This paper offers an account of data manipulation in scientific experiments. It will be shown that in many cases raw, unprocessed data is not produced, but rather a form of processed data that will be referred to as a data model. The language of data models will be used to provide a framework within which to understand a recent debate about the status of data and data manipulation. It will be seen that (...) a description in terms of data models allows one to understand cases in which data acquisition and data manipulation cannot be separated into two independent activities. (shrink)

While Bachman's role data model is often cited, it appears that its contribution, the introduction of role types and their conception as unions of entity types that occupy the places of relationship types, has mostly been ignored. This is unfortunate since it has led to countless reinventions of the wheel, and sometimes even to regress. With this homage, the author wishes to shed some light on the natural elegance of Bachman's role concept, and to make clear why he believes (...) that it is valid until this day. (shrink)

We critically engage two traditional views of scientific data and outline a novel philosophical view that we call the pragmatic-representational view of data. On the PR view, data are representations that are the product of a process of inquiry, and they should be evaluated in terms of their adequacy or fitness for particular purposes. Some important implications of the PR view for data assessment, related to misrepresentation, context-sensitivity, and complementary use, are highlighted. The PR view provides (...) insight into the common but little-discussed practices of iteratively reusing and repurposing data, which result in many datasets’ having a phylogeny—an origin and complex evolutionary history—that is relevant to their evaluation and future use. We relate these insights to the open-data and data-rescue movements, and highlight several future avenues of research that build on the PR view of data. (shrink)

Many decades ago Patrick Suppes argued rather convincingly that theoretical hypotheses are not confronted with the direct, raw results of an experiment, rather, they are typically compared with models of data. What exactly is a data model however? And how do the interactions of particles at the subatomic scale give rise to the huge volumes of data that are then moulded into a polished data model? The aim of this paper is to answer these questions (...) by presenting a detailed case study of the construction of data models at the LHCb for testing Lepton Flavour Universality in rare decays of B-mesons. The close examination of the scientific practice at the LHCb leads to the following four main conclusions: raw data in their pure form are practically useless for the comparison of experimental results with theory, and processed data are in some cases epistemically more reliable, real and simulated data are involved in the co-production of the final data model and cannot be easily distinguished, theory-ladenness emerges at three different levels depending on the scope and the purpose for which background theory guides the overall experimental process and the overall process of acquiring and analysing data in high energy physics is too complicated to be fully captured by a generic methodological description of the experimental practice. (shrink)

Translation abstractZusammenfassung: Daten, Modelle und Erdgeschichte ineinander gefaltet: Paläo‐ Simulationen und ihre epistemologische Unruhe. Klima‐ und Erdsystemmodelle werden nicht nur verwendet, um künftige klimatische Bedingungen zu prognostizieren, sondern auch, um vergangene Klima‐Ereignisse zu rekonstruieren. Dieser Beitrag ist der zweite in einer Reihe, welche die Paläoklimatologie – die Wissenschaft der Klimate vor Anbeginn direkter, instrumentenbasierter Messungen – als eine epistemisch radikale Praxis vorstellt, die in direkter und offener Weise die Unterscheidung zwischen Daten und Modell aufhebt sowie den Begriff des Experiments rekonfiguriert (...) und im Zuge dessen unsere Vorstellungen einer tiefen, d. h. geologischen, und einer flachen, d. h. historischen Zeit miteinander vermengt. Der vorliegende Artikel zeigt erstens, wie sich im Zuge der Einführung von General Circulation Models nicht nur die Wissensproduktion über die komplexen Ursachen und Abläufe früherer Klimawandel‐Ereignisse auf numerische Simulationen verlagerte, sondern diese Einführung selbst von paläoklimatologischen Fragen begleitet, wenn nicht gar eminent beeinflusst wurde. Zweitens erweisen sich die numerischen Experimente inzwischen als potente Katalysatoren eines neuen Verständnisses für die spezifische Zeitlichkeit, in der sich unsere Gegenwart im Übergang zum Anthropozän entfaltet. In der Diskussion der historischen Entwicklung der Paläoklimamodellierung und durch die Untersuchung der produktiven heuristischen Eigenschaften ihrer Praxis stellt der Artikel unsere Fixierung auf die Kategorie der Unsicherheit infrage. (shrink)

