory of computational implementation. The aim of this paper is to shore up its conceptual foundations by responding to several recent challenges. After outlining and responding to a recent proposal from Kuokkanen (2022a), I suggest that computational description should be conceptualised as a form of idealisation (selectively attending to modified subsets of model features) rather than abstraction (selectively attending to subsets of features within a target system). I argue that this (...) conceptualisation not only offers the best way of making sense of computational implementation, but also a way of resolving each of the outstanding challenges facing the mechanistic account. The idealisation view allows the mechanistic account to make sense of the omission process found in computational descriptions without leaving the relationship between physical and computational properties mysterious. (shrink)

A new AI system, called AlphaGeometry, trained under synthetic data has demonstrated the ability to solve geometric problems at the International Olympiad level. This essay considers the fact that human abilities to learn and do math as well as many other tasks are increasingly augmented with AI. Clearly, smart technologies like AlphaGeometry are redefining a number of concepts and institutions such as learning, schools, education, teacher-student relationships, creativity, etc, which are so fundamental for what we’ve thought of as modern society, (...) economy, and culture. What are we preparing for in mathematics education in the face of the possibility of the continued emergence of disruptive technologies, signaling the beginning of the end, if a new reality of mathematics is not taken into account? (shrink)

This brief article explores recent ethical issues surrounding digital technology, illustrating the need to ensure everyone's digital ethics awareness.

Ethically significant consequences of artificially intelligent artifacts will stem from their effects on existing social relations. Artifacts will serve in a variety of socially important roles—as personal companions, in the service of elderly and infirm people, in commercial, educational, and other socially sensitive contexts. The inevitable disruptions that these technologies will cause to social norms, institutions, and communities warrant careful consideration. As we begin to assess these effects, reflection on degrees and kinds of social agency will be required to make (...) properly informed decisions concerning the deployment of artificially intelligent artifacts in important settings. The social agency of these systems is unlike a human social agency, and this paper provides a methodological framework that is more suited for inquiry into artificial social agents than conventional philosophical treatments of the concept of agency. Separate aspects and dimensions of agency can be studied without assuming that agency must always look like adult human agency. This revised approach to the agency of artifacts is conducive to progress in the topics studied by AI ethics. (shrink)

Humankind is currently confronted with a critical challenge that determines its very existence, not only on an individual, racial, or national level but as a whole species: the fight against climate change and environmental degradation. To win this battle, humanity needs innovations and non-linear thinking. Nature has long been a substantial information source for unthinkable discoveries that save human lives. The paper suggests that by understanding the nature, emergence, and mechanism of serendipity, the survival skill of humans, humanity can capitalize (...) on it to learn from nature and generate nature-based innovations that can address the climate and environmental degradation crises. In the era of Artificial Intelligence (AI) proliferation, AI will be a vital tool providing navigational and useful information to leverage the power of serendipity in climate and sustainability science. However, an eco-surplus culture that incorporates core values pertaining to nature and sees environmental sustainability as the conscience of the times needs to be built to direct and control the immense power of AI-leveraged serendipity. Insights applied in this paper are drawn from the book titled “A New Theory of Serendipity: Nature, Emergence and Mechanism”. (shrink)

This article presents a fairness principle for evaluating decision-making based on predictions: a decision rule is unfair when the individuals directly impacted by the decisions who are equal with respect to the features that justify inequalities in outcomes do not have the same statistical prospects of being benefited or harmed by them, irrespective of their socially salient morally arbitrary traits. The principle can be used to evaluate prediction-based decision-making from the point of view of a wide range of antecedently specified (...) substantive views about justice in outcome distributions. (shrink)

The explanation for everything in Nature, everything in human history, future, and-or, past, is the conservation of a circle, proven by, the circular-linear relationship between, the solstice and the equinox.

