Epistemic Components Research Articles

The Importance of Being Consistent: Attribution of Mental States in Strategic Human-Robot Interactions.

This article investigates the attribution of mental state (AMS) to an anthropomorphic robot by humans in a strategic interaction. We conducted an experiment in which human subjects are paired with either a human or an anthropomorphic robot to play an iterated Prisoner's Dilemma game, and we tested whether AMS is dependent on the robot "consistency," that is, the correspondence between the robot's verbal reaction and its behavior after a nonoptimal social outcome of the game is obtained. We find that human partners are attributed a higher mental state level than robotic partners, regardless of the partner's consistency between words and actions. Conversely, the level of AMS assigned to the robot is significantly higher when the robot is consistent in its words and actions. This finding is robust to the inclusion of psychological factors such as risk attitude and trust, and it holds regardless of subjects' initial beliefs about the adaptability of the robot. Finally, we find that when the robot apologizes for its behavior and defects in the following stage, the epistemic component of the AMS significantly increases.

Research repertoires and boundary work in the search for extra-terrestrial intelligence (SETI)

ABSTRACT The idea of scouting the sky in search of extra-terrestrial signals (SETI) was first proposed in the late 1950s; soon afterwards, its scope was formalised in the so-called Drake Equation, a probabilistic argument to gauge the number of ‘civilisations’ capable of radio communication on the basis of astronomical, biological and cultural factors. Since then, radio telescopes around the world have been searching for an artificial signal from elsewhere in the universe, a sign of extraterrestrial communication. How have SETI practitioners claimed scientific legitimacy and epistemic authority over the search for extra-terrestrial life, a matter often regarded to be at the very periphery of science? And how have they succeeded (or failed) in maintaining them over time? The historical trajectory of SETI offers the opportunity to investigate how, over time, an interdisciplinary community coalesced around a research repertoire and navigated ebbs and flows of interest by rearticulating the methods and scope of their research. The ‘probability repertoire’ around which the SETI community coalesced is an evolving entity, entailing significant boundary work. As with every successful repertoire, it is a flexible but always entailing a recognisable alignment of material, social and epistemic components instrumental for a research community in gaining recognition and negotiating its epistemic positioning within the always shifting boundaries of science.

Informing synthetic passive microwave predictions through Bayesian deep learning with uncertainty decomposition

Abstract Space-borne passive microwave (PMW) data provide rich information on atmospheric state, including cloud structure and underlying surface properties. However, PMW data are sparse and limited due to low Earth orbit collection, resulting in coarse Earth system sampling. This study demonstrates that Bayesian deep learning (BDL) is a promising technique for predicting synthetic microwave (MW) data and its uncertainties from more ubiquitously available geostationary infrared observations. Our BDL models decompose predicted uncertainty into aleatoric (irreducible) and epistemic (reducible) components, providing insights into uncertainty origin and guiding model improvement. Low and high aleatoric uncertainty values are characteristic of clear sky and cloudy regions, respectively, suggesting that expanding the input feature vector to allow richer information content could improve model performance. The initially high average epistemic uncertainty metrics quantified by most models indicate that the training process would benefit from a greater data volume, leading to improved performance at most studied MW frequencies. Using quantified epistemic uncertainty to select the most useful additional training data (a training dataset size increase of 3.6%), the study reduced the mean absolute error and root mean squared error by 1.74% and 1.38%, respectively. The broader impact of this study is the demonstration of how predicted epistemic uncertainty can be used to select targeted training data. This allows for the curation of smaller, more optimized training datasets and also allows for future active learning studies.

A model for enhancing creativity, collaboration and pre-professional identities in technology-supported cross-organizational communities of practice

This research proposes that technology-supported cross-organizational (university-industry) Communities of Practice (CoPs), which are integrated into the Design Studies curriculum in Higher Education, can foster robust university-industry collaborations. These can help bridge the reported gap between the actual versus the expected soft skills and personae of young graduates transitioning to the creative industries today. CoPs are groups of people who share a common interest in an area of ‘endeavor’ and connect to co-create competence in that area through their practice.This paper makes two overarching research contributions. First, it informs about the design, enactment, and evaluation of a student CoP in an undergraduate Design course which was expanded to include members from the industry as clients, alumni mentors, and expert evaluators. Drawing from rich empirical data, the paper explains the designed and emergent learning phenomena of CoP participation and its effects on the students’ creative and socio-epistemic outcomes, as well as their pre-professional identities. Second, it presents a governance model with three sets of actionable guidelines, namely the Set (technology), the Social (collaborative), and the Epistemic (learning) components. The entire body of work validates the critical interlocking of these components to form a robust social learning model that appropriates the complex practices of cross-organizational CoPs in Higher Education Design studies.

Narrative authority in online video: Legacy media and digital native production styles

This project examines the way journalists establish narrative authority in online video news through their strategic uses of epistemological production elements. Journalists have historically constructed video news stories with a combination of linguistic and photographic epistemic components to assert their role as authoritative arbiters of factual reality. Online news makes extensive use of video, but production formats and their associated epistemic strategies vary, largely in accordance with the traditions and values of different organizations. This study represents one part of a larger content analysis of video news narrative. It examines the epistemological choices in online video news narratives as produced by legacy print organizations (LP), TV stations (TV), or digital natives (DN). The results show that legacy traditions still divide print and TV organizations, while authority in DN video is constructed in a more diffuse and participatory style reflective of online culture.