Information is considered to be the major part of an organization. With the enhancement of technology, the knowledge level is increasing with the passage of time. This increase of information is in volume, velocity, and variety. Extracting meaningful insights is the dire need of an individual from such information and knowledge. Visualization is a key tool and has become one of the most significant platforms for interpreting, extracting, and communicating information. The current study is an endeavour toward data modelling (...) and user knowledge by using a rough set approach for extracting meaningful insights. The technique has used different rough set algorithms such as K-nearest neighbours, decision rules, decomposition tree, and local transfer function classifier for an experimental setup. The approach has found its accuracy for the optimal use of data modelling and user knowledge. The experimental setup of the proposed method is validated by using the dataset available in the UCI web repository. Results of the proposed study show that the model is effective and efficient with an accuracy of 96% for KNN, 87% for decision rules, 91% for decision trees, 85.04% for cross validation architecture, and 94.3% for local transfer function classifier. The validity of the proposed classification algorithms is tested using different performance metrics such as F-score, precision, accuracy, recall, specificity, and misclassification rates. For all these performance metrics, the KNN classifier outperformed, and this high performance shows the applicability of the KNN classifier in the proposed problem. (shrink)

In modern day technology, the level of knowledge is increasing day by day. This increase is in terms of volume, velocity, and variety. Understanding of such knowledge is a dire need of an individual to extract meaningful insight from it. With the advancement in computer and image-based technologies, visualization becomes one of the most significant platforms to extract, interpret, and communicate information. In data modelling, visualization is the process of extracting knowledge to reveal the detail data structure and (...) process of data. The proposed study aim is to know about the user knowledge, data modelling, and visualization by handling through the fuzzy logic-based approach. The experimental setup is validated through the data user modelling dataset available in the UCI web repository. The results show that the model is effective and efficient in situations where uncertainty and complexity arise. (shrink)

Predictive risk modelling is a computational method used to generate probabilities correlating events. The output of such systems is typically represented by a statistical score derived from various related and often arbitrary datasets. In many cases, the information generated by such systems is treated as a form of evidence to justify further action. This paper examines the nature of the information generated by such systems and compares it with more orthodox notions of evidence found in epistemology. The paper focuses on (...) a specific example to illustrate the issues: The New Zealand Government has proposed implementing a predictive risk modelling system which purportedly identifies children at risk of a maltreatment event before the age of five. Timothy Williamson’s conception of epistemology places a requirement on knowledge that it be explanatory. Furthermore, Williamson argues that knowledge is equivalent to evidence. This approach is compared to the claim that the output of such computational systems constitutes evidence. While there may be some utility in using predictive risk modelling systems, I argue, since an explanatory account of the output of such algorithms that meets Williamson’s requirements cannot be given, doubt is cast upon the resulting statistical scores as constituting evidence on generally accepted epistemic grounds. The algorithms employed in such systems are geared towards identifying patterns which turn out to be good correlations. However, rather than providing information about specific individuals and their exposure to risk, a more valid explanation of a high probability score is that the particular variables related to incidents of maltreatment are just higher amongst certain subgroups in a population than they are amongst others. The paper concludes that any justification of the information generated by such systems is generalised and pragmatic at best and the application of this information to individual cases raises various ethical issues. (shrink)

This article describes some potential uses of Bayesian estimation for time-series and panel data models by incorporating information from prior probabilities (i.e., priors) in addition to observed data. Drawing on econometrics and other literatures we illustrate the use of informative “shrinkage” or “small variance” priors (including so-called “Minnesota priors”) while extending prior work on the general cross-lagged panel model (GCLM). Using a panel dataset of national income and subjective well-being (SWB) we describe three key benefits of these (...) priors. First, they shrink parameter estimates toward zero or toward each other for time-varying parameters, which lends additional support for an income → SWB effect that is not supported with maximum likelihood (ML). This is useful because, second, these priors increase model parsimony and the stability of estimates (keeping them within more reasonable bounds) and thus improve out-of-sample predictions and interpretability, which means estimated effect should also be more trustworthy than under ML. Third, these priors allow estimating otherwise under-identified models under ML, allowing higher-order lagged effects and time-varying parameters that are otherwise impossible to estimate using observed data alone. In conclusion we note some of the responsibilities that come with the use of priors which, departing from typical commentaries on their scientific applications, we describe as involving reflection on how best to apply modeling tools to address matters of worldly concern. (shrink)