Why would we want to develop artificial human-like arithmetical intelligence, when computers already outperform humans in arithmetical calculations? Aside from arithmetic consisting of much more than mere calculations, one suggested reason is that AI research can help us explain the development of human arithmetical cognition. Here I argue that this question needs to be studied already in the context of basic, non-symbolic, numerical cognition. Analyzing recent machine learning research on artificial neural networks, I show how AI studies could potentially shed (...) light on the development of human numerical abilities, from the proto-arithmetical abilities of subitizing and estimating to counting procedures. Although the current results are far from conclusive and much more work is needed, I argue that AI research should be included in the interdisciplinary toolbox when we try to explain the development and character of numerical cognition and arithmetical intelligence. This makes it relevant also for the epistemology of mathematics. (shrink)

Theories of rational decision making often abstract away from computational and other resource limitations faced by real agents. An alternative approach known as resource rationality puts such matters front and center, grounding choice and decision in the rational use of finite resources. Anticipated by earlier work in economics and in computer science, this approach has recently seen rapid development and application in the cognitive sciences. Here, the theory of rationality plays a dual role, both as a framework for normative assessment (...) and as a source of scientific hypotheses about how mental processes in fact work. The latter project, often called rational analysis, depends for its success on a fine-grained characterization of the computational problem facing a decision maker, which may in turn depend on realistic assumptions about what the relevant agent is like. As a consequence, resource rationality involves a delicate, but often fruitful interplay between the normative and the descriptive. (shrink)

Is a machine running artificial intelligence algorithms a real artificial intelligence or does it only simulate the operations of an intelligent creature? After specifying the concept of a mechanical simulation of human intelligence, we propose a criterion or a test to determine in which cases a machine embedding AI technology should be considered, not as a mechanical simulator of intelligent operations, but as a real artificial intelligence: we call this test the test of deference.

Cheating is a growing academic and ethical concern in higher education. This article examines the rise of artificial intelligence (AI) generative chatbots for use in education and provides a review of research literature and relevant scholarship concerning the cheating-related issues involved and their implications for pedagogy. The technological “arms race” that involves cheating-detection system developers versus technology savvy students is attracting increased attention to cheating. AI has added new dimensions to academic cheating challenges as students (as well as faculty and (...) staff) can easily access powerful systems for generating content that can be presented in assignments, exams, or published papers as their own. AI methodology is also providing some emerging anti-cheating approaches, including facial recognition and water- marking. This article provides an overview of human/AI collaboration approaches and frames some educational misuses of such AI generative systems as ChatGPT and Bard as forms of “misattributed co-authorship.” As with other kinds of collaborations, the work that students produce with AI assistance can be presented in transparent and straightforward modes or (unfortunately) in opaquer and ethically-problematic ways. However, rather than just for catching or entrapping students, the emerging varieties of technological cheating-detection strategies can be used to assist students in learning how to document and attribute properly their AI-empowered as well as human-human collaborations. Construing misuses of AI generative systems as misattributed co-authorship can recognize the growing capabilities of these tools and how stressing responsible and mindful usage by students can help prepare them for a highly collaborative, AI-saturated future. (shrink)

This book provides a framework for thinking about foundational philosophical questions surrounding machine learning as an approach to artificial intelligence. Specifically, it links recent breakthroughs in deep learning to classical empiricist philosophy of mind. In recent assessments of deep learning's current capabilities and future potential, prominent scientists have cited historical figures from the perennial philosophical debate between nativism and empiricism, which primarily concerns the origins of abstract knowledge. These empiricists were generally faculty psychologists; that is, they argued that the active (...) engagement of general psychological faculties-such as perception, memory, imagination, attention, and empathy-enables rational agents to extract abstract knowledge from sensory experience. This book explains a number of recent attempts to model roles attributed to these faculties in deep neural network based artificial agents by appeal to the faculty psychology of philosophers such as Aristotle, Ibn Sina (Avicenna), John Locke David Hume, William James, and Sophie de Grouchy. It illustrates the utility of this interdisciplinary connection by showing how it can provide benefits to both philosophy and computer science: computer scientists can continue to mine the history of philosophy for ideas and aspirational targets to hit on the way to more robustly rational artificial agents, and philosophers can see how some of the historical empiricists' most ambitious speculations can be realized in specific computational systems. (shrink)

The medium‐independence of computational descriptions has shaped common conceptions of computational explanation. So long as our goal is to explain how a system successfully carries out its computations, then we only need to describe the abstract series of operations that achieve the desired input–output mapping, however they may be implemented. It is argued that this abstract conception of computational explanation cannot be applied to so‐called real‐time computing systems, in which meeting temporal deadlines imposed by the systems with which a device (...) interfaces are constitutive of the computing tasks that a device performs. Instead, real‐time computing reveals the need for alternative conceptions of computational explanation, as well as computational implementation, that eschew medium‐independence. (shrink)

Con số vượt qua mốc 8000 cũng phần nào cho thấy bayesvl thu hút được một lượng quan tâm nhất định của giới nghiên cứu.