It is increasingly recognised that evidence generated using “real-world data” is crucial for assessing the safety and effectiveness of health-related interventions. This, however, raises a number of issues, including those related to the quality of RWD, and of the scientific methods used to generate evidence from it, and the potential for those gathering and using RWD be driven by commercial, political, professional or personal self-interest. This article is an application of the framework presented in this issue of (...) ABR. Please refer to that article for more information on how this framework is to be used, including a full explanation of the key values involved and the balancing approach used in the case study at the end to demonstrate how those generating or using RWD can make use of values when deliberating about ethical issues. (shrink)

An emerging approach to complement randomised controlled trial (RCT) data in the development of radiotherapy treatments is to use routinely collected ‘real-world’ data (RWD). RWD is the data collected as standard-of-care about all patients during their usual cancer care pathway. Given the nature of this data, important questions remain about the permissibility and acceptability of using RWD in routine practice. We involved and engaged with patients, carers and the public in a two-day citizens’ jury (...) to understand their views and obtain decisions regarding two key issues: (1) preferred approaches to consent for the use of RWD within the context of patients receiving radiotherapy for lung cancer in RAPID-RT and (2) how RWD use should be best communicated to patients. Individual views were polled using questionnaires at various stages of the jury, whilst group discussion activities prompted further dialogue about the rationale behind choices of consent. Key decisions obtained from the jury include: (1) an opt-out approach to consent for the use of RWD; (2) the opt-out approach to consent should be informed. Furthermore, it was advised that information and communication regarding the consent process and use of RWD should be accessible, clear and available in a variety of formats. It is important that the consent process for patient data use is underpinned by principles of autonomy and transparency with clear channels of communication between those asking for and giving consent. Moreover, the process of seeking consent from patients should be proportionate to the risks presented from their participation. (shrink)

Volume 20, Issue 7, July 2020, Page 89-92.

In their open peer commentary: “Making It Count: Extracting Real World Data from Compassionate Use and Expanded Access Programs”, Rozenberg and Greenbaum discuss impo...

For centuries, fiction—particularly fables—has seamlessly combined storytelling, moral lessons, and societal reflections to engage readers on both emotional and intellectual levels. Despite extensive research on the benefits of reading and the emotional responses it evokes, a critical gap remains in understanding what drives readers to form deep emotional connections with specific works. This study seeks to identify the characteristics of a book that foster such connections. Using Bayesian Mindsponge Framework analytics, we analyzed a dataset of 129 Amazon reviews of Wild (...) Wise Weird, a collection of 42 fables that intertwine traditional storytelling with contemporary sensibilities, offering life lessons, humor, and social commentary. Of these reviews, approximately 66% expressed love for the book. The findings reveal that readers who describe the book as unique, whimsical, quirky, or innovative are more likely to express affection for it. Similarly, readers who are drawn to the book’s illustrations and characters tend to form stronger emotional connections. The cultural richness and representativeness of the book also emerged as significant factors in fostering readers’ love. The book successfully captures and conveys the depth, charm, and moral essence of Vietnamese culture to a global audience. Readers have expressed heartfelt gratitude after engaging with the book, underscoring its success in fostering meaningful emotional connections. Thus, insights from this analysis offer actionable recommendations for stakeholders in the literary ecosystem, from authors and publishers to marketers, to enhance readers’ emotional attachment and commitment to books. (shrink)

The evidence compiled in the target article demonstrates that the assumptions of cultural group selection theory are often met, and it is therefore a useful framework for generating plausible hypotheses. However, more can be said about how we can test the predictions of CGS hypotheses against competing explanations using historical, archaeological, and anthropological data.

This article probes the visual spatial représentations underlying the creative conceptual design of complex objects.

Decisions based on algorithmic, machine learning models can be unfair, reproducing biases in historical data used to train them. While computational techniques are emerging to address aspects of these concerns through communities such as discrimination-aware data mining and fairness, accountability and transparency machine learning, their practical implementation faces real-world challenges. For legal, institutional or commercial reasons, organisations might not hold the data on sensitive attributes such as gender, ethnicity, sexuality or disability needed to diagnose and (...) mitigate emergent indirect discrimination-by-proxy, such as redlining. Such organisations might also lack the knowledge and capacity to identify and manage fairness issues that are emergent properties of complex sociotechnical systems. This paper presents and discusses three potential approaches to deal with such knowledge and information deficits in the context of fairer machine learning. Trusted third parties could selectively store data necessary for performing discrimination discovery and incorporating fairness constraints into model-building in a privacy-preserving manner. Collaborative online platforms would allow diverse organisations to record, share and access contextual and experiential knowledge to promote fairness in machine learning systems. Finally, unsupervised learning and pedagogically interpretable algorithms might allow fairness hypotheses to be built for further selective testing and exploration. Real-world fairness challenges in machine learning are not abstract, constrained optimisation problems, but are institutionally and contextually grounded. Computational fairness tools are useful, but must be researched and developed in and with the messy contexts that will shape their deployment, rather than just for imagined situations. Not doing so risks real, near-term algorithmic harm. (shrink)