Despite an enormous philosophical literature on models in science, surprisingly little has been written about data models and how they are constructed. In this paper, I examine the case of how paleodiversity data models are constructed from the fossil data. In particular, I show how paleontologists are using various model-based techniques to correct the data. Drawing on this research, I argue for the following related theses: first, the ‘purity’ of a data model (...) is not a measure of its epistemic reliability. Instead it is the fidelity of the data that matters. Second, the fidelity of a data model in capturing the signal of interest is a matter of degree. Third, the fidelity of a data model can be improved ‘vicariously’, such as through the use of post hoc model-based correction techniques. And, fourth, data models, like theoretical models, should be assessed as adequate for particular purposes. (shrink)

Several quantitative techniques for choosing among data models are available. Among these are techniques based on algorithmic information theory, minimum description length theory, and the Akaike information criterion. All these techniques are designed to identify a single model of a data set as being the closest to the truth. I argue, using examples, that many data sets in science show multiple patterns, providing evidence for multiple phenomena. For any such data set, there is more than (...) one data model that must be considered close to the truth. I conclude that, since the established techniques for choosing among data models are unequipped to handle these cases, they cannot be regarded as adequate. ‡I presented a previous version of this paper at the 20th Biennial Meeting of the Philosophy of Science Association, Vancouver, November 2006. I am grateful to the audience for constructive discussion. I thank Leiden University students Marjolein Eysink Smeets and Lenneke Schrier for suggesting the cortisol example, and Remko van der Geest for comments on a draft. †To contact the author, please write to: Faculty of Philosophy, University of Leiden, P.O. Box 9515, 2300 RA Leiden, The Netherlands; e-mail: [email protected]. (shrink)

Subjective experience is transformed into objective reality for societal members through cultural idea systems that can be represented with theory and data models. A theory model shows relationships and their logical implications that structure a cultural idea system. A data model expresses patterning found in ethnographic observations regarding the behavioral implementation of cultural idea systems. An example of this duality for modeling cultural idea systems is illustrated with Arabic proverbs that structurally link friend and enemy as concepts (...) through a culturally defined computational system. Computational systems also generate new concepts, as will be illustrated through a theory model for the structure of a .. (shrink)

The distinction between data and phenomena introduced by Bogen and Woodward (Philosophical Review 97(3):303–352, 1988) was meant to help accounting for scientific practice, especially in relation with scientific theory testing. Their article and the subsequent discussion is primarily viewed as internal to philosophy of science. We shall argue that the data/phenomena distinction can be used much more broadly in modelling processes in philosophy.

The article examines four models of data governance emerging in the current platform society. While major attention is currently given to the dominant model of corporate platforms collecting and economically exploiting massive amounts of personal data, other actors, such as small businesses, public bodies and civic society, take also part in data governance. The article sheds light on four models emerging from the practices of these actors: data sharing pools, data cooperatives, public (...) class='Hi'>data trusts and personal data sovereignty. We propose a social science-informed conceptualisation of data governance. Drawing from the notion of data infrastructure we identify the models as a function of the stakeholders’ roles, their interrelationships, articulations of value, and governance principles. Addressing the politics of data, we considered the actors’ competitive struggles for governing data. This conceptualisation brings to the forefront the power relations and multifaceted economic and social interactions within data governance models emerging in an environment mainly dominated by corporate actors. These models highlight that civic society and public bodies are key actors for democratising data governance and redistributing value produced through data. Through the discussion of the models, their underpinning principles and limitations, the article wishes to inform future investigations of socio-technical imaginaries for the governance of data, particularly now that the policy debate around data governance is very active in Europe. (shrink)

A novel data-driven model-free adaptive control approach is first proposed by combining the advantages of model-free adaptive control and data-driven optimal iterative learning control, and then its stability and convergence analysis is given to prove algorithm stability and asymptotical convergence of tracking error. Besides, the parameters of presented approach are adaptively adjusted with fuzzy logic to determine the occupied proportions of MFAC and DDOILC according to their different control performances in different control stages. Lastly, the proposed fuzzy DDMFAC (...) approach is applied to the control of particle quality in drug development phase of spray fluidized-bed granulation process, and its control effect is compared with MFAC and DDOILC and their fuzzy forms, in which the parameters of MFAC and DDOILC are adaptively adjusted with fuzzy logic. The effectiveness of the presented FDDMFAC approach is verified by a series of simulations. (shrink)

Several key areas in modeling the cardiovascular and respiratory control systems are reviewed and examples are given which reflect the research state of the art in these areas. Attention is given to the interrelated issues of data collection, experimental design, and model application including model development and analysis. Examples are given of current clinical problems which can be examined via modeling, and important issues related to model adaptation to the clinical setting.

