Machine Learning

We consider an extended notion of reinforcement learning in which the environment can simulate the agent and base its outputs on the agent's hypothetical behavior. Since good performance usually requires paying attention to whatever things the environment's outputs are based on, we argue that for an agent to achieve on-average good performance across many such extended environments, it is necessary for the agent to self-reflect. Thus weighted-average performance over the space of all suitably well-behaved extended environments could be considered a (...) way of measuring how self-reflective an agent is. We give examples of extended environments and introduce a simple transformation which experimentally seems to increase some standard RL agents' performance in a certain type of extended environment. (shrink)

At its strongest, Hume's problem of induction denies the existence of any well justified assumptionless inductive inference rule. At the weakest, it challenges our ability to articulate and apply good inductive inference rules. This paper examines an analysis that is closer to the latter camp. It reviews one answer to this problem drawn from the VC theorem in statistical learning theory and argues for its inadequacy. In particular, I show that it cannot be computed, in general, whether we are in (...) a situation where the VC theorem can be applied for the purpose we want it to. (shrink)

In their paper 'Reward is enough', Silver et al conjecture that the creation of sufficiently good reinforcement learning (RL) agents is a path to artificial general intelligence (AGI). We consider one aspect of intelligence Silver et al did not consider in their paper, namely, that aspect of intelligence involved in designing RL agents. If that is within human reach, then it should also be within AGI's reach. This raises the question: is there an RL environment which incentivises RL agents to (...) design RL agents? (shrink)

Inspired by recent progress in multi-agent Reinforcement Learning (RL), in this work we examine the collective intelligent behaviour of theoretical universal agents by introducing a weighted mixture operation. Given a weighted set of agents, their weighted mixture is a new agent whose expected total reward in any environment is the corresponding weighted average of the original agents' expected total rewards in that environment. Thus, if RL agent intelligence is quantified in terms of performance across environments, the weighted mixture's intelligence is (...) the weighted average of the original agents' intelligences. This operation enables various interesting new theorems that shed light on the geometry of RL agent intelligence, namely: results about symmetries, convex agent-sets, and local extrema. We also show that any RL agent intelligence measure based on average performance across environments, subject to certain weak technical conditions, is identical (up to a constant factor) to performance within a single environment dependent on said intelligence measure. (shrink)

Machine learning operates at the intersection of statistics and computer science. This raises the question as to its underlying methodology. While much emphasis has been put on the close link between the process of learning from data and induction, the falsificationist component of machine learning has received minor attention. In this paper, we argue that the idea of falsification is central to the methodology of machine learning. It is commonly thought that machine learning algorithms infer general prediction rules from past (...) observations. This is akin to a statistical procedure by which estimates are obtained from a sample of data. But machine learning algorithms can also be described as choosing one prediction rule from an entire class of functions. In particular, the algorithm that determines the weights of an artificial neural network operates by empirical risk minimization and rejects prediction rules that lack empirical adequacy. It also exhibits a behavior of implicit regularization that pushes hypothesis choice toward simpler prediction rules. We argue that taking both aspects together gives rise to a falsificationist account of artificial neural networks. (shrink)

Many thinkers, past, and present, have tried to solve the underlying mystery of Life. Yet, no one has ever categorically expressed its exact concrete essence, scope, or meaning until a new school of thought known as Originemology was conceptualized in 1988 by Joey Lawsin. Life and consciousness can not be explained properly since their theoretical and philosophical bases are wrong. When the bases are incorrect, the outcomes are incorrect. The words associated with life such as alive, aware, conscious, intelligent, and (...) self-conscious are always interchanged and misinterpreted to be all one and the same. Now, they are properly corrected and redefined in this paper. -/- In addition, results uncovered in Autognorics, the science of engineered life forms (SELFS), are also presented here. One of its findings revealed that Life, like the emergence of the Self, is an evolutionary process. It is a sequential progression that begins from being alive to being with life. Rather than as a detailed thesis, the seven sequential signatures of life are addressed here in outline form - from (i)Mechanical Aliveness, (ii)Sensoric Awareness, (iii)Aneural Consciousness, (iv)Logical Intuitiveness, (v)Information Inlearness, (vi)Symbiotic Living to (vii)Self Emergence. To be conscious, one only needs the first three criteria: alive, aware, and aneural(a brain without the brain). The third requirement is based on the Theory of Codexation. And to emerge as "me" (Self), the seven evolutionary orders of life are necessary. -/- A seed specimen called ELFS, or engineered life forms, is also part of the experiment developed primarily to address predictions, anticipate behaviors, discover new perspectives, and rewrite the true essence of life. Aside from the Aneural prototype, neural specimens such as dogs, mice, birds, fish, plants, and molds are utilized as well in this research to eliminate the misguided overrated concept that consciousness emerges solely in creatures with brains. Like consciousness, thoughts, emotions, animations, dreams, music, colors, senses, and self, Life is a generated interim. It is a by-product that only emerges due to the presence of all objects that cause it exists. According to the Theory of Generated Emergence, everything only exists because other things cause it to exist, or else everything never existed at all. (shrink)