ABSTRACTInspired by the discovery of positive age trends in emotional well-being across adulthood, lifespan researchers have uncovered fascinating age differences in cognition–emotion interactions in healthy adult samples, for example in emotion processing, memory, reactivity, perception, and regulation. Taking stock of this body of research, I identify four trends and five remaining gaps in our understanding of emotional functioning in adulthood. In particular, I suggest that the field should pay stronger attention to the prediction of real-world behaviour. Using the (...) sample case of work functioning, I outline gaps in current knowledge, including the lack of data on middle-aged adults, the neglect of relevant cognitive-emotional mechanisms, and the unclear role of life experience. Filling these gaps will enable progress in research on emotional aging in and beyond the work setting and enhance its practical utility for individuals, organisations, and society. (shrink)

Conflict resolution practitioners consistently struggle with access to structured armed conflict data, a dataset already rife with uncertainty, inconsistency, and politicization. Due to the lack of a standardized approach to collating conflict data, publicly available armed conflict datasets often require manipulation depending upon the needs of end users. Transformation of armed conflict data tends to be a manual, time-consuming task that nonprofits with limited budgets struggle to keep up with. In this paper, we explore the use of (...) a deep natural language processing model to aid the transformation of armed conflict data for conflict analysis. Our model drastically reduces the time spent on manual data transformations and improves armed conflict event classification by identifying multiple incidence types. This minimizes the human supervision cost and allows nonprofits to access a broader range of conflict data sources to reduce reporting bias. Thus, our model contributes to the incorporation of technology in the peace building and conflict resolution sector. (shrink)

The visual system is known to extract summary representations of visually similar objects which bias the perception of individual objects toward the ensemble average. Although vision plays a large role in guiding action, less is known about whether ensemble representation is informative for action. Motor behavior is tuned to the veridical dimensions of objects and generally considered resistant to perceptual biases. However, when the relevant grasp dimension is not available or is unconstrained, ensemble perception may be informative to behavior by (...) providing gist information about surrounding objects. In the present study, we examined if summary representations of a surrounding ensemble display influenced grip aperture and orientation when participants reached-to-grasp a central circular target which had an explicit size but importantly no explicit orientation that the visuomotor system could selectively attend to. Maximum grip aperture and grip orientation were not biased by ensemble statistics during grasping, although participants were able to perceive and provide manual estimations of the average size and orientation of the ensemble display. Support vector machine classification of ensemble statistics achieved above-chance classification accuracy when trained on kinematic and electromyography data of the perceptual but not grasping conditions, supporting our univariate findings. These results suggest that even along unconstrained grasping dimensions, visually-guided behaviors toward real-world objects are not biased by ensemble processing. (shrink)

Humankind is concerned with scientific enquiry because humans want to understand the milieu in which they find themselves. They want to engineer and reliably control or cooperatively modulate as much of the environment as possible to sustain, enrich, and propagate their lives. Following this path, the goal of science is to gain a reliable description of all natural phenomena so as to allow accurate prediction (within appropriate limits) of nature’s behavior as a function of an ever-changing environment. As such, science (...) is incapable of providing us with absolute truth. Rather, it provides relative knowledge, internally self-consistent knowledge, about the relationships between different phenomena and between different things. The goal of engineering, on the other hand, is to build on this fundamental understanding to generate new materials, devices, structures, attitudes, moralities, philosophies, and so on, for producing tangible order, harnessing the latent potential in nature’s phenomena and expanding human capabilities in an ever-changing environment. In this context, medicine is to human biology as engineering is to physical science. As each of us rides the “river of life,” the great consciousness adventure, we perceive events occurring around us, but more often than not, we do not perceive the total information content inherent in those events nor the true reality of those events. The latter point is so because what we take as the reality of an observation is actually a convolution between (1) what our sensory system actually senses and (2) our mindset or belief structure that filters and/or selectively amplifies segments of the basic gathered data stream. Further, only what we call the five physical senses are well-developed and integrated in our overall sensory system, so only a portion of the total available data becomes our basic internal data stream. We are thus always making personal observations through the distorting and spectrally-limited lens of our mindsets, and we have no way at present to perform a deconvolution and perceive the pure information inherent in our basic input data stream. By using designed instruments to access information patterns in nature, we gain a more objective perspective of these events. However, we must always remember that these instruments were designed on the basis of the logic of our average cognitive development and therefore probably also have only a limited access to the total information spectrum for these events occurring in nature. In particular, these instruments only respond to positive energies. Over the course of the past four to five centuries, we have learned how to conduct true scientific investigations, first under the rubric of classical mechanics and, more recently, quantum mechanics. Let us now look a little more deeply to see what this means. (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)