With the rapid and unbalanced development of industry, a large amount of cultivated land is converted into industrial land with lower efficiency. The existing research is extensively concerned with industrial land use and industrial development in isolation, but little attention has been paid to the relationship between them. To help address this gap, the paper creates a new efficiency measure method for industrial land use combining Subvector Data Envelope Analysis with spatial analysis approach. The proposed model has been verified (...) by using the industrial land use data of 30 Chinese provinces from 2001 to 2013. The spatial autocorrelation relationship between industrial development and industrial land use efficiency is explored. Furthermore, this paper examines the effects of industrial development on industrial land use efficiency by spatial panel data model. The results indicate that the industrial land use efficiency and the industrial development level in the provinces of eastern region are higher than those of the western region. The spatial distribution of industrial land use efficiency shows remarkable positive spatial autocorrelation. However, the level of industrial development has obvious negative spatial autocorrelation since 2009. The improvement of industrial development has a significant positive impact on the industrial land use efficiency. (shrink)

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)

The contribution of Big Data to social science is not limited to data availability but includes the introduction of analytical approaches that have been developed in computer science, and in particular in machine learning. This brings about a new ‘culture’ of statistical modelling that bears considerable potential for the social scientist. This argument is illustrated with a brief discussion of model-based recursive partitioning which can bridge the theory and data-driven approach. Such a method is an example of (...) how this new approach can help revise models that work for the full dataset: it can be used for evaluating different models, a traditional weakness of the ‘traditional’ statistical approach used in social science. (shrink)

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)

Late in a career of more than sixty years, Thomas Burch, an internationally known social demographer, undertook a wide-ranging methodological critique of demography. This open access volume contains a selection of resulting papers, some previously unpublished, some published but not readily accessible [from past meetings of The International Union for the Scientific Study of Population and its research committees, or from other small conferences and seminars]. Rejecting the idea that demography is simply a branch of applied statistics, his work views (...) it as an autonomous and complete scientific discipline. When viewed from the perspective of modern philosophy of science, specifically the semantic or model-based school, demography is a balanced discipline, with a rich body of techniques and data, but also with more and better theories than generally recognized. As demonstrated in this book, some demographic techniques can also be seen as theoretical models, and some substantive/behavioral models, commonly rejected as theory because of inconsistent observations, are now seen as valuable theoretical models, for example demographic transition theory. This book shows how demography can build a strong theoretical edifice on its broad and deep empirical foundation by adoption of the model-based approach to science. But the full-fruits of this approach will require demographers to make greater use of computer modeling [both macro- and micro-simulation], in the statement and manipulation of theoretical ideas, as well as for numerical computation. This book is open access under a CC BY license. (shrink)

In this brief chapter, Lloyd sets the stage for the following three papers, most centrally, Santer et al., which discusses whether the satellite data fit with climate models. Its target is a paper by Douglass et al., which claimed that satellite and weather balloon data showed that the climate models were wrong and could not be trusted. The Santer and Wigley “Fact Sheet” gives a nontechnical summary of what is wrong with the Douglass paper, while the (...) full story is in Chap. 5, a reprint of the Santer et al. paper. The context and import of this work is discussed in Lloyd. In this introduction, Lloyd emphasizes the timeliness of this work, as climate deniers testify in Congress presently. (shrink)

In this paper I consider some data of Covid-19 infection in Italy from the 20th of February to the 29th of June 2020. Data are analyzed using some fits based on mathematical models. This analysis may be proposed to students of the last class of the Liceo Scientifico in order to debate a real problem with mathematical tools.

I introduce a distinction between general and local model-data symbiosis and offer three examples of local symbiosis in the fields of meteorology and climate science. Local model-data symbiosis ref...