A visual learning analytics approach for knowledge building: Impact on students' epistemic understanding of discourse, productive inquiry and domain knowledge

AbstractIn this study, we examine the design process for a visual learning analytics (VLA) approach in addition to its effect on developing students' epistemic understanding of discourse and relationships to promote knowledge‐building inquiry. Over a 3‐month period, an experimental class of secondary school students took a visual arts course using a knowledge‐building design augmented by VLA, while a comparison class took the same course within a regular knowledge‐building environment. The VLA approach implemented in the experimental class emphasized collective reflections using assessment tools to visualize the students' online and classroom discussions. The analysis of the students' questionnaire responses and Knowledge Forum discourse showed that the VLA approach helped the students develop a deeper epistemic understanding of discourse for productive inquiry and greater domain knowledge. In addition, the students' epistemic understanding was associated with their performance in productive Knowledge Forum inquiry. Qualitative analysis of the classroom and group discussions, interview responses and learning artefacts showed that the VLA design helped them understand the nature of discourse deeply by encouraging them to notice community orientations, increase their awareness of productive discourse moves and advance idea development. The implications of the VLA design in improving epistemic understanding of discourse are discussed. We contribute to the literature by examining the epistemic dimensions of discourse and implementing VLA as an epistemic tool for knowledge building. Practitioner notesWhat is already known about this topic Supporting students to develop a meta‐level understanding, ie, their epistemic understanding of the nature of discourse, is important for promoting productive inquiry but continues to be challenging in education, and limited research has focused on the epistemic aspects of discourse. Little research has examined the role of a computer‐supported knowledge‐building environment augmented by a visual learning analytics approach on students' epistemic understanding of discourse, productive inquiry and domain knowledge. What this paper adds A knowledge‐building design augmented by a visual learning analytics approach linking online and face‐to‐face discourse supported students' epistemic understanding of discourse, productive inquiry and domain knowledge. VLA‐supported collective reflection in knowledge‐building facilitated students' noticing of community orientations, awareness of productive discourse moves and their idea development. These metacognitive processes helped students develop deeper epistemic understanding of discourse, productive inquiry and greater domain knowledge. Implications for practice and/or policy Epistemic understanding of the nature of discourse acts as a medium for promoting productive knowledge‐building inquiry and learning outcomes, and it is critical to help students develop meta‐level understanding by combining discourse content and epistemic components. It is important to create a VLA‐supported community that provides VLA‐related data‐supported collective reflection opportunities to help students engage in productive inquiry and knowledge building. Teachers can scaffold students to link online and face‐to‐face discourse and help them inquire into and reflect on discourse processes.

Markets and knowledge commons: Is there a difference between private and community governance of markets?

It is well-established that market governance can be provided by both public (state) and private organizations. However, the concept of private governance has been used, this article contends, to refer to two distinct forms of non-state governance: private governance and community governance. We distinguish between these two forms, arguing that private governance should be understood as the provision of market governance by (external) private parties, while community governance refers to a process where a group, a community, or society has the autonomy to govern its own affairs without interference from external authorities. The former internalizes the externalities associated with governance, while the latter comes about mainly as an unintended externality of social interaction in markets. To further illuminate the differences, and the relative strengths of these types of non-state governance, we distinguish among three elements of market governance: (1) the formation and interpretation of rules, (2) the administration of rules of ownership and exchange, and (3) the enforcement of rules. We argue that community governance is of great relevance for the formation and interpretation of the rules of ownership and exchange, which is consequently very hard to outsource to external parties, private or public. Community governance also plays a frequently overlooked role in administration and enforcement through the process of co-production. Rule formation and interpretation are theorized as the epistemic components of market governance, which can be analyzed within the Governing Knowledge Commons framework.

Epistemic mag revisited: Assessing the (ir)relevance of propositions

This paper investigates the epistemic use of the German modal mag. Recent studies have suggested that it might be developing into a concessive marker, possibly losing its epistemic component of assessing the factuality of a proposition. Using a corpus‑based analysis of its context features, this paper argues that mag still retains some of its epistemic characteristics and is functionally close to markers of irrelevance.

Comparative evaluation of uncertainty estimation and decomposition methods on liver segmentation

PurposeDeep neural networks need to be able to indicate error likelihood via reliable estimates of their predictive uncertainty when used in high-risk scenarios, such as medical decision support. This work contributes a systematic overview of state-of-the-art approaches for decomposing predictive uncertainty into aleatoric and epistemic components, and a comprehensive comparison for Bayesian neural networks (BNNs) between mutual information decomposition and the explicit modelling of both uncertainty types via an additional loss-attenuating neuron.MethodsExperiments are performed in the context of liver segmentation in CT scans. The quality of the uncertainty decomposition in the resulting uncertainty maps is qualitatively evaluated, and quantitative behaviour of decomposed uncertainties is systematically compared for different experiment settings with varying training set sizes, label noise, and distribution shifts.ResultsOur results show the mutual information decomposition to robustly yield meaningful aleatoric and epistemic uncertainty estimates, while the activation of the loss-attenuating neuron appears noisier with non-trivial convergence properties. We found that the addition of a heteroscedastic neuron does not significantly improve segmentation performance or calibration, while slightly improving the quality of uncertainty estimates.ConclusionsMutual information decomposition is simple to implement, has mathematically pleasing properties, and yields meaningful uncertainty estimates that behave as expected under controlled changes to our data set. The additional extension of BNNs with loss-attenuating neurons provides no improvement in terms of segmentation performance or calibration in our setting, but marginal benefits regarding the quality of decomposed uncertainties.

Active Bayesian Deep Learning With Vector Sensor for Passive Sonar Sensing of the Ocean

Multiple studies have explored and reported on using passive sonar scalar pressure fields for both biological and man-made target classification. Here, inclusion of the vector field is considered and evaluated as an input feature in the classification of ship noises. We find that the inclusion of vector field information can significantly improve the understanding of the classification model performance. Based on an analysis of the epistemic and aleatoric uncertainties of probabilistic classification models, using vector field information as additional inputs leads to an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula> score of 0