In real life, behavior is influenced by dynamically changing contextual factors and is rarely limited to simple tasks and binary choices. For a meaningful interpretation of brain dynamics underlying more natural cognitive processing in active humans, ecologically valid test scenarios are essential. To understand whether brain dynamics in restricted artificial lab settings reflect the neural activity in complex natural environments, we systematically tested the auditory event-related P300 in both settings. We developed an integrative approach comprising an initial P300-study in (...) a highly controlled laboratory set-up and a subsequent validation within a realistic driving scenario. Using a simulated dialog with a speech-based input system, increased P300 amplitudes reflected processing of infrequent and incorrect auditory feedback events in both the laboratory setting and the real world setup. Environmental noise and movement-related activity in the car driving scenario led to higher data rejection rates but revealed comparable theta and alpha frequency band pattern. Our results demonstrate the possibility to investigate cognitive functions like context updating in highly adverse driving scenarios and encourage the consideration of more realistic task settings in prospective brain imaging approaches. (shrink)

In the 21st century, digitalization is a global challenge of mankind. Even for the public, it is obvious that our world is increasingly dominated by powerful algorithms and big data. But, how computable is our world? Some people believe that successful problem solving in science, technology, and economies only depends on fast algorithms and data mining. Chances and risks are often not understood, because the foundations of algorithms and information systems are not studied rigorously. Actually, they (...) are deeply rooted in logics, mathematics, computer science and philosophy. Therefore, this book studies the foundations of mathematics, computer science, and philosophy, in order to guarantee security and reliability of the knowledge by constructive proofs, proof mining and program extraction. We start with the basics of computability theory, proof theory, and information theory. In a second step, we introduce new concepts of information and computing systems, in order to overcome the gap between the digital world of logical programming and the analog world of real computing in mathematics and science. The book also considers consequences for digital and analog physics, computational neuroscience, financial mathematics, and the Internet of Things (IoT). (shrink)

This article presents a computational model of the learning of diagnostic knowledge, based on observations of human operators engaged in real-world troubleshooting tasks. We present a model of problem solving and learning in which the reasoner introspects about its own performance on the problem-solving task, identifies what it needs to learn to improve its performance, formulates learning goals to acquire the required knowledge, and pursues its learning goals using multiple learning strategies. The model is implemented in a computer (...) system which provides a case study based on observations of troubleshooting operators and protocol analysis of the data gathered in the test area of an operational electronics manufacturing plant. The model not only addresses issues in human learning, but, in addition, is computationally justified as a uniform, extensible framework for multistrategy learning. (shrink)

Despite its effectiveness, limited research has examined the provision of telemental health and how practices may vary according to treatment paradigm. We surveyed 276 community mental health providers registered with a commercial telemedicine platform. Most providers reported primarily offering TMH services to adults with anxiety, depression, and trauma-and stressor-related disorders in individual therapy formats. Approximately 82% of TMH providers reported endorsing the use of Cognitive Behavioral Therapy in their remote practice. The most commonly used in-session and between-session exercises included coping (...) and emotion regulation, problem solving, mindfulness, interpersonal skills, relaxation, and modifying and addressing core beliefs. CBT TMH providers had a higher odds of using in-session and homework exercises and assigning them through postal mail, email or fax methods, as compared to non-CBT TMH providers. TMH providers, regardless of treatment paradigm, felt that assigning homework was neither easy nor difficult and they believed their patients were somewhat-to-moderately compliant to their assigned exercises. CBT TMH providers also collected clinical information from their patients more often than non-CBT TMH providers. They reported being less satisfied with their method, which was identified most often as paper-based surveys and forms. Overall, TMH providers employ evidence-based treatments to their patients remotely, with CBT TMH providers most likely to do so. Findings highlight the need for innovative solutions to improve how TMH providers that endorse following the CBT treatment paradigm remotely assign homework and collect clinical data to increase their satisfaction via telemedicine. (shrink)