Not all models are explanatory. Some models are data summaries. Some models sketch explanations but leave crucial details unspecified or hidden behind filler terms. Some models are used to conjecture a how-possibly explanation without regard to whether it is a how-actually explanation. I use the Hodgkin and Huxley model of the action potential to illustrate these ways that models can be useful without explaining. I then use the subsequent development of the explanation of the (...) action potential to show what is required of an adequate mechanistic model. Mechanistic models are explanatory. (shrink)

A recurring problem in conceptual modelling and ontology development is the representation of part-whole relations, with a requirement to be able to distinguish between essential and mandatory parts. To solve this problem, we formally characterize the semantics of these shareability notions by resorting to the temporal conceptual model E RVT and its formalization in the description logic DLRUS.

The transformers that drive chatbots and other AI systems constitute large language models (LLMs). These are currently the focus of a lively discussion in both the scientific literature and the popular media. This discussion ranges from hyperbolic claims that attribute general intelligence and sentience to LLMs, to the skeptical view that these devices are no more than “stochastic parrots”. I present an overview of some of the weak arguments that have been presented against LLMs, and I consider several of (...) the more compelling criticisms of these devices. The former significantly underestimate the capacity of transformers to achieve subtle inductive inferences required for high levels of performance on complex, cognitively significant tasks. In some instances, these arguments misconstrue the nature of deep learning. The latter criticisms identify significant limitations in the way in which transformers learn and represent patterns in data. They also point out important differences between the procedures through which deep neural networks and humans acquire knowledge of natural language. It is necessary to look carefully at both sets of arguments in order to achieve a balanced assessment of the potential and the limitations of LLMs. (shrink)

Recent frameworks, models, and reports highlight the critical need to address social determinants of health for achieving health equity in the United States and around the globe. In the United States, data play an important role in better understanding community‐level and population‐level disparities particularly for local health departments. However, data‐driven decision‐making—the use of data for public health activities such as program implementation, policy development, and resource allocation—is often presented theoretically or through case studies in the literature. (...) We sought to develop a preliminary model that identifies the factors that contribute to data‐driven decision‐making in US local health departments and describe relationships between them. Guided by implementation science literature, we examined organizational‐level capacity and individual‐level factors contributing to using data for decision‐making related to social determinants of health and the reduction of county‐level disparities. This model has the potential to improve implementation of public health interventions and programs aimed at upstream structural factors, by elucidating the factors critical to incorporating data in decision‐making. (shrink)

With the surge in the number of data and datafied governance initiatives, arrangements, and practices across the globe, understanding various types of such initiatives, arrangements, and their structural causes has become a daunting task for scholars, policy makers, and the public. This complexity additionally generates substantial difficulties in considering different data(fied) governances commensurable with each other. To advance the discussion, this study argues that existing scholarship is inclined to embrace an organization-centric perspective that primarily concerns factors and dynamics (...) regarding data and datafication at the organizational level at the expense of macro-level social, political, and cultural factors of both data and governance. To explicate the macro, societal dimension of data governance, this study then suggests the term “social data governance” to bring forth the consideration that data governance not only reflects the society from which it emerges but also (re)produces the policies and practices of the society in question. Drawing on theories of political science and public management, a model of social data governance is proposed to elucidate the ideological and conceptual groundings of various modes of governance from a comparative perspective. This preliminary model, consisting of a two-dimensional continuum, state intervention and societal autonomy for the one, and national cultures for the other, accounts for variations in social data governance across societies as a complementary way of conceptualizing and categorizing data governance beyond the European standpoint. Finally, we conduct an extreme case study of governing digital contact-tracing techniques during the pandemic to exemplify the explanatory power of the proposed model of social data governance. (shrink)

Coherence and correspondence are classical contenders as theories of truth. In this paper we examine them instead as interacting factors in the dynamics of belief across epistemic networks. We construct an agent-based model of network contact in which agents are characterized not in terms of single beliefs but in terms of internal belief suites. Individuals update elements of their belief suites on input from other agents in order both to maximize internal belief coherence and to incorporate ‘trickled in’ elements of (...) truth as correspondence. Results, though often intuitive, prove more complex than in simpler models (Hegselmann & Krause 2002, 2006; Grim, Singer, Reade & Fisher 2015). The optimistic finding is that pressures toward internal coherence can exploit and expand on small elements of truth as correspondence is introduced into epistemic networks. Less optimistic results show that pressures for coherence can also work directly against the incorporation of truth, particularly when coherence is established first and new data is introduced later. (shrink)