Wittgenstein’s Private Language Argument holds that language requires rule-following, rule following requires the possibility of error, error is precluded in pure introspection, and inner mental life is known only by pure introspection, thus language cannot exist entirely within inner mental life. Fodor defends his Language of Thought program against the Private Language Argument with a dilemma: either privacy is so narrow that internal mental life can be known outside of introspection, or so broad that computer language serves as a counter-example. (...) I suggest that the developing field of artificial intelligence (deep learning neural networks) tends to vitiate Fodor’s defense and hence vindicate the Private Language Argument. The first horn of Fodor’s dilemma requires language to encompass genuinely internal mental life, i.e. non-projected intentional states, which are not exhibited in classical machine learning but only by deep learning neural networks (artificial intelligence). Such networks act as black boxes, however, whose state cannot be understood by tracking the changes in their supervenience bases without shared context, and that shared context introduces the possibility of error. The language of artificial intelligence is not private. (shrink)

As our epistemic ambitions grow, the common and scientific endeavours are becoming increasingly dependent on Machine Learning (ML). The field rests on a single experimental paradigm, which consists of splitting the available data into a training and testing set and using the latter to measure how well the trained ML model generalises to unseen samples. If the model reaches acceptable accuracy, an a posteriori contract comes into effect between humans and the model, supposedly allowing its deployment to target environments. Yet (...) the latter part of the contract depends on human inductive predictions or generalisations, which infer a uniformity between the trained ML model and the targets. The paper asks how we justify the contract between human and machine learning. It is argued that the justification becomes a pressing issue when we use ML to reach ‘elsewheres’ in space and time or deploy ML models in non-benign environments. The paper argues that the only viable version of the contract can be based on optimality (instead of on reliability which cannot be justified without circularity) and aligns this position with Schurz’s optimality justification. It is shown that when dealing with inaccessible/unstable ground-truths (‘elsewheres’ and non-benign targets), the optimality justification undergoes a slight change, which should reflect critically on our epistemic ambitions. Therefore, the study of ML robustness should involve not only heuristics that lead to acceptable accuracies on testing sets. The justification of human inductive predictions or generalisations about the uniformity between ML models and targets should be included as well. Without it, the assumptions about inductive risk minimisation in ML are not addressed in full. (shrink)

Recently, a connection has been established between two branches of computability theory, namely between algorithmic randomness and algorithmic learning theory. Learning-theoretical characterizations of several notions of randomness were discovered. We study such characterizations based on the asymptotic density of positive answers. In particular, this note provides a new learning-theoretic definition of weak 2-randomness, solving the problem posed by (Zaffora Blando, Rev. Symb. Log. 2019). The note also highlights the close connection between these characterizations and the problem of convergence on random (...) sequences. (shrink)

Douven (in press) observes that Schurz's meta-inductive justification of induction cannot explain the great empirical success of induction, and offers an explanation based on computer simulations of the social and evolutionary development of our inductive practices. In this paper, I argue that Douven's account does not address the explanatory question that Schurz's argument leaves open, and that the assumption of the environment's induction-friendliness that is inherent to Douven's simulations is not justified by Schurz's argument.

Definition of the problem The umbrella term “explicability” refers to the reduction of opacity of artificial intelligence (AI) systems. These efforts are challenging for medical AI applications because higher accuracy often comes at the cost of increased opacity. This entails ethical tensions because physicians and patients desire to trace how results are produced without compromising the performance of AI systems. The centrality of explicability within the informed consent process for medical AI systems compels an ethical reflection on the trade-offs. Which (...) levels of explicability are needed to obtain informed consent when utilizing medical AI? Arguments We proceed in five steps: First, we map the terms commonly associated with explicability as described in the ethics and computer science literature, i.e., disclosure, intelligibility, interpretability, and explainability. Second, we conduct a conceptual analysis of the ethical requirements for explicability when it comes to informed consent. Third, we distinguish hurdles for explicability in terms of epistemic and explanatory opacity. Fourth, this then allows to conclude the level of explicability physicians must reach and what patients can expect. In a final step, we show how the identified levels of explicability can technically be met from the perspective of computer science. Throughout our work, we take diagnostic AI systems in radiology as an example. Conclusion We determined four levels of explicability that need to be distinguished for ethically defensible informed consent processes and showed how developers of medical AI can technically meet these requirements. (shrink)

Recent influential accounts of temporal representation—the use of mental representations with explicit temporal contents, such as before and after relations and durations—sharply distinguish representation from mere sensitivity. A common, important picture of inter-temporal rationality is that it consists in maximizing total expected discounted utility across time. By analyzing reinforcement learning algorithms, this article shows that, given such notions of temporal representation and inter-temporal rationality, it would be possible for an agent to achieve inter-temporal rationality without temporal representation. It then explores (...) potential upshots of this result for theorizing about rationality and representation. (shrink)

Methods of machine learning (ML) are gradually complementing and sometimes even replacing methods of classical statistics in science. This raises the question whether ML faces the same methodological problems as classical statistics. This paper sheds light on this question by investigating a long-standing challenge to classical statistics: the reference class problem (RCP). It arises whenever statistical evidence is applied to an individual object, since the individual belongs to several reference classes and evidence might vary across them. Thus, the problem consists (...) in choosing a suitable reference class for the individual. I argue that deep neural networks (DNNs) are able to overcome specific instantiations of the RCP. Whereas the criteria of narrowness, reliability, and homogeneity, that have been proposed to determine a suitable reference class, pose an inextricable tradeoff to classical statistics, DNNs are able to satisfy them in some situations. On the one hand, they can exploit the high dimensionality in big-data settings. I argue that this corresponds to the criteria of narrowness and reliability. On the other hand, ML research indicates that DNNs are generally not susceptible to overfitting. I argue that this property is related to a particular form of homogeneity. Taking both aspects together reveals that there are specific settings in which DNNs can overcome the RCP. (shrink)

La scorciatoia - Come le macchine sono diventate intelligenti senza pensare in modo umano -/- Le nostre creature sono diverse da noi e talvolta più forti. Per poterci convivere dobbiamo imparare a conoscerle Vagliano curricula, concedono mutui, scelgono le notizie che leggiamo: le macchine intelligenti sono entrate nelle nostre vite, ma non sono come ce le aspettavamo. Fanno molte delle cose che volevamo, e anche qualcuna in più, ma non possiamo capirle o ragionare con loro, perché il loro comportamento è (...) in realtà guidato da relazioni statistiche ricavate da quantità sovrumane di dati. Eppure possono essere in certi casi più potenti di noi: ci osservano continuamente, e prendono decisioni al nostro posto. E allora come incorporarle nella nostra società senza rischi ed effetti collaterali? Questo libro - rigoroso, pungente, originale nell'approccio - ci spiega come siamo arrivati sin qui, e indica il percorso che ci aspetta prima di poterci fidare di questi nuovi agenti «alieni». La tecnologia non basta, occorre un dialogo tra scienze naturali e umane: è il passaggio cruciale per una convivenza sicura con questa nuova forma di intelligenza. (Il Mulino). (shrink)

If Large Language Models can make real scientific contributions, then they can genuinely use language, be systematically wrong, and be held responsible for their errors. AI models which can make scientific contributions thereby meet the criteria for scientific authorship.

We present a game mechanic called pseudo-visibility for games inhabited by non-player characters (NPCs) driven by reinforcement learning (RL). NPCs are incentivized to pretend they cannot see pseudo-visible players: the training environment simulates an NPC to determine how the NPC would act if the pseudo-visible player were invisible, and penalizes the NPC for acting differently. NPCs are thereby trained to selectively ignore pseudo-visible players, except when they judge that the reaction penalty is an acceptable tradeoff (e.g., a guard might accept (...) the penalty in order to protect a treasure because losing the treasure would hurt even more). We describe an RL agent transformation which allows RL agents that would not otherwise do so to perform some limited self-reflection to learn the training environments in question. (shrink)

This paper presents a new trilemma with respect to resolving the control and alignment problems in machine ethics. Section 1 outlines three possible types of artificial moral agents (AMAs): (1) 'Inhuman AMAs' programmed to learn or execute moral rules or principles without understanding them in anything like the way that we do; (2) 'Better-Human AMAs' programmed to learn, execute, and understand moral rules or principles somewhat like we do, but correcting for various sources of human moral error; and (3) 'Human-Like (...) AMAs' programmed to understand and apply moral values in broadly the same way that we do, with a human-like moral psychology. Sections 2–4 then argue that each type of AMA generates unique control and alignment problems that have not been fully appreciated. Section 2 argues that Inhuman AMAs are likely to behave in inhumane ways that pose serious existential risks. Section 3 then contends that Better-Human AMAs run a serious risk of magnifying some sources of human moral error by reducing or eliminating others. Section 4 then argues that Human-Like AMAs would not only likely reproduce human moral failures, but also plausibly be highly intelligent, conscious beings with interests and wills of their own who should therefore be entitled to similar moral rights and freedoms as us. This generates what I call the New Control Problem: ensuring that humans and Human-Like AMAs exert a morally appropriate amount of control over each other. Finally, Section 5 argues that resolving the New Control Problem would, at a minimum, plausibly require ensuring what Hume and Rawls term ‘circumstances of justice’ between humans and Human-Like AMAs. But, I argue, there are grounds for thinking this will be profoundly difficult to achieve. I thus conclude on a skeptical note. Different approaches to developing ‘safe, ethical AI’ generate subtly different control and alignment problems that we do not currently know how to adequately resolve, and which may or may not be ultimately surmountable. (shrink)

Machine learning researchers distinguish between reinforcement learning and supervised learning and refer to reinforcement learning systems as “agents”. This paper vindicates the claim that systems trained by reinforcement learning are agents while those trained by supervised learning are not. Systems of both kinds satisfy Dretske’s criteria for agency, because they both learn to produce outputs selectively in response to inputs. However, reinforcement learning is sensitive to the instrumental value of outputs, giving rise to systems which exploit the effects of outputs (...) on subsequent inputs to achieve good performance over episodes of interaction with their environments. Supervised learning systems, in contrast, merely learn to produce better outputs in response to individual inputs. (shrink)