In recent years, a multitude of datasets of human–human conversations has been released for the main purpose of training conversational agents based on data‐hungry artificial neural networks. In this paper, we argue that datasets of this sort represent a useful and underexplored source to validate, complement, and enhance cognitive studies on human behavior and language use. We present a method that leverages the recent development of powerful computational models to obtain the fine‐grained annotation required to apply metrics and techniques (...) from Cognitive Science to large datasets. Previous work in Cognitive Science has investigated the question‐asking strategies of human participants by employing different variants of the so‐called 20‐question‐game setting and proposing several evaluation methods. In our work, we focus on GuessWhat, a task proposed within the Computer Vision and Natural Language Processing communities that is similar in structure to the 20‐question‐game setting. Crucially, the GuessWhat dataset contains tens of thousands of dialogues based on real‐world images, making it a suitable setting to investigate the question‐asking strategies of human players on a large scale and in a natural setting. Our results demonstrate the effectiveness of computational tools to automatically code how the hypothesis space changes throughout the dialogue in complex visual scenes. On the one hand, we confirm findings from previous work on smaller and more controlled settings. On the other hand, our analyses allow us to highlight the presence of “uninformative” questions (in terms of Expected Information Gain) at specific rounds of the dialogue. We hypothesize that these questions fulfill pragmatic constraints that are exploited by human players to solve visual tasks in complex scenes successfully. Our work illustrates a method that brings together efforts and findings from different disciplines to gain a better understanding of human question‐asking strategies on large‐scale datasets, while at the same time posing new questions about the development of conversational systems. (shrink)

Based on empirical data from a qualitative study, this paper explores the complexity of ‘real world’ management in Hong Kong’s public sector, as contrasted with various paradigmatic claims under ‘new public management’ (NPM). A plurality of sub-worlds within the broad public sector is identified, which makes the management roles and responsibilities much less ‘homogenised’ than depicted in NPM exhortations. The instrumental rationality underpinning NPM is identified as too restrictive in understanding the way in which public managers reach (...) decisions. When the daily challenges of reconciling values and practices arising from the complexities of politics, policies and service delivery are considered it is necessary to incorporate ideas related to procedural and expressive rationality to fully appreciate the nature of management in public organisations. (shrink)

Background Implicit biases are present in the general population and among professionals in various domains, where they can lead to discrimination. Many interventions are used to reduce implicit bias. However, uncertainties remain as to their effectiveness. -/- Methods We conducted a systematic review by searching ERIC, PUBMED and PSYCHINFO for peer-reviewed studies conducted on adults between May 2005 and April 2015, testing interventions designed to reduce implicit bias, with results measured using the Implicit Association Test (IAT) or sufficiently similar methods. (...) -/- Results 30 articles were identified as eligible. Some techniques, such as engaging with others’ perspective, appear unfruitful, at least in short term implicit bias reduction, while other techniques, such as exposure to counterstereotypical exemplars, are more promising. Robust data is lacking for many of these interventions. -/- Conclusions Caution is thus advised when it comes to programs aiming at reducing biases. This does not weaken the case for implementing widespread structural and institutional changes that are multiply justified. (shrink)

This chapter examines the role of the situation model in referential processing and how it can link what appear to be incompatible results from studies of monologue and dialogue as well as studies of reading and visual-world eye tracking. It shows that data from experiments on pronoun resolution in reading indicate a two-step model, in which candidate antecedents for an anaphor are first identified on the basis of gender matching and number matching, then evaluated with respect to the (...) overall situation model. In similar visual-world studies, these stages are seemingly collapsed together, which can be explained in terms of differences in how the situation model is related to the text versus the visual display of potential referents. The chapter suggests that comparing these different types of studies can elucidate the processes of reference resolution. (shrink)

Universities and funders in many countries have been using Journal Impact Factor (JIF) as an indicator for research and grant assessment despite its controversial nature as a statistical representation of scientific quality. This study investigates how the changes of JIF over the years can affect its role in research evaluation and science management by using JIF data from annual Journal Citation Reports (JCR) to illustrate the changes. The descriptive statistics find out an increase in the median JIF for the (...) top 50 journals in the JCR, from 29.300 in 2017 to 33.162 in 2019. Moreover, on average, elite journal families have up to 27 journals in the top 50. In the group of journals with a JIF of lower than 1, the proportion has shrunk by 14.53% in the 2015–2019 period. The findings suggest a potential ‘JIF bubble period’ that science policymaker, university, public fund managers, and other stakeholders should pay more attention to JIF as a criterion for quality assessment to ensure more efficient science management. (shrink)