Empirical evidence suggests that while people hold the capacity to control their data in high regard, they increasingly experience a loss of control over their data in the online world. The capacity to exert control over the generation and flow of personal information is a fundamental premise to important values such as autonomy, privacy, and trust. In healthcare and clinical research this capacity is generally achieved indirectly, by agreeing to specific conditions of informational exposure. Such conditions can be (...) openly stated in informed consent documents or be implicit in the norms of confidentiality that govern the relationships of patients and healthcare professionals. However, with medicine becoming a data-intense enterprise, informed consent and medical confidentiality, as mechanisms of control, are put under pressure. In this paper we explore emerging models of informational control in data-intense healthcare and clinical research, which can compensate for the limitations of currently available instruments. More specifically, we discuss three approaches that hold promise in increasing individual control: the emergence of data portability rights as means to control data access, new mechanisms of informed consent as tools to control data use, and finally, new participatory governance schemes that allow individuals to control their data through direct involvement in data governance. We conclude by suggesting that, despite the impression that biomedical big data diminish individual control, the synergistic effect of new data management models can in fact improve it. (shrink)

The purpose of this paper is to identify the factors that affect the intention to use Big Data Applications in companies. Research into Big Data usage intention and adoption is scarce and much less from the perspective of the use of these techniques in companies. That is why this research focuses on analyzing the adoption of Big Data Applications by companies. Further to a review of the literature, it is proposed to use a UTAUT model as a (...) starting model with the update and incorporation of other variables such as resistance to use and perceived risk, and then to perform a neural network to predict this adoption. With respect to this non-parametric technique, we found that the multilayer perceptron model (MLP) for the use of Big Data Applications in companies obtains higher AUC values, and a better confusion matrix. This paper is a pioneering study using this hybrid methodology on the intention to use Big Data Applications. The result of this research has important implications for the theory and practice of adopting Big Data Applications. (shrink)

Large Language Models (LLMs) have been transformative. They are pre-trained foundational models that are self-supervised and can be adapted with fine tuning to a wide range of natural language tasks, each of which previously would have required a separate network model. This is one step closer to the extraordinary versatility of human language. GPT-3 and more recently LaMDA can carry on dialogs with humans on many topics after minimal priming with a few examples. However, there has been a (...) wide range of reactions and debate on whether these LLMs understand what they are saying or exhibit signs of intelligence. This high variance is exhibited in three interviews with LLMs reaching wildly different conclusions. A new possibility was uncovered that could explain this divergence. What appears to be intelligence in LLMs may in fact be a mirror that reflects the intelligence of the interviewer, a remarkable twist that could be considered a Reverse Turing Test. If so, then by studying interviews we may be learning more about the intelligence and beliefs of the interviewer than the intelligence of the LLMs. As LLMs become more capable they may transform the way we interact with machines and how they interact with each other. Increasingly, LLMs are being coupled with sensorimotor devices. LLMs can talk the talk, but can they walk the walk? A road map for achieving artificial general autonomy is outlined with seven major improvements inspired by brain systems. LLMs could be used to uncover new insights into brain function by downloading brain data during natural behaviors. (shrink)

As AI technologies are rolled out into healthcare, academia, human resources, law, and a multitude of other domains, they become de-facto arbiters of truth. But truth is highly contested, with many different definitions and approaches. This article discusses the struggle for truth in AI systems and the general responses to date. It then investigates the production of truth in InstructGPT, a large language model, highlighting how data harvesting, model architectures, and social feedback mechanisms weave together disparate understandings of veracity. (...) It conceptualizes this performance as an operationalization of truth, where distinct, often-conflicting claims are smoothly synthesized and confidently presented into truth-statements. We argue that these same logics and inconsistencies play out in Instruct’s successor, ChatGPT, reiterating truth as a non-trivial problem. We suggest that enriching sociality and thickening “reality” are two promising vectors for enhancing the truth-evaluating capacities of future language models. We conclude, however, by stepping back to consider AI truth-telling as a social practice: what kind of “truth” do we as listeners desire? (shrink)

