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.

Recent scholarship in philosophy of science and technology has shown that scientific and technological decision making are laden with values, including values of a social, political, and/or ethical character. This paper examines the role of value judgments in the design of machine-learning systems generally and in recidivism-prediction algorithms specifically. Drawing on work on inductive and epistemic risk, the paper argues that ML systems are value laden in ways similar to human decision making, because the development and design of ML systems (...) requires human decisions that involve tradeoffs that reflect values. In many cases, these decisions have significant—and, in some cases, disparate—downstream impacts on human lives. After examining an influential court decision regarding the use of proprietary recidivism-prediction algorithms in criminal sentencing, Wisconsin v. Loomis, the paper provides three recommendations for the use of ML in penal systems. (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)

The idea that human cognition is, or can be understood as, a form of computation is a useful conceptual tool for cognitive science. It was a foundational assumption during the birth of cognitive science as a multidisciplinary field, with Artificial Intelligence (AI) as one of its contributing fields. One conception of AI in this context is as a provider of computational tools (frameworks, concepts, formalisms, models, proofs, simulations, etc.) that support theory building in cognitive science. The contemporary field of AI, (...) however, has taken the theoretical possibility of explaining human cognition as a form of computation to imply the practical feasibility of realising human(-like or -level) cognition in factual computational systems, and the field frames this realisation as a short-term inevitability. Yet, as we formally prove herein, creating systems with human(-like or -level) cognition is intrinsically computationally intractable. This means that any factual AI systems created in the short-run are at best decoys. When we think these systems capture something deep about ourselves and our thinking, we induce distorted and impoverished images of ourselves and our cognition. In other words, AI in current practice is deteriorating our theoretical understanding of cognition rather than advancing and enhancing it. The situation could be remediated by releasing the grip of the currently dominant view on AI and by returning to the idea of AI as a theoretical tool for cognitive science. In reclaiming this older idea of AI, however, it is important not to repeat conceptual mistakes of the past (and present) that brought us to where we are today. (shrink)

Philosophy Compass, Volume 17, Issue 6, June 2022.

There has been a surge of interest in explainable artificial intelligence (XAI). It is commonly claimed that explainability is necessary for trust in AI, and that this is why we need it. In this paper, I argue that for some notions of trust it is plausible that explainability is indeed a necessary condition. But that these kinds of trust are not appropriate for AI. For notions of trust that are appropriate for AI, explainability is not a necessary condition. I thus (...) conclude that explainability is not necessary for trust in AI that matters. (shrink)

Models developed using machine learning are increasingly prevalent in scientific research. At the same time, these models are notoriously opaque. Explainable AI aims to mitigate the impact of opacity by rendering opaque models transparent. More than being just the solution to a problem, however, Explainable AI can also play an invaluable role in scientific exploration. This paper describes how post-hoc analytic techniques from Explainable AI can be used to refine target phenomena in medical science, to identify starting points for future (...) investigations of causal relationships, and to generate possible explanations of target phenomena in cognitive science. In this way, this paper describes how Explainable AI—over and above machine learning itself—contributes to the efficiency and scope of data-driven scientific research. (shrink)

Some machine learning models, in particular deep neural networks (DNNs), are not very well understood; nevertheless, they are frequently used in science. Does this lack of understanding pose a problem for using DNNs to understand empirical phenomena? Emily Sullivan has recently argued that understanding with DNNs is not limited by our lack of understanding of DNNs themselves. In the present paper, we will argue, _contra_ Sullivan, that our current lack of understanding of DNNs does limit our ability to understand with (...) DNNs. Sullivan’s claim hinges on which notion of understanding is at play. If we employ a weak notion of understanding, then her claim is tenable, but rather weak. If, however, we employ a strong notion of understanding, particularly explanatory understanding, then her claim is not tenable. (shrink)

Contemporary AI models owe much of their success and discontents to connectionism, a framework in cognitive science that has been (and continues to be) highly influential. Herein, we analyze artificial neural networks (ANNs): a) when used as scientific instruments of study; and b) when functioning as emergent arbiters of the zeitgeist in the cognitive, computational, and neural sciences. Building on our previous work with respect to analogizing between ANNs and cognition, brains, or behaviour (Guest & Martin, 2023), we use metatheoretical (...) analysis techniques (Guest, 2024), including formal logic, to characterise two distinct tendencies within connectionism that we dub classical and modern, with divergent properties, e.g. goals, mechanisms, scientific questions. We also demonstrate how we, as a field, often fail to follow important lines of argument to their end --- this results in a paradoxical praxis. By engaging more deeply with (meta)theory surrounding ANNs, our field can obviate the cycle of AI winters and summers, which need not be inevitable. (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)

The counterfactual approach to explainable AI (XAI) seeks to provide understanding of AI systems through the provision of counterfactual explanations. In a recent systematic review, Chou et al. (Inform Fus 81:59–83, 2022) argue that the counterfactual approach does not clearly provide causal understanding. They diagnose the problem in terms of the underlying framework within which the counterfactual approach has been developed. To date, the counterfactual approach has not been developed in concert with the approach for specifying causes developed by Pearl (...) (Causality: Models, reasoning, and inference. Cambridge University Press, 2000) and Woodward (Making things happen: A theory of causal explanation. Oxford University Press, 2003). In this paper, I build on Chou et al.’s work by applying the Pearl-Woodward approach. I argue that the standard counterfactual approach to XAI is capable of delivering causal understanding, but that there are limitations on its capacity to do so. I suggest a way to overcome these limitations. (shrink)

As machine learning has gradually entered into ever more sectors of public and private life, there has been a growing demand for algorithmic explainability. How can we make the predictions of complex statistical models more intelligible to end users? A subdiscipline of computer science known as interpretable machine learning (IML) has emerged to address this urgent question. Numerous influential methods have been proposed, from local linear approximations to rule lists and counterfactuals. In this article, I highlight three conceptual challenges that (...) are largely overlooked by authors in this area. I argue that the vast majority of IML algorithms are plagued by (1) ambiguity with respect to their true target; (2) a disregard for error rates and severe testing; and (3) an emphasis on product over process. Each point is developed at length, drawing on relevant debates in epistemology and philosophy of science. Examples and counterexamples from IML are considered, demonstrating how failure to acknowledge these problems can result in counterintuitive and potentially misleading explanations. Without greater care for the conceptual foundations of IML, future work in this area is doomed to repeat the same mistakes. (shrink)

In the natural and social sciences, it is common to use toy models—extremely simple and highly idealized representations—to understand complex phenomena. Some of the simple surrogate models used to understand opaque machine learning (ML) models, such as rule lists and sparse decision trees, bear some resemblance to scientific toy models. They allow non-experts to understand how an opaque ML model works globally via a much simpler model that highlights the most relevant features of the input space and their effect on (...) the output. The obvious difference is that the common target of a toy and a full-scale model in the sciences is some phenomenon in the world, while the target of a surrogate model is another model. This essential difference makes toy surrogate models (TSMs) a new object of study for theories of understanding, one that is not easily accommodated under current analyses. This paper provides an account of what it means to understand an opaque ML model globally with the aid of such simple models. (shrink)

Explainable artificial intelligence (XAI) is an emerging, multidisciplinary field of research that seeks to develop methods and tools for making AI systems more explainable or interpretable. XAI researchers increasingly recognise explainability as a context-, audience- and purpose-sensitive phenomenon, rather than a single well-defined property that can be directly measured and optimised. However, since there is currently no overarching definition of explainability, this poses a risk of miscommunication between the many different researchers within this multidisciplinary space. This is the problem we (...) seek to address in this paper. We outline a framework, called Explanatory Pragmatism, which we argue has two attractive features. First, it allows us to conceptualise explainability in explicitly context-, audience- and purpose-relative terms, while retaining a unified underlying definition of explainability. Second, it makes visible any normative disagreements that may underpin conflicting claims about explainability regarding the purposes for which explanations are sought. Third, it allows us to distinguish several dimensions of AI explainability. We illustrate this framework by applying it to a case study involving a machine learning model for predicting whether patients suffering disorders of consciousness were likely to recover consciousness. (shrink)

The contention that medical artificial intelligence (AI) should be ‘explainable’ is widespread in contemporary philosophy and in legal and best practice documents. Yet critics argue that ‘explainability’ is not a stable concept; non-explainable AI is often more accurate; mechanisms intended to improve explainability do not improve understanding and introduce new epistemic concerns; and explainability requirements are ad hoc where human medical decision-making is often opaque. A recent ‘political response’ to these issues contends that AI used in high-stakes scenarios, including medical (...) AI, must be explainable to meet basic standards of legitimacy: People are owed reasons for decisions that impact their vital interests, and this requires explainable AI. This article demonstrates why the political response fails. Attending to systemic considerations, as its proponents desire, suggests that the political response is subject to the same criticisms as other arguments for explainable AI and presents new issues. It also suggests that decision-making about non-explainable medical AI can meet public reason standards. The most plausible version of the response amounts to a simple claim that public reason demands reasons why AI is permitted. But that does not actually support explainable AI or respond to criticisms of strong requirements for explainable medical AI. (shrink)

Deep neural networks are said to be opaque, impeding the development of safe and trustworthy artificial intelligence, but where this opacity stems from is less clear. What are the sufficient properties for neural network opacity? Here, I discuss five common properties of deep neural networks and two different kinds of opacity. Which of these properties are sufficient for what type of opacity? I show how each kind of opacity stems from only one of these five properties, and then discuss to (...) what extent the two kinds of opacity can be mitigated by explainability methods. (shrink)

Unsupervised learning algorithms are widely used for many important statistical tasks with numerous applications in science and industry. Yet despite their prevalence, they have attracted remarkably little philosophical scrutiny to date. This stands in stark contrast to supervised and reinforcement learning algorithms, which have been widely studied and critically evaluated, often with an emphasis on ethical concerns. In this article, I analyze three canonical unsupervised learning problems: clustering, abstraction, and generative modeling. I argue that these methods raise unique epistemological and (...) ontological questions, providing data-driven tools for discovering natural kinds and distinguishing essence from contingency. This analysis goes some way toward filling the lacuna in contemporary philosophical discourse on unsupervised learning, as well as bringing conceptual unity to a heterogeneous field more often described by what it isnot(i.e., supervised or reinforcement learning) than by what itis. I submit that unsupervised learning is not just a legitimate subject of philosophical inquiry but perhaps the most fundamental branch of all AI. However, an uncritical overreliance on unsupervised methods poses major epistemic and ethical risks. I conclude by advocating for a pragmatic, error-statistical approach that embraces the opportunities and mitigates the challenges posed by this powerful class of algorithms. (shrink)

This paper develops an account of how the implementation of ML models into healthcare settings requires revising the methodological apparatus of philosophical bioethics. On this account, ML models are cognitive interventions that provide decision-support to physicians and patients. Due to reliability issues, opaque reasoning processes, and information asymmetries, ML models pose inferential problems for them. These inferential problems lay the grounds for many ethical problems that currently claim centre-stage in the bioethical debate. Accordingly, this paper argues that the best way (...) to make progress in remedying these ethical problems is to distil their epistemic core and to identify appropriate epistemic norms as guardrails. The viability of this approach will be highlighted based on four key issues: trust, responsibility, paternalism, and fairness. In that respect, the paper also contributes to a more pronounced view of the specific methodological challenges for ethical theorizing, when the point of reference are complex statistical models that make predictions, rather than individual agents, acting for reasons. (shrink)

I argue that ML models used in science function as highly idealized toy models. If we treat ML models as a type of highly idealized toy model, then we can deploy standard representational and epistemic strategies from the toy model literature to explain why ML models can still provide epistemic success despite their lack of similarity to their targets.

Artificial intelligent (AI) systems used in medicine are often very reliable and accurate, but at the price of their being increasingly opaque. This raises the question whether a system’s opacity undermines the ability of medical doctors to acquire knowledge on the basis of its outputs. We investigate this question by focusing on a case in which a patient’s risk of recurring breast cancer is predicted by an opaque AI system. We argue that, given the system’s opacity, as well as the (...) possibility of malfunctioning AI systems, practitioners’ inability to check the correctness of their outputs, and the high stakes of such cases, the knowledge of medical practitioners is indeed undermined. They are lucky to form true beliefs based on the AI systems’ outputs, and knowledge is incompatible with luck. We supplement this claim with a specific version of the safety condition on knowledge, Safety*. We argue that, relative to the perspective of the medical doctor in our example case, his relevant beliefs could easily be false, and this despite his evidence that the AI system functions reliably. Assuming that Safety* is necessary for knowledge, the practitioner therefore doesn’t know. We address three objections to our proposal before turning to practical suggestions for improving the epistemic situation of medical doctors. (shrink)

AI technologies have recently shown remarkable capabilities in various scientific fields, such as drug discovery, medicine, climate modeling, and archaeology, primarily through their pattern recognition abilities. They can also generate hypotheses and suggest new research directions. While acknowledging AI’s potential to aid in scientific breakthroughs, the paper shows that current AI models do not meet the criteria for making independent scientific discoveries. Discovery is an epistemic achievement that requires a level of competence and self-reflectivity that AI does not yet possess.

In climate science, climate models are one of the main tools for understanding phenomena. Here, we develop a framework to assess the fitness of a climate model for providing understanding. The framework is based on three dimensions: representational accuracy, representational depth, and graspability. We show that this framework does justice to the intuition that classical process-based climate models give understanding of phenomena. While simple climate models are characterized by a larger graspability, state-of-the-art models have a higher representational accuracy and representational (...) depth. We then compare the fitness-for-providing understanding of process-based to data-driven models that are built with machine learning. We show that at first glance, data-driven models seem either unnecessary or inadequate for understanding. However, a case study from atmospheric research demonstrates that this is a false dilemma. Data-driven models can be useful tools for understanding , specifically for phenomena for which scientists can argue from the coherence of the models with background knowledge to their representational accuracy and for which the model complexity can be reduced such that they are graspable to a satisfactory extent. When citing this paper, please use the full journal title Studies in History and Philosophy of Science. (shrink)

Machine learning methods have recently created high expectations in the climate modelling context in view of addressing climate change, but they are often considered as non-physics-based ‘black boxes’ that may not provide any understanding. However, in many ways, understanding seems indispensable to appropriately evaluate climate models and to build confidence in climate projections. Relying on two case studies, we compare how machine learning and standard statistical techniques affect our ability to understand the climate system. For that purpose, we put five (...) evaluative criteria of understanding to work: intelligibility, representational accuracy, empirical accuracy, coherence with background knowledge, and assessment of the domain of validity. We argue that the two families of methods are part of the same continuum where these various criteria of understanding come in degrees, and that therefore machine learning methods do not necessarily constitute a radical departure from standard statistical tools, as far as understanding is concerned. (shrink)

There is a current debate about if, and in what sense, machine learning systems used in the medical context need to be explainable. Those arguing in favor contend these systems require post hoc explanations for each individual decision to increase trust and ensure accurate diagnoses. Those arguing against suggest the high accuracy and reliability of the systems is sufficient for providing epistemic justified beliefs without the need for explaining each individual decision. But, as we show, both solutions have limitations—and it (...) is unclear either address the epistemic worries of the medical professionals using these systems. We argue these systems do require an explanation, but an institutional explanation. These types of explanations provide the reasons why the medical professional should rely on the system in practice—that is, they focus on trying to address the epistemic concerns of those using the system in specific contexts and specific occasions. But ensuring that these institutional explanations are fit for purpose means ensuring the institutions designing and deploying these systems are transparent about the assumptions baked into the system. This requires coordination with experts and end-users concerning how it will function in the field, the metrics used to evaluate its accuracy, and the procedures for auditing the system to prevent biases and failures from going unaddressed. We contend this broader explanation is necessary for either post hoc explanations or accuracy scores to be epistemically meaningful to the medical professional, making it possible for them to rely on these systems as effective and useful tools in their practices. (shrink)

This paper attempts to describe and address a specific puzzle related to compositionality in artificial networks such as Deep Neural Networks and machine learning in general. The puzzle identified here touches on a larger debate in Artificial Intelligence related to epistemic opacity but specifically focuses on computational applications of human level linguistic abilities or properties and a special difficulty with relation to these. Thus, the resulting issue is both general and unique. A partial solution is suggested.

I examine the construction and evaluation of machine learning binary classification models. These models are increasingly used for societal applications such as classifying patients into two categories according to the presence or absence of a certain disease like cancer and heart disease. I argue that the construction of ML classification models involves an optimisation process aiming at the minimization of the inductive risk associated with the intended uses of these models. I also argue that the construction of these models is (...) underdetermined by the available data, and that this makes it necessary for ML modellers to make social value judgments in determining the error costs used in ML optimization. I thus suggest that the assessment of the inductive risk with respect to the social values of the intended users is an integral part of the construction and evaluation of ML classification models. I also discuss the implications of this conclusion for the philosophical debate concerning inductive risk. (shrink)

By looking at it you come to know that a thing is an apple. How? A natural answer is that this is down to how it looks – its superficial visual appearance. Looks Views treat our acquaintance with such looks as accounting for how visual knowledge is secured. Here I argue that for many pairings of properties and perceivers Looks Views will turn out not to work. We can visually track many properties through huge variation in things’ visual appearances. For (...) such properties no kind of look will play the explanatory role Looks Views demanded. Sometimes, then, you might secure visual knowledge that something is an apple but not by way of its superficial visual appearance. That conclusion may seem difficult to swallow. I argue that at least many of our intuitions about the roles of looks can nonetheless be assuaged. (shrink)

To learn about real world phenomena, scientists have traditionally used models with clearly interpretable elements. However, modern machine learning (ML) models, while powerful predictors, lack this direct elementwise interpretability (e.g. neural network weights). Interpretable machine learning (IML) offers a solution by analyzing models holistically to derive interpretations. Yet, current IML research is focused on auditing ML models rather than leveraging them for scientific inference. Our work bridges this gap, presenting a framework for designing IML methods—termed ’property descriptors’—that illuminate not just (...) the model, but also the phenomenon it represents. We demonstrate that property descriptors, grounded in statistical learning theory, can effectively reveal relevant properties of the joint probability distribution of the observational data. We identify existing IML methods suited for scientific inference and provide a guide for developing new descriptors with quantified epistemic uncertainty. Our framework empowers scientists to harness ML models for inference, and provides directions for future IML research to support scientific understanding. (shrink)

Manifold learning and effective model building are generally viewed as fundamentally different types of procedure. After all, in one we build a simplified model of the data, in the other, we construct a simplified model of the another model. Nonetheless, I argue that certain kinds of high-dimensional effective model building, and effective field theory construction in quantum field theory, can be viewed as special cases of manifold learning. I argue that this helps to shed light on all of these techniques. (...) First, it suggests that the effective model building procedure depends upon a certain kind of algorithmic compressibility requirement. All three approaches assume that real-world systems exhibit certain redundancies, due to regularities. The use of these regularities to build simplified models is essential for scientific progress in many different domains. (shrink)

This paper investigates how unsupervised machine learning methods might make hermeneutic interpretive text analysis more objective in the social sciences. Through a close examination of the uses of topic modeling—a popular unsupervised approach in the social sciences—it argues that the primary way in which unsupervised learning supports interpretation is by allowing interpreters to discover unanticipated information in larger and more diverse corpora and by improving the transparency of the interpretive process. This view highlights that unsupervised modeling does not eliminate the (...) researchers’ judgments from the process of producing evidence for social scientific theories. The paper shows this by distinguishing between two prevalent attitudes toward topic modeling, i.e., topic realism and topic instrumentalism. Under neither can modeling provide social scientific evidence without the researchers’ interpretive engagement with the original text materials. Thus the unsupervised text analysis cannot improve the objectivity of interpretation by alleviating the problem of underdetermination in interpretive debate. The paper argues that the sense in which unsupervised methods can improve objectivity is by providing researchers with the resources to justify to others that their interpretations are correct. This kind of objectivity seeks to reduce suspicions in collective debate that interpretations are the products of arbitrary processes influenced by the researchers’ idiosyncratic decisions or starting points. The paper discusses this view in relation to alternative approaches to formalizing interpretation and identifies several limitations on what unsupervised learning can be expected to achieve in terms of supporting interpretive work. (shrink)

Are intelligent machines really intelligent? Is the underlying philosoph- ical concept of intelligence satisfactory for describing how the present systems work? Is understanding a necessary and sufficient condition for intelligence? If a machine could understand, should we attribute subjectivity to it? This paper addresses the problem of deciding whether the so-called ”intelligent machines” are capable of understanding, instead of merely processing signs. It deals with the relationship between syntax and semantics. The main thesis concerns the inevitability of semantics for any (...) discussion about the possibility of building conscious machines, condensed into the following two tenets: ”If a machine is capable of understanding (in the strong sense), then it must be capable of combining rules and intuitions”; “If semantics cannot be reduced to syntax, then a machine cannot understand.” Our conclusion states that it is not necessary to attribute understanding to a machine in order to explain its exhibited “intelligent” behavior; a merely syntactic and mechanistic approach to intelligence as a task-solving tool suffices to justify the range of operations that it can display in the current state of technological development. (shrink)

This paper approaches the interaction of a health professional with an AI system for diagnostic purposes as a hybrid decision making process and conceptualizes epistemo-ethical constraints on this process. We argue for the importance of the understanding of the underlying machine epistemology in order to raise awareness of and facilitate realistic expectations from AI as a decision support system, both among healthcare professionals and the potential benefiters. Understanding the epistemic abilities and limitations of such systems is essential if we are (...) to integrate AI into the decision making processes in a way that takes into account its applicability boundaries. This will help to mitigate potential harm due to misjudgments and, as a result, to raise the trust—understood here as a belief in reliability of—in the AI system. We aim at a minimal requirement for AI meta-explanation which should distinguish machine epistemic processes from similar processes in human epistemology in order to avoid confusion and error in judgment and application. An informed approach to the integration of AI systems into the decision making for diagnostic purposes is crucial given its high impact on health and well-being of patients. (shrink)

This paper examines two prominent formal trade-offs in artificial intelligence (AI)---between predictive accuracy and fairness, and between predictive accuracy and interpretability. These trade-offs have become a central focus in normative and regulatory discussions as policymakers seek to understand the value tensions that can arise in the social adoption of AI tools. The prevailing interpretation views these formal trade-offs as directly corresponding to tensions between underlying social values, implying unavoidable conflicts between those social objectives. In this paper, I challenge that prevalent (...) interpretation by introducing a sociotechnical approach to examining the value implications of trade-offs. After providing a set-theoretic understanding of these model trade-offs, I argue that their implications cannot be understood in isolation from the technical, psychological, organizational, and social factors that shape the integration of AI models into decision-making pipelines. Specifically, I identify three key considerations---validity and instrumental relevance, compositionality, and dynamics---for contextualizing and characterizing these implications. These considerations reveal that the relationship between model trade-offs and corresponding values depends on critical choices and assumptions. Crucially, judicious sacrifices in one model property for another can, in fact, promote both sets of corresponding values. The proposed sociotechnical perspective thus shows that we can and should aspire to higher epistemic and ethical possibilities than the prevalent interpretation suggests, while offering practical guidance for achieving those outcomes. Finally, I draw out the broader implications of this perspective for AI design and governance, highlighting the need to broaden normative engagement across the AI lifecycle, develop legal and auditing tools sensitive to sociotechnical considerations, and rethink the vital role and appropriate structure of interdisciplinary collaboration in fostering a responsible AI workforce. (shrink)

Machine learning (ML) models recently led to major breakthroughs in predictive tasks in the natural sciences. Yet their benefits for the social sciences are less evident, as even high-profile studies on the prediction of life trajectories have shown to be largely unsuccessful – at least when measured in traditional criteria of scientific success. This paper tries to shed light on this remarkable performance gap. Comparing two social science case studies to a paradigm example from the natural sciences, we argue that, (...) in addition to explanation, prediction is an important goal of social science – and we identify constraints that impede pure ML prediction from being successful in that field. As a remedy, we outline elements of an integrative modelling approach that combines explanatory models and predictive ML models. (shrink)

The practice of science appears to involve “model-talk”. Scientists, one thinks, are in the business of giving accounts of reality. Scientists, in the process of furnishing such accounts, talk about what they call “models”. Philosophers of science have inspected what this talk of models suggests about how scientific theories manage to represent reality. There are, it seems, at least three distinct philosophical views on the role of scientific models in science’s portrayal of reality: the abstractionist view, the indirect fictionalist view, (...) and the direct fictionalist view. In this essay, I try to articulate a question about what makes a scientific model more or less appropriate for a specific domain of reality. More precisely, I ask, “What accounts for the fact that given a determinate target domain, some scientific models, but not others, are thought to be “appropriate” for that domain?” I then consider whether and the degree to which each of the mentioned views on scientific models institutes a satisfactory response to this question. I conclude that, amongst those views, the direct fictionalist view seems to have the most promising response. I then utilize this argument to develop a more precise account of the problem of differential importability, and ultimately offer a more general and less presumptive argument that the problem seems to be optimally solved by justifying comparative evaluation of model-importabilities solely in terms of comparative evaluations of what I characterize as models’ “holistic” predictive success. (shrink)

This paper examines two prominent formal trade-offs in artificial intelligence (AI)---between predictive accuracy and fairness, and between predictive accuracy and interpretability. These trade-offs have become a central focus in normative and regulatory discussions as policymakers seek to understand the value tensions that can arise in the social adoption of AI tools. The prevailing interpretation views these formal trade-offs as directly corresponding to tensions between underlying social values, implying unavoidable conflicts between those social objectives. In this paper, I challenge that prevalent (...) interpretation by introducing a sociotechnical approach to examining the value implications of trade-offs. After providing a set-theoretic understanding of these model trade-offs, I argue that their implications cannot be understood in isolation from the technical, psychological, organizational, and social factors that shape the integration of AI models into decision-making pipelines. Specifically, I identify three key considerations---validity and instrumental relevance, compositionality, and dynamics---for contextualizing and characterizing these implications. These considerations reveal that the relationship between model trade-offs and corresponding values depends on critical choices and assumptions. Crucially, judicious sacrifices in one model property for another can, in fact, promote both sets of corresponding values. The proposed sociotechnical perspective thus shows that we can and should aspire to higher epistemic and ethical possibilities than the prevalent interpretation suggests, while offering practical guidance for achieving those outcomes. Finally, I draw out the broader implications of this perspective for AI design and governance, highlighting the need to broaden normative engagement across the AI lifecycle, develop legal and auditing tools sensitive to sociotechnical considerations, and rethink the vital role and appropriate structure of interdisciplinary collaboration in fostering a responsible AI workforce. (shrink)

We claim that scientists working with deep learning (DL) models exhibit a form of pragmatic understanding that is not reducible to or dependent on explanation. This pragmatic understanding comprises a set of learned methodological principles that underlie DL model design-choices and secure their reliability. We illustrate this action-oriented pragmatic understanding with a case study of AlphaFold2, highlighting the interplay between background knowledge of a problem and methodological choices involving techniques for constraining how a model learns from data. Building successful models (...) requires pragmatic understanding to apply modelling strategies that encourage the model to learn data patterns that will facilitate reliable generalisation. (shrink)

Evolutionary explanations are not only common in the biological sciences, but also widespread outside biology. But an account of how evolutionary explanations perform their explanatory work is still lacking. This paper develops such an account. I argue that available accounts of explanations in evolutionary science miss important parts of the role of history in evolutionary explanations. I argue that the historical part of evolutionary science should be taken as having genuine explanatory force, and that it provides how‐possibly explanations sensu Dray. (...) I propose an account of evolutionary explanations as comparative‐composite explanations consisting of two distinct kinds of explanations, one processual and one historical, that are connected via the explanandum's evolvability to show how the explanandum is the product of its evolutionary past. The account is both a reconstruction of how evolutionary explanations in biology work and a guideline specifying what kind of explanations evolutionary research programs should develop. (shrink)

Artificial Intelligence and Machine Learning (AI/ML) models introduce unique considerations when determining their epistemic value. Fortunately, existing work on the epistemic features of AI/ML can...

It is now well-appreciated by philosophers that contemporary large-scale ‘-omics’ studies in biology stand in non-trivial relationships to more orthodox hypothesis-driven approaches. These relationships have been clarified by Ratti (2015); however, there remains much more to be said regarding how an important field of genomics cited in that work—‘genome-wide association studies’ (GWAS)—fits into this framework. In the present article, we propose a revision to Ratti’s framework more suited to studies such as GWAS. In the process of doing so, we introduce (...) to the philosophical literature novel exploratory experiments in (phospho)proteomics, and demonstrate how these experiments interplay with the above considerations. (shrink)

Bioethical consensus promotes a shared decision making model, which requires healthcare professionals to partner their knowledge with that of their patients—who, at a minimum, contribute their valu...

It has often been argued that we face a trade-off between accuracy and opacity in deep learning models. The idea is that we can only harness the accuracy of deep learning models by simultaneously accepting that the grounds for the models’ decision-making are epistemically opaque to us. In this paper, we ask the following question: what are the prospects of making deep learning models transparent without compromising on their accuracy? We argue that the answer to this question depends on which (...) kind of opacity we have in mind. If we focus on the standard notion of opacity, which tracks the internal complexities of deep learning models, we argue that existing explainable AI (XAI) techniques show us that the prospects look relatively good. But, as it has recently been argued in the literature, there is another notion of opacity that concerns factors external to the model. We argue that there are at least two types of external opacity—link opacity and structure opacity—and that existing XAI techniques can to some extent help us reduce the former but not the latter. (shrink)

Artificial testimony generated by large language models (LLMs) can be a source of knowledge. However, the requirement that artificial testifiers must satisfy for successful knowledge acquisition is different from the requirement that human testifiers must satisfy. Correspondingly, the epistemic ground of artificial testimonial knowledge is not the well-known and accepted ones suggested by renowned epistemological theories of (human) testimony. Based on Thomas Reid’s old teaching, we suggest a novel epistemological theory of artificial testimony that for receivers to justifiably believe artificially (...) generated statements, testifiers of the statement should robustly perform the propensities of veracity and cautiousness. The theory transforms the weakness of Reid’s view to an advantage of its own. It sets an achievable standard for LLMs and clarifies the improvement that current LLMs should make for meeting the standard. Moreover, it indicates a pluralistic nature of testimonial justification pertaining to the pluralistic nature of possible testifiers for knowledge transmission. (shrink)

As the demand for explainable deep learning grows in the evaluation of language technologies, the value of a principled grounding for those explanations grows as well. Here we study the state-of-the-art in explanation for neural models for natural-language processing (NLP) tasks from the viewpoint of philosophy of science. We focus on recent evaluation work that finds brittleness in explanations obtained through attention mechanisms.We harness philosophical accounts of explanation to suggest broader conclusions from these studies. From this analysis, we assert the (...) impossibility of causal explanations from attention layers over text data. We then introduce NLP researchers to contemporary philosophy of science theories that allow robust yet non-causal reasoning in explanation, giving computer scientists a vocabulary for future research. (shrink)

The philosophical debate around the impact of machine learning in science is often framed in terms of a choice between AI and classical methods as mutually exclusive alternatives involving difficult epistemological trade-offs. A common worry regarding machine learning methods specifically is that they lead to opaque models that make predictions but do not lead to explanation or understanding. Focusing on the field of molecular biology, I argue that in practice machine learning is often used with explanatory aims. More specifically, I (...) argue that machine learning can be tightly integrated with other, more traditional, research methods and in a clear sense can contribute to insight into the causal processes underlying phenomena of interest to biologists. One could even say that machine learning is not the end of theory in important areas of biology, as has been argued, but rather a new beginning. I support these claims with a detailed discussion of a case study involving gene regulation by microRNAs. (shrink)