The COVID-19 pandemic significantly impacted medical education, causing a shift towards online learning. However, this transition posed challenges in administering online assessments, particularly in proctoring and detecting academic misconduct. This study aimed to investigate the prevalence of academic misconduct among medical students during remote online examinations using hierarchical clustering and comparing self-reported confessions and recorded video reviews. The results of the study confirmed the existence of academic misconduct in remote online examinations, as evidenced by both self-report and video footage review. (...) Hierarchical clustering was able to predict academic misconduct with an accuracy rate of approximately 55%. These findings highlight the need for medical educators to develop and implement ethics and morality-focused courses to address the issues of dishonesty among medical students in online assessments. In conclusion, this study shed light on the challenges of remote online assessments in medical education, particularly in detecting academic misconduct. Using a data-driven approach to analyse the prevalence of academic misconduct, this work provided insights into the need for ethical and moral education in medical curricula, which can help address the dishonesty among medical students during online assessments. (shrink)

Open peer commentary on the article “Modeling Subjects’ Experience While Modeling the Experimental Design: A Mild-Neurophenomenology-Inspired Approach in the Piloting Phase” by Constanza Baquedano & Catalina Fabar. Upshot: The authors show in their pilots how open it is to participants not to obey the instructions during an experiment. Their findings leave us to choose between two options: either we accept that subjective confounds are inevitable and stronger than we think, but in this case, why should we continue trying to measure (...) subjective experience?; or strive at designing better experiments in order to control for these fluctuations. I will argue for option and propose an alternative model to go beyond the first- and third-person data gap, namely “predictive processing.”. (shrink)

Phylogenetic trees based on gene sequence data contain information about the evolutionary processes responsible for their genesis. Methods have now been developed which help to reveal those processes. The methods are based on simple models of evolutionary change but, when applied across individuals in a population, rather than across species in a higher‐level taxon, they can reveal the past history of population change. Examples from salamanders and viruses are used to illustrate how the past history of changes in speciation (...) rate and the origin of epidemics can be inferred in the absence of fossil material or historical documentation. (shrink)

Pulmonary rehabilitation has demonstrated patients’ physiological and psychosocial improvements, symptoms reduction and health-economic benefits whilst enhances the ability of the whole family to adjust to illness. However, PR remains highly inaccessible due to lack of awareness of its benefits, poor referral and availability mostly in hospitals. Novel models of PR delivery are needed to enhance its implementation while maintaining cost-efficiency. We aim to implement an innovative community-based PR programme and assess its cost-benefit. A 12-week community-based PR will be implemented in (...) primary healthcare centres where programmes are not available. Healthcare professionals will be trained. 73 patients with CRD and their caregivers will compose the experimental group. The control group will include dyads age- and disease-matched willing to collaborate in data collection but not in PR. Patients/family-centred outcomes will be dyspnoea, fatigue, cough and sputum, impact of the disease, emotional state, number of exacerbations, healthcare utilisation, health-related quality of life and family adaptability/cohesion. Other clinical outcomes will be peripheral and respiratory muscle strength, muscle thickness and cross sectional area, exercise capacity, balance and physical activity. Data will be collected at baseline, at 12 weeks, at 3- and 6-months post-PR. Changes in the outcome measures will be compared between groups, after multivariate adjustment for possible confounders, and effect sizes will be calculated. A cost-benefit analysis will be conducted. This study will enhance patients access to PR, by training healthcare professionals in the local primary healthcare centres to conduct such programmes and actively involving caregivers. The cost-benefit analysis of this intervention will provide an evidence-based insight into the economic benefit of community-based PR in chronic respiratory diseases. The trial was registered in the ClinicalTrials. gov U.S. National Library of Medicine, on 10th January, 2019. (shrink)

Nowadays, protecting trust in social sciences also means engaging in open community dialogue, which helps to safeguard robustness and improve efficiency of research methods. The combination of open data, open review and open dialogue may sound simple but implementation in the real world will not be straightforward. However, in view of Begley and Ellis’s (2012) statement that, “the scientific process demands the highest standards of quality, ethics and rigour,” they are worth implementing. More importantly, they are feasible (...) to work on and likely will help to restore plausibility to social sciences research. Therefore, I feel it likely that the triplet of open data, open review and open dialogue will gradually emerge to become policy requirements regardless of the research funding source. (shrink)