Subjects exposed to strings of letters generated by a finite state grammar can later classify grammatical and nongrammatical test strings, even though they cannot adequately say what the rules of the grammar are (e.g., Reber, 1989). The MINERVA 2 (Hintzman, 1986) and Medin and Schaffer (1978) memory‐array models and a number of connectionist outoassociator models are tested against experimental data by deriving mainly parameter‐free predictions from the models of the rank order of classification difficulty of test (...) strings. The importance of different assumptions regarding the coding of features (How should the absence of a feature be coded? Should single letters or digrams be coded?), the learning rule used (Hebb rule vs. delta rule), and the connectivity (Should features be predicted only by previous features in the string, or by all features simultaneously?) is investigated by determining the performance of the models with and without each assumption. Only one class of connectionist model (the simultaneous delta rule) passes all the tests. It is shown that this class of model can be regarded by abstracting a set of representative but incomplete rules of the grammar. (shrink)

According to the hierarchy of models account of scientific experimentation developed by Patrick Suppes and elaborated by Deborah Mayo, theoretical considerations about the phenomena of interest are involved in an experiment through theoretical models that in turn relate to experimental data through data models, via the linkage of experimental models. In this paper, I dispute the HoM account in the context of present-day high-energy physics experiments. I argue that even though the HoM account aims (...) to characterize experimentation as a model-based activity, it does not involve a modeling concept for the process of data acquisition, and it thus fails to provide a model-based characterization of the theory-experiment relationship underlying this process. In order to characterize the foregoing relationship, I propose the concept of a model of data acquisition and illustrate it in the case of the ATLAS experiment at CERN’s Large Hadron Collider, where the Higgs boson was discovered in 2012. I show that the process of data acquisition in the ATLAS experiment is performed according to a model of data acquisition that specifies and organizes the experimental procedures necessary to select the data according to a predetermined set of selection criteria. I also point out that this data acquisition model is theory-laden, in the sense that the underlying data selection criteria are determined by considering the testable predictions of the theoretical models that the ATLAS experiment is aimed to test. I take this sense of theory-ladenness to indicate that the relationship between the procedures of the ATLAS experiment and the theoretical models of the phenomena of interest is first established, prior to the formation of data models, through the data acquisition model of the experiment, thus not requiring the intermediary of other types of models as suggested by the HoM account. I therefore conclude that in the context of present-day HEP experiments, the HoM account does not consistently extend to the process of data acquisition so as to include models of data acquisition. (shrink)

In the last decades, supervised machine learning has seen the widespread growth of highly complex, non-interpretable models, of which deep neural networks are the most typical representative. Due to their complexity, these models have showed an outstanding performance in a series of tasks, as in image recognition and machine translation. Recently, though, there has been an important discussion over whether those non-interpretable models are able to provide any sort of understanding whatsoever. For some scholars, only interpretable (...) class='Hi'>models can provide understanding. More popular, however, is the idea that understanding can come from a careful analysis of the dataset or from the model’s theoretical basis. In this paper, I wish to examine the possible forms of obtaining understanding of such non-interpretable models. Two main strategies for providing understanding are analyzed. The first involves understanding without interpretability, either through external evidence for the model’s inner functioning or through analyzing the data. The second is based on the artificial production of interpretable structures, through three main forms: post hoc models, hybrid models, and quasi-interpretable structures. Finally, I consider some of the conceptual difficulties in the attempt to create explanations for these models, and their implications for understanding. (shrink)

This paper addresses a relatively new mode of ecological research: data synthesis studies. Data synthesis studies involve reusing data to create a general model as well as a reusable, aggregated dataset. Using a case from movement ecology, I analyse the trade-offs and strategies involved in data synthesis. Like theoretical ecological modelling, I find that synthesis studies involve a modelling trade-off between generality, precision and realism; they deal with this trade-off by adopting a pragmatic kludging strategy. I (...) also identify an additional trade-off, the synthesis trade-off, between making data easy to synthesise for a particular project, on the one hand, and facilitating data reuse for other projects, on the other. In response to this synthesis trade-off, researchers create flexible datasets that are relatively easy to use for particular projects and can be adjusted to suit some other purposes. The flexibility compromise is also found in broader open data efforts, making it a significant element in the future of data-intensive ecology. (shrink)