Le déficit d’explicabilité des techniques d’apprentissage machine (AM) pose des problèmes opérationnels, juridiques et éthiques. Un des principaux objectifs de notre projet est de fournir des explications éthiques des sorties générées par une application fondée sur de l’AM, considérée comme une boîte noire. La première étape de ce projet, présentée dans cet article, consiste à montrer que la validation de ces boîtes noires diffère épistémologiquement de celle mise en place dans le cadre d’une modélisation mathéma- tique et causale d’un phénomène (...) physique. La différence majeure est qu’une méthode d’AM ne prétend pas représenter une causalité entre les paramètres d’entrées et ceux de sortie. Après avoir proposé une clarification et une adaptation des notions d’inter- prétabilité et d’explicabilité telles qu’on les rencontre dans la littérature déjà abondante sur le sujet, nous montrons dans cet article l’intérêt de mettre en œuvre les distinctions épistémologiques entre les différentes fonctions épistémiques d’un modèle, d’une part, et entre la fonction épistémique et l’usage d’un modèle, d’autre part. Enfin, la dernière partie de cet article présente nos travaux en cours sur l’évaluation d’une explication, qui peut être plus persuasive qu’informative, ce qui peut ainsi causer des problèmes d’ordre éthique. (shrink)

Many AI systems that make important decisions are black boxes: how they function is opaque even to their developers. This is due to their high complexity and to the fact that they are trained rather than programmed. Efforts to alleviate the opacity of black box systems are typically discussed in terms of transparency, interpretability, and explainability. However, there is little agreement about what these key concepts mean, which makes it difficult to adjudicate the success or promise of opacity alleviation methods. (...) I argue for a unified account of these key concepts that treats the concept of understanding as fundamental. This allows resources from the philosophy of science and the epistemology of understanding to help guide opacity alleviation efforts. A first significant benefit of this understanding account is that it defuses one of the primary, in-principle objections to post hoc explainable AI (XAI) methods. This “rationalization objection” argues that XAI methods provide mere rationalizations rather than genuine explanations. This is because XAI methods involve using a separate “explanation” system to approximate the original black box system. These explanation systems function in a completely different way than the original system, yet XAI methods make inferences about the original system based on the behavior of the explanation system. I argue that, if we conceive of XAI methods as idealized scientific models, this rationalization worry is dissolved. Idealized scientific models misrepresent their target phenomena, yet are capable of providing significant and genuine understanding of their targets. (shrink)

Should we fear a future in which the already tricky world of academic publishing is increasingly crowded out by super-intelligent artificial general intelligence (AGI) systems writing papers on phenomenology and ethics? What are the chances that AGI advances to a stage where a human philosophy instructor is similarly removed from the equation? If Jobst Landgrebe and Barry Smith are correct, we have nothing to fear.

Although machine intelligence is increasingly employed in healthcare, the realm of decision-making in medical ethics remains largely unexplored from a technical perspective. We propose an approach based on fuzzy cognitive maps (FCMs), which builds on Beauchamp and Childress’ prima-facie principles. The FCM’s weights are optimized using a genetic algorithm to provide recommendations regarding the initiation, continuation, or withdrawal of medical treatment. The resulting model approximates the answers provided by our team of medical ethicists fairly well and offers a high degree (...) of interpretability. Possible applications of such a system include informal guidance on medical ethics dilemmas as well as educational purposes. (shrink)

This essay describes trends in the organization of work that have laid the groundwork for the adoption of interactive AI-driven instruction tools, and the technological innovations that will make intelligent tutoring systems truly competitive with human teachers. Since the origin of occupational specialization, the collection and transmission of knowledge have been tied to individual careers and job roles, specifically doctors, teachers, clergy, and lawyers, the paradigmatic knowledge professionals. But these roles have also been tied to texts and organizations that can (...) disseminate knowledge independently from professionals. Professionals and organizations turn knowledge into texts and tools that enable lay people to access knowledge without the intermediation of professionals or organizations. In the 21st century, one emerging tool for transmitting knowledge is the intelligent tutoring system. This paper examines how technological, epistemic, and economic trends in education are supporting the routinization, proletarianization, and automation of the occupation of teaching, leading to the increasing substitution of intelligent tutoring systems for human instruction. Some trends, such as standardized curricula and testing, both restrict teachers’ professional autonomy and facilitate the creation of pedagogical tools. Other trends reduce teachers’ ability to resist automation. The growth of adjunct teaching and paraprofessional roles in higher education allows organizations to take over and rationalize parts of the traditional teacher role. Faculty evaluations and learning outcomes assessment weaken professional claims to be the sole arbiters of instructional quality and student learning. The widespread use of intelligent tutoring systems also depends on the sophistication of software capable of performing the social-emotional and cognitive roles that educators perform. Eventually, pedagogical software will be able to interactively individualize curricula to the needs and interests of every learner, more cheaply, quickly, and accurately than any human teacher. Assessment of learning will be continuous, and certification of learning will be for specific skills instead of broad area competencies. Intelligent tutoring systems will help transition education from its medieval and industrial-era model to more accessible and flexible continuing education for employment and life enrichment. (shrink)

Unless and until our society recognizes cyber bullying for what it is, the suffering of thousands of silent victims will continue.” ~ Anna Maria Chavez. There had been series of research on cyber bullying which are unable to provide reliable solution to cyber bullying. In this research work, we were able to provide a permanent solution to this by developing a model capable of detecting and intercepting bullying incoming and outgoing messages with 92% accuracy. We also developed a chatbot automation (...) messaging system to test our model leading to the development of Artificial Intelligence powered anti-cyber bullying system using machine learning algorithm of Multinomial Naïve Bayes (MNB) and optimized linear Support Vector Machine(SVM). Our model is able to detect and intercept bullying outgoing and incoming bullying messages and take immediate action. (shrink)

Ever since the era of internet had ushered in cloud computing, there had been increase in the demand for the unlimited data available through cloud computing for data analysis, pattern recognition and technology advancement. With this also bring the problem of scalability, efficiency and security threat. This research paper focuses on how data can be dynamically mine in real time for pattern detection in a secure cloud computing environment using combination of decision tree algorithm and Random Forest over a restful (...) Application Programming Interface (API). We are able to successfully Implement data mining on cloud computing bypassing or avoiding direct interaction with data warehouse and without any terminal involve by using combination of IBM Cloud storage facility, Amazing Web Service, Application Programming Interface and Window service along with a decision tree and Random Forest algorithm for our classifier. We were able to successfully bypass direct connection with the data warehouse and cloud terminal with 94% accuracy in our model. (shrink)

The proposed system for coping negative emotions arising during the learning process is considered as an embedded part of the complex intelligent learning system realized in a digital environment. By applying data-driven procedures on the current and retrospective data the main didactic-based stimuli provoking emotion generation are identified. They are examined as dominant negative emotions in the context of learning. Due to the presence of strong internal and output interconnections between teaching and emotional states, an intelligent decoupling multidimensional control scheme (...) is accepted to overcome the lack of sample effective independent control actions that separately affect the states of learning and emotions. To avoid existing drawbacks in emotion-focused partial control, an approach with integrating emotions with a low-dimensional representation is accepted. Two-stage procedure is proposed to compensate the negative impact of emotions on the learning process. (shrink)

The study shows the existence of a broad conformity between Whitehead’s organismic cosmology and the contemporary theory of complex systems at a relevant level of abstraction. One of the most promising directions of educational transformation in the age of big data and artificial intelligence – personalized learning – is conceived as a system of systems and reveals its close congruence with a number of basic Whiteheadian concepts. A new functional structure of personalized learning systems is proposed, including all the core (...) elements of a full learning sequence. A multiobjective optimization problem, which is subject to strong constraints, uncertain outcomes, and continued development, is under consideration. It is argued that many of Whitehead’s concepts can be used constructively in designing and implementing advanced personalized learning systems after being adapted and expanded and account for the requirements of emerging big data and artificial intelligence research. Special attention is paid to the main factors that determine the multi-modality of personalized learning – learning styles, contexts, and didactic variability. The effecttiveness of personalized learning is a data-driven problem. (shrink)

Racial, Gender, and Intersectional Biases in AI / -/- Dominant View of Intersectional Fairness in the AI Literature / -/- Three Fundamental Problems with the Dominant View / 1. Overemphasis on Intersections of Attributes / 2. Dilemma between Infinite Regress and Fairness Gerrymandering / 3. Narrow Understanding of Fairness as Parity / -/- Rethinking AI Fairness: from Weak to Strong Fairness.

A growing number of studies on fairness in artificial intelligence (AI) use the notion of intersectionality to measure AI fairness. Most of these studies take intersectional fairness to be a matter of statistical parity among intersectional subgroups: an AI algorithm is “intersectionally fair” if the probability of the outcome is roughly the same across all subgroups defined by different combinations of the protected attributes. This paper identifies and examines three fundamental problems with this dominant interpretation of intersectional fairness in AI. (...) First, the dominant approach is so preoccupied with the intersection of attributes/categories (e.g., race, gender) that it fails to address the intersection of oppression (e.g., racism, sexism), which is more central to intersectionality as a critical framework. Second, the dominant approach faces a dilemma between infinite regress and fairness gerrymandering: it either keeps splitting groups into smaller subgroups or arbitrarily selects protected groups. Lastly, the dominant view fails to capture what it really means for AI algorithms to be fair, in terms of both distributive and non-distributive fairness. I distinguish a strong sense of AI fairness from a weak sense that is prevalent in the literature, and conclude by envisioning paths towards strong intersectional fairness in AI. (shrink)

There is a common misconception across the lntelligence Community (IC) to the effect that information trapped within multiple heterogeneous data silos can be semantically integrated by the sorts of meaning-blind statistical methods employed in much of artificial intelligence (Al) and natural language processlng (NLP). This leads to the misconception that incoming data can be analysed coherently by relying exclusively on the use of statistical algorithms and thus without any shared framework for classifying what the data are about. Unfortunately, such approaches (...) do not yield sustainable results where we are dealing with widely distributed, highly heterogeneous and often changing bodies of data. We argue here that the needed integration requires the use of what we call an lntegrating Semantic Framework (ISF), which provldes a consistent set of categories and relationships that can be reused over and over again to tag successive bodies of data in ways which foster more coherent representation and reasoning. (shrink)

This paper studies algorithmic learning theory applied to algebraic structures. In previous papers, we have defined our framework, where a learner, given a family of structures, receives larger and larger pieces of an arbitrary copy of a structure in the family and, at each stage, is required to output a conjecture about the isomorphism type of such a structure. The learning is successful if there is a learner that eventually stabilizes to a correct conjecture. Here, we analyze the number of (...) mind changes that are needed to learn a given family K. We give a descriptive set-theoretic interpretation of such mind change complexity. We also study how bounding the Turing degree of learners affects the mind change complexity of a given family of algebraic structures. (shrink)

We argue that machine learning algorithms can inflict microaggressions on members of marginalized groups and that recognizing these harms as instances of microaggressions is key to effectively addressing the problem. The concept of microaggression is also illuminated by being studied in algorithmic contexts. We contribute to the microaggression literature by expanding the category of environmental microaggressions and highlighting the unique issues of moral responsibility that arise when we focus on this category. We theorize two kinds of algorithmic microaggression, stereotyping and (...) erasure microaggressions, and argue that corporations are responsible for the microaggressions their algorithms create. As a case study, we look at the problems faced by Google’s autocomplete prediction and at the insufficiency of their solutions. The case study of autocomplete demonstrates our core claim that microaggressions constitute a distinct form of algorithmic bias and that identifying them as such allows us to avoid seeming solutions that recreate the same kinds of harms. Google has a responsibility to make information freely available, without exposing users to degradation. To fulfill its duties to marginalized groups, Google must abandon the fiction of neutral prediction and instead embrace the liberatory power of suggestion. (shrink)

Can machine intelligence do clinical ethics? And if so, would applying it to actual medical cases be desirable? In a recent target article (Meier et al. 2022), we described the piloting of our advisory algorithm METHAD. Here, we reply to commentaries published in response to our project. The commentaries fall into two broad categories: concrete criticism that concerns the development of METHAD; and the more general question as to whether one should employ decision-support systems of this kind—the debate we set (...) out to ignite with our target article. (shrink)

In a recent article, Erasmus, Brunet, and Fisher (2021) argue that Artificial Neural Networks (ANNs) are explainable. They survey four influential accounts of explanation: the Deductive-Nomological model, the Inductive-Statistical model, the Causal-Mechanical model, and the New-Mechanist model. They argue that, on each of these accounts, the features that make something an explanation is invariant with regard to the complexity of the explanans and the explanandum. Therefore, they conclude, the complexity of ANNs (and other Machine Learning models) does not make them (...) less explainable. In this reply, it is argued that Erasmus et al. left out one influential account of explanation from their discussion: the Unificationist model. It is argued that, on the Unificationist model, the features that makes something an explanation is sensitive to complexity. Therefore, on the Unificationist model, ANNs (and other Machine Learning models) are not explainable. It is emphasized that Erasmus et al.’s general strategy is correct. The literature on explainable Artificial Intelligence can benefit by drawing from philosophical accounts of explanation. However, philosophical accounts of explanation do not settle the problem of whether ANNs are explainable because they do not unanimously declare that explanation is invariant with regard to complexity. (shrink)

We can illustrate the phenomenon of model-induced escape by examining the phenomenon of spam filters. Spam filter A is, we can assume, very effective at blocking spam. Indeed it is so effective that it motivates the authors of spam to invent new types of spam that will beat the filters of spam filter A. -/- An example of this phenomenon in the realm of philosophy is illustrated in the work of Nyíri on Wittgenstein's political beliefs. Nyíri writes a paper demonstrating (...) convincingly that there are strong signals of a conservative strain of thought in the writings of Wittgenstein. This has initially a tiny effect; but then a more significant effect sets in as the authors of Wittgenstein secondary literature, consciously or unconsciously, draw attention to features of Wittgenstein which cast the conservatism thesis in a negative light. This is an example of what I call model-induced escape: the Nyíri model of Wittgenstein initiates a process which undermines the Nyíri model of Wittgenstein. -/- Wittgenstein, as obiectum philosophiae, is reduced to a constantly mutating set of interpretations of a certain body of work. -/- Perhaps this is what makes philosophy so problematic when viewed from the perspective of results, or in other words of signs of progress commonly accepted across the discipline. There are just too many ways of inducing escape from any given putative discovery; too many dimensions along which an interpretative or definitional arms race can be triggered. (shrink)

We study computable PAC (CPAC) learning as introduced by Agarwal et al. (2020). First, we consider the main open question of finding characterizations of proper and improper CPAC learning. We give a characterization of a closely related notion of strong CPAC learning, and provide a negative answer to the COLT open problem posed by Agarwal et al. (2021) whether all decidably representable VC classes are improperly CPAC learnable. Second, we consider undecidability of (computable) PAC learnability. We give a simple general (...) argument to exhibit such ndecidability, and initiate a study of the arithmetical complexity of learnability. We briefly discuss the relation to the undecidability result of Ben-David et al. (2019), that motivated the work of Agarwal et al. (shrink)

Under what conditions does machine learning (ML) model opacity inhibit the possibility of explaining and understanding phenomena? In this article, I argue that nonepistemic values give shape to the ML opacity problem even if we keep researcher interests fixed. Treating ML models as an instance of doing model-based science to explain and understand phenomena reveals that there is (i) an external opacity problem, where the presence of inductive risk imposes higher standards on externally validating models, and (ii) an internal opacity (...) problem, where greater inductive risk demands a higher level of transparency regarding the inferences the model makes. (shrink)

Simple idealized models seem to provide more understanding than opaque, complex, and hyper-realistic models. However, an increasing number of scientists are going in the opposite direction by utilizing opaque machine learning models to make predictions and draw inferences, suggesting that scientists are opting for models that have less potential for understanding. Are scientists trading understanding for some other epistemic or pragmatic good when they choose a machine learning model? Or are the assumptions behind why minimal models provide understanding misguided? In (...) this paper, using the case of deep neural networks, I argue that it is not the complexity or black box nature of a model that limits how much understanding the model provides. Instead, it is a lack of scientific and empirical evidence supporting the link that connects a model to the target phenomenon that primarily prohibits understanding. (shrink)

One of the primary, if not most critical, difficulties in the design and implementation of autonomous systems is the black-boxed nature of the decision-making structures and logical pathways. How human values are embodied and actualised in situ may ultimately prove to be harmful if not outright recalcitrant. For this reason, the values of stakeholders become of particular significance given the risks posed by opaque structures of intelligent agents (IAs). This paper explores how decision matrix algorithms, via the belief-desire-intention model for (...) autonomous vehicles, can be designed to minimize the risks of opaque architectures. Primarily through an explicit orientation towards designing for the values of explainability and verifiability. In doing so, this research adopts the Value Sensitive Design (VSD) approach as a principled framework for the incorporation of such values within design. VSD is recognized as a potential starting point that offers a systematic way for engineering teams to formally incorporate existing technical solutions within ethical design, while simultaneously remaining pliable to emerging issues and needs. It is concluded that the VSD methodology offers at least a strong enough foundation from which designers can begin to anticipate design needs and formulate salient design flows that can be adapted to the changing ethical landscapes required for utilisation in autonomous vehicles. (shrink)

Deep neural networks have become increasingly successful in applications from biology to cosmology to social science. Trained DNNs, moreover, correspond to models that ideally allow the prediction of new phenomena. Building in part on the literature on ‘eXplainable AI’, I here argue that these models are instrumental in a sense that makes them non-explanatory, and that their automated generation is opaque in a unique way. This combination implies the possibility of an unprecedented gap between discovery and explanation: When unsupervised models (...) are successfully used in exploratory contexts, scientists face a whole new challenge in forming the concepts required for understanding underlying mechanisms. (shrink)

A review of Gary Smith and Jay Cordes: The Phantom Pattern Problem: The Mirage of Big Data. New York: Oxford University Press, 2020.

Can humans and artificial intelligences share concepts and communicate? Making AI Intelligible shows that philosophical work on the metaphysics of meaning can help answer these questions. Herman Cappelen and Josh Dever use the externalist tradition in philosophy to create models of how AIs and humans can understand each other. In doing so, they illustrate ways in which that philosophical tradition can be improved. The questions addressed in the book are not only theoretically interesting, but the answers have pressing practical implications. (...) Many important decisions about human life are now influenced by AI. In giving that power to AI, we presuppose that AIs can track features of the world that we care about (for example, creditworthiness, recidivism, cancer, and combatants). If AIs can share our concepts, that will go some way towards justifying this reliance on AI. This ground-breaking study offers insight into how to take some first steps towards achieving Interpretable AI. (shrink)

We introduce and discuss a knowledge-driven distillation approach to explaining black-box models by means of two kinds of interpretable models. The first is perceptron (or threshold) connectives, which enrich knowledge representation languages such as Description Logics with linear operators that serve as a bridge between statistical learning and logical reasoning. The second is Trepan Reloaded, an ap- proach that builds post-hoc explanations of black-box classifiers in the form of decision trees enhanced by domain knowledge. Our aim is, firstly, to target (...) a model-agnostic distillation approach exemplified with these two frameworks, secondly, to study how these two frameworks interact on a theoretical level, and, thirdly, to investigate use-cases in ML and AI in a comparative manner. Specifically, we envision that user-studies will help determine human understandability of explanations generated using these two frameworks. (shrink)

Typically, fair machine learning research focuses on a single decision maker and assumes that the underlying population is stationary. However, many of the critical domains motivating this work are characterized by competitive marketplaces with many decision makers. Realistically, we might expect only a subset of them to adopt any non-compulsory fairness-conscious policy, a situation that political philosophers call partial compliance. This possibility raises important questions: how does partial compliance and the consequent strategic behavior of decision subjects affect the allocation outcomes? (...) If k% of employers were to voluntarily adopt a fairness-promoting intervention, should we expect k% progress (in aggregate) towards the benefits of universal adoption, or will the dynamics of partial compliance wash out the hoped-for benefits? How might adopting a global (versus local) perspective impact the conclusions of an auditor? In this paper, we propose a simple model of an employment market, leveraging simulation as a tool to explore the impact of both interaction effects and incentive effects on outcomes and auditing metrics. Our key findings are that at equilibrium: (1) partial compliance by k% of employers can result in far less than proportional (k%) progress towards the full compliance outcomes; (2) the gap is more severe when fair employers match global (vs local) statistics; (3) choices of local vs global statistics can paint dramatically different pictures of the performance vis-a-vis fairness desiderata of compliant versus non-compliant employers; (4) partial compliance based on local parity measures can induce extreme segregation. Finally, we discuss implications for auditors and insights concerning the design of regulatory frameworks. (shrink)

It has been argued that ethical frameworks for data science often fail to foster ethical behavior, and they can be difficult to implement due to their vague and ambiguous nature. In order to overcome these limitations of current ethical frameworks, we propose to integrate the analysis of the connections between technical choices and sociocultural factors into the data science process, and show how these connections have consequences for what data subjects can do, accomplish, and be. Using healthcare as an example, (...) attention to sociocultural conversion factors relevant to health can help in navigating technical choices that require broader considerations of the sociotechnical system, such as metric tradeoffs in model validation, resulting in better ethical and technical choices. This approach promotes awareness of the ethical dimension of technical choices by data scientists and others, and that can foster the cultivation of 'ethical skills' as integral to data science. (shrink)

The ethical decision-making and behaviour of artificially intelligent systems is increasingly important given the prevalence of these systems and the impact they can have on human well-being. Many current approaches to implementing machine ethics utilize top-down approaches, that is, ensuring the ethical decision-making and behaviour of an agent via its adherence to explicitly defined ethical rules or principles. Despite the attractiveness of this approach, this chapter explores how all top-down approaches to implementing machine ethics are fundamentally limited and how bottom-up (...) approaches, in particular, reinforcement learning methods, are not beset by the same problems as top-down approaches. Bottom-up approaches possess significant advantages that make them better suited for implementing machine ethics. (shrink)

Coronary Heart Disease (CHD) is one of the leading causes of death nowadays. Prediction of the disease at an early stage is crucial for many health care providers to protect their patients and save lives and costly hospitalization resources. The use of machine learning in the prediction of serious disease events using routine medical records has been successful in recent years. In this paper, a comparative analysis of different machine learning techniques that can accurately predict the occurrence of CHD events (...) from clinical data was performed. Four machine learning classifiers, namely Logistic Regression, Support Vector Machine (SVM), K- Nearest Neighbor (KNN), and Multi-Layer Perceptron (MLP) Neural Networks were identified and applied to a dataset of 462 medical instances and 9 features as well as the class feature from the South African Heart Disease data retrieved from the KEEL repository. The dataset consists of 302 records of healthy patients and 160 records of patients who suffer from CHD. In order to handle the imbalanced classification problem, the K-means algorithm along with Synthetic Minority Oversampling TEchnique (SMOTE) was used in this study. The empirical results of applying the four machine learning classifiers on the oversampled dataset have been very promising. The results reported using different evaluation metrics showed that SVM has achieved the highest overall prediction performance. (shrink)

Popular culture continues fuelling public imagination with things, human and non-human, that we might beco-me or confront. Besides robots, other significant tropes in popular fiction that generated images include non-human humans and cyborgs, wired into his-torically varying sociocultural realities. Robots and artificial intelligence are re-defining the natural order and its hierar-chical structure. This is not surprising, as natural order is always in flux, shaped by new scientific discoveries, especially the reading of the genetic code, that reveal and redefine relationships between (...) life forms. However, for the first time, a new, artificial, species is being introduced into the existing diagrams, and, for the first time, it appears to hybridise the anthro-pocentric natural world order. (shrink)

Artificial intelligence (AI) research enjoyed an initial period of enthusiasm in the 1970s and 80s. But this enthusiasm was tempered by a long interlude of frustration when genuinely useful AI applications failed to be forthcoming. Today, we are experiencing once again a period of enthusiasm, fired above all by the successes of the technology of deep neural networks or deep machine learning. In this paper we draw attention to what we take to be serious problems underlying current views of artificial (...) intelligence encouraged by these successes, especially in the domain of language processing. We then show an alternative approach to language-centric AI, in which we identify a role for philosophy. (shrink)