We face reality presented with the data of conscious experience and nothing else. The project of early modern philosophy was to build a complete theory of the world from this starting point, with no cheating. Crucial to this starting point is the data of conscious sensory experience – sense data. Attempts to avoid this project often argue that the very idea of sense data is confused. But the sense-data way of talking, the sense-data (...) language, can be freed from every blemish using ideas from contemporary metaontology. We can adopt a sense-data framework that vindicates the traditional claims of sense-data theories and leads to plausible theories of perception and color. We can, we should, and in a sense, we must. Yet when we do, we face the traditional problem of external world skepticism, head-on. The real challenge of skepticism is forced upon us; it cannot honestly be avoided using externalist tricks, burden of proof shifting, or other razzle dazzle. But the challenge can be met: we are rational to posit the external world as the best explanation of the many synchronic and diachronic patterns over our sense-data. This paper argues for all of these points and ends with a plea for analytic empiricism – a traditional sense-data version of empiricism that uses all of the tools of analytic philosophy while avoiding the more questionable doctrines of the Vienna Circle (phenomenalism, verificationism). Analytic empiricism is our last best hope for completing the great philosophical project of early modern philosophy. (shrink)

The very expertise with which psychologists wield their tools for achieving laboratory control may have had the unwelcome effect of blinding psychologists to the possibilities of discovering principles of behavior without conducting experiments. When creatively interrogated, a diverse range of large, real-world data sets provides powerful diagnostic tools for revealing principles of human judgment, perception, categorization, decision-making, language use, inference, problem solving, and representation. Examples of these data sets include patterns of website links, dictionaries, logs of (...) group interactions, collections of images and image tags, text corpora, history of financial transactions, trends in twitter tag usage and propagation, patents, consumer product sales, performance in high-stakes sporting events, dialect maps, and scientific citations. The goal of this issue is to present some exemplary case studies of mining naturally existing data sets to reveal important principles and phenomena in cognitive science, and to discuss some of the underlying issues involved with conducting traditional experiments, analyses of naturally occurring data, computational modeling, and the synthesis of all three methods. (shrink)

Ethical decision-making frameworks assist in identifying the issues at stake in a particular setting and thinking through, in a methodical manner, the ethical issues that require consideration as well as the values that need to be considered and promoted. Decisions made about the use, sharing, and re-use of big data are complex and laden with values. This paper sets out an Ethics Framework for Big Data in Health and Research developed by a working group convened by the Science, (...) Health and Policy-relevant Ethics in Singapore Initiative. It presents the aim and rationale for this framework supported by the underlying ethical concerns that relate to all health and research contexts. It also describes a set of substantive and procedural values that can be weighed up in addressing these concerns, and a step-by-step process for identifying, considering, and resolving the ethical issues arising from big data uses in health and research. This Framework is subsequently applied in the papers published in this Special Issue. These papers each address one of six domains where big data is currently employed: openness in big data and data repositories, precision medicine and big data, real-world data to generate evidence about healthcare interventions, AI-assisted decision-making in healthcare, public-private partnerships in healthcare and research, and cross-sectoral big data. (shrink)

In recent years, the healthcare field welcomed an emerging field of practices captured under the umbrella term ‘Big Data’. This term is surrounded with positive rhetoric and promises about the ability to analyse real-world data quickly and comprehensively. Such rhetoric is highly consequential in shaping debates on Big Data. While the fields of Science and Technology Studies and Critical Data Studies have been instrumental in elaborating the neglected and problematic dimensions of Big Data, (...) it remains an open question how and to what extent such insights become embedded in other fields. In this paper, we analyse the epistemological claims that accompany Big Data in the healthcare domain. We systematically searched scientific literature and selected 206 editorials as these reflect on developments in the domain. Through an interpretive analysis, we construct five ideal-typical discourses that all frame Big Data in specific ways. Three of the discourses frame Big Data in positive terms and disseminate a compelling rhetoric. Metaphors of ‘capturing’, ‘illuminating’ and ‘harnessing’ data presume that Big Data are benign and leading to valid knowledge. The scientist and critical-interpretive discourses question the objectivity and effectivity claims of Big Data. Metaphors of ‘selecting’ and ‘constructing’ data illustrate another political message, framing Big Data as limited. We conclude that work in the critical-interpretive discourse has not broadly infiltrated the medical domain. Ways to better integrate aspects of the discourse in the healthcare domain are urgently needed. (shrink)

This article argues that the fact that an action will compound a prior injustice counts as a reason against doing the action. I call this reason The Anti-Compounding Injustice principle or aci. Compounding injustice and the aci principle are likely to be relevant when analyzing the moral issues raised by “big data” and its combination with the computational power of machine learning and artificial intelligence. Past injustice can infect the data used in algorithmic decisions in two distinct ways. (...) Sometimes prior injustice undermines the accuracy of the data itself. In these contexts, improving accuracy will also help to avoid compounding injustice. Other times, past injustice produces real-world differences among people with regard to skills, health, wealth, and other traits that employers, lenders, and others seek to measure. When decisions are based on accurate data that itself results from prior injustice, these decisions can also compound injustice. This second dynamic has received less attention than the first but is especially important because improving the accuracy of data will not mitigate this unfairness. (shrink)

In a context of rapid technological innovation and expensive new products, the paper calls for the generation of real-world data to inform decision-making and an international discussion on the affordability of new medicines, particularly for low- and middle-income countries. Without these, the challenges of health judicialization will continue to grow.