Prelog’s model was one of the first empirical models to explain the stereoselectivity of the Grignard reactions of 2-oxocarboxylic acid esters bearing a chiral alcohol. Prelog constructed his model based on some assumptions regarding the conformation of chiral 2-oxocarboxylic acid esters to explain the relationship in configuration between the chiral alcohol starting materials and the 2-hydroxycarboxylic acid products. Construction of the model involves four steps: (1) mentally analyzing the reactants to identify the basic stereochemical structures, (2) assuming the conformations (...) of the chiral alcohol moiety as the main reasons for stereoselectivity, (3) modeling the transition states to explain the relationship in configuration between the chiral alcohols and 2-hydroxycarboxylic acids, and (4) validating the evidence to revise the model. The process of constructing this historically pivotal model may have implications for science education, especially for developing theories from extensive data. The four steps from the Prelog’s model were compared with the models for boiling points of hydrocarbons and solubility of amino acids. (shrink)

The biological sciences have become increasingly reliant on so-called 'model organisms'. I argue that in this domain, the concept of a descriptive model is essential for understanding scientific practice. Using a case study, I show how such a model was formulated in a preexplanatory context for subsequent use as a prototype from which explanations ultimately may be generated both within the immediate domain of the original model and in additional, related domains. To develop this concept of a descriptive model, I (...) focus on use of the nematode worm Caenorhabditis elegans and the wiring diagrams that were developed as models of its neural structure. In addition, implications of the concept of a descriptive model, particularly its relevance for the data-phenomena distinction as well as its relation to long-standing debates on realism, are briefly examined. (shrink)

Mechanistic models are built using knowledge as the primary information source, with well-established biological and physical laws determining the causal relationships within the model. Once the causal structure of the model is determined, parameters must be defined in order to accurately reproduce relevant data. Determining parameters and their values is particularly challenging in the case of models of pathophysiology, for which data for calibration is sparse. Multiple data sources might be required, and data may (...) not be in a uniform or desirable format. We describe a calibration strategy to address the challenges of scarcity and heterogeneity of calibration data. Our strategy focuses on parameters whose initial values cannot be easily derived from the literature, and our goal is to determine the values of these parameters via calibration with constraints set by relevant data. When combined with a covariance matrix adaptation evolution strategy, this step-by-step approach can be applied to a wide range of biological models. We describe a stepwise, integrative and iterative approach to multiscale mechanistic model calibration, and provide an example of calibrating a pathophysiological lung adenocarcinoma model. Using the approach described here we illustrate the successful calibration of a complex knowledge-based mechanistic model using only the limited heterogeneous datasets publicly available in the literature. (shrink)

Following Brunswik (1952), social judgement theorists have long relied on regression models to describe both an individual's judgements and the environment about which such judgements are made. However, social judgement theory is not synonymous with these compensatory, static, structural models. We compared the characterisations of physicians' judgements using a regression model with that of a non-compensatory process model (called fast and frugal). We found that both models fit the data equally well. Both models suggest that (...) physicians use few cues, that they disagree among themselves, and that their stated cue use is discrepant with the models' stated cue use. However, the fast and frugal model is easier to convey to physicians and is also more psychologically plausible. Implications for how social judgement theorists conceptualise the process of vicarious functioning are discussed. (shrink)

I propose a framework that explicates and distinguishes the epistemic roles of data and models within empirical inquiry through consideration of their use in scientific practice. After arguing that Suppes’ characterization of data models falls short in this respect, I discuss a case of data processing within exploratory research in plant phenotyping and use it to highlight the difference between practices aimed to make data usable as evidence and practices aimed to use data (...) to represent a specific phenomenon. I then argue that whether a set of objects functions as data or models does not depend on intrinsic differences in their physical properties, level of abstraction or the degree of human intervention involved in generating them, but rather on their distinctive roles towards identifying and characterizing the targets of investigation. The paper thus proposes a characterization of data models that builds on Suppes’ attention to data practices, without however needing to posit a fixed hierarchy of data and models or a highly exclusionary definition of data models as statistical constructs. (shrink)

This paper advances further an ongoing project to understand the history of stellar structure modelling and its inferential practice. It does so by taking a harder look at the data: how it is collected, analysed statistically, and represented in HR diagrams and stellar structure models alike. The focus is ultimately on the sorts of strong observational constraints revealed in the last two decades within the new and expanding field of asteroseismology. It is argued that the typical inferential practices (...) in asteroseismology, while richly loaded with modelling assumptions of their own, do not raise any circularity worries that may compromise the quality or value of the data. (shrink)