BackgroundResearch using data from medical care promises to advance medical science and improve healthcare. Academia is not the only sector that expects such research to be of great benefit. The research-based health industry is also interested in so-called ‘real-world’ health data to develop new drugs, medical technologies or data-based health applications. While access to medical data is handled very differently in different countries, and some empirical data suggest people are uncomfortable with the idea (...) of companies accessing health information, this paper aims to advance the ethical debate about secondary use of medical data generated in the public healthcare sector by for-profit companies for medical research (ReuseForPro).MethodsWe first clarify some basic concepts and our ethical-normative approach, then discuss and ethically evaluate potential claims and interests of relevant stakeholders: patients as data subjects in the public healthcare system, for-profit companies, the public, and physicians and their healthcare institutions. Finally, we address the tensions between legitimate claims of different stakeholders in order to suggest conditions that might ensure ethically sound ReuseForPro.ResultsWe conclude that there are good reasons to grant for-profit companies access to medical data if they meet certain conditions: among others they need to respect patients’ informational rights and their actions need to be compatible with the public’s interest in health benefit from ReuseForPro. (shrink)

Background. The SA DTA was updated to better serve the South African research community by providing clarity on exactly when – at what stage during research – institutions hold rights to the data they generate in collaborative research contexts where raw data are received and integrated with other data. SA DTA 1.1 introduces significant enhancements in data governance, focusing on the explicit definition and management of ‘inferential data’. Objectives. To introduce SA DTA 1.1 and demonstrate (...) its practical application in real-world research contexts. Methods. Using a descriptive design, the study presents two practical case studies: one involving two universities and another a university and a pharmaceutical company. The first case focuses on identifying genetic markers for neurodegenerative diseases, while the second addresses genetic markers for chronic diseases. Results. In the first case study, the entity conducting the analysis generated inferential data independently, thereby gaining sole rights in such data. In the second case study, both collaborating entities contributed to the data analysis, leading to joint rights in the inferential data. These findings demonstrate the flexibility and clarity provided by SA DTA 1.1 in managing data ownership and intellectual property rights. Conclusion. SA DTA 1.1 significantly advances data governance by providing clear and adaptable guidelines for data rights in collaborative research. It supports ethical and efficient data sharing, protecting researchers’ interests and fostering global scientific innovation. (shrink)

This article explores some of the roles of computer simulation in measurement. A model-based view of measurement is adopted and three types of measurement—direct, derived, and complex—are distinguished. It is argued that while computer simulations on their own are not measurement processes, in principle they can be embedded in direct, derived, and complex measurement practices in such a way that simulation results constitute measurement outcomes. Atmospheric data assimilation is then considered as a case study. This practice, which involves combining (...) information from conventional observations and simulation-based forecasts, is characterized as a complex measuring practice that is still under development. The case study reveals challenges that are likely to resurface in other measuring practices that embed computer simulation. It is also noted that some practices that embed simulation are difficult to classify; they suggest a fuzzy boundary between measurement and non-measurement. 1 Introduction2 A Contemporary View of Measurement3 Three Types of Measurement4 Can Computer Simulations Measure Real-World Target Systems?5 Case Study: Atmospheric Data Assimilation5.1 Why data assimilation?5.2 A complex measuring practice under development5.3 Epistemic iteration6 The Boundaries of Measurement7 Epistemology, Not Terminology. (shrink)

The explosion of data grows at a rate of roughly five trillion bits a second, giving rise to greater urgency in conceptualizing the infosphere and understanding its implications for knowledge and public policy. Philosophers of technology and information technologists alike who wrestle with ontological and epistemological questions of digital information tend to emphasize, as Floridi does, information as our new ecosystem and human beings as interconnected informational organisms, inforgs at home in ambient intelligence. But the linguistic and conceptual representations (...) of Big Data—the massive volume of both structured and unstructured data—and the real world practice of data-mining for patterns and meaningful interpretation of evidence reveal tension and ambiguity in the bold promise of data analytics. This paper explores the tacit epistemology of the rhetoric and representation of Big Data and suggests a richer account of its ambiguities and the paradox of its real world materiality. We argue that Big Data should be recognized as manifesting multiple and conflicting trajectories that reflect human intentionality and particular patterns of power and authority. Such patterns require attentive exploration and moral appraisal if we are to resist simplistic informationist ontologies of Big Data, and the subtle forms of control in the political ecology of Big Data that undermine its promise as transformational knowledge. (shrink)

We offer a particularist defense of conspiratorial thinking. We explore the possibility that the presence of a certain kind of evidence—what we call "fortuitous data"—lends rational credence to conspiratorial thinking. In developing our argument, we introduce conspiracy theories and motivate our particularist approach (§1). We then introduce and define fortuitous data (§2). Lastly, we locate an instance of fortuitous data in one real world conspiracy, the Watergate scandal (§3).

Using biased data to train Artificial Intelligence (AI) algorithms will lead to biased decisions, discriminating against certain groups or individuals. Bias can be explicit (one or several protected features directly influence the decisions) or implicit (one or several protected features indirectly influence the decisions). Unsurprisingly, biased patterns are difficult to detect and mitigate. This paper investigates the extent to which explicit and implicit against one or more protected features in structured classification data sets can be mitigated simultaneously while (...) retaining the data’s discriminatory power. The main contribution of this paper concerns an optimization-based bias mitigation method that reweights the training instances. The algorithm operates with numerical and nominal data and can mitigate implicit and explicit bias against several protected features simultaneously. The trade-off between bias mitigation and accuracy loss can be controlled using parameters in the objective function. The numerical simulations using real-world data sets show a reduction of up to 77% of implicit bias and a complete removal of explicit bias against protected features at no cost of accuracy of a wrapper classifier trained on the data. Overall, the proposed method outperforms the state-of-the-art bias mitigation methods for the selected data sets. (shrink)

It is commonplace to encounter nonstationary or heterogeneous data, of which the underlying generating process changes over time or across data sets. Such a distribution shift feature presents both challenges and opportunities for causal discovery. In this paper we develop a principled framework for causal discovery from such data, called Constraint-based causal Discovery from Nonstationary/heterogeneous Data, which addresses two important questions. First, we propose an enhanced constraint-based procedure to detect variables whose local mechanisms change and recover (...) the skeleton of the causal structure over observed variables. Second, we present a way to determine causal orientations by making use of independence changes in the data distribution implied by the underlying causal model, benefiting from information carried by changing distributions. Experimental results on various synthetic and real-world data sets are presented to demonstrate the efficacy of our methods. (shrink)

We present an improved method for clustering in the presence of very limited supervisory information, given as pairwise instance constraints. By allowing instance-level constraints to have spacelevel inductive implications, we are able to successfully incorporate constraints for a wide range of data set types. Our method greatly improves on the previously studied constrained -means algorithm, generally requiring less than half as many constraints to achieve a given accuracy on a range of real-world data, while also being (...) more robust when over-constrained. We additionally discuss an active learning algorithm which increases the value of constraints even further. (shrink)

The business environment is increasingly uncertain due to the rapid development of disruptive information technologies, the changing global economy, and the COVID-19 pandemic. This brings great uncertainties to investors to predict the performance changes and risks of companies. This research proposes a sequential data-based framework that aggregates data from multiple sources including both structured and unstructured data to predict the performance changes. It leverages data generated from the early risk warning system in China stock market to (...) measure and predict organization performance changes based on the risk warning status changes of public companies. Different from the models in existing literature that focus on the prediction of risk warning of companies, our framework predicts a portfolio of organization performance changes, including business decline and recovery, thus helping investors to not only predict public company risks, but also discover investment opportunities. By incorporating sequential data, our framework achieves 92.3% macro-F1 value on real-world data from listed companies in China, outperforming other static models. (shrink)

Data science is characterized by engaging heterogeneous data to tackle real world questions and problems. But data science has no data of its own and must seek it within real world domains. We call this search for data “prospecting” and argue that the dynamics of prospecting are pervasive in, even characteristic of, data science. Prospecting aims to render the data, knowledge, expertise, and practices of worldly domains available and tractable (...) to data science method and epistemology. Prospecting precedes data synthesis, analysis, or visualization, and is constituted by the upstream work of discovering disordered or inaccessible data resources, thereafter to be ordered and rendered available for computation. Through this work, data science positions itself in the middle of all things—capable of engaging this, that, or any domain—and thus prospecting is a key driver of data science’s ongoing formation as a universal science. (shrink)

We offer a particularist defense of conspiratorial thinking.We explore the possibility that the presence of a certain kind of evidence—what we call “fortuitous data”—lends rational credence to conspiratorial thinking. In developing our argument, we introduce conspiracy theories and motivate our particularist approach (§1).We then introduce and define fortuitous data (§2). Lastly, we locate an instance of fortuitous data in one real world conspiracy, the Watergate scandal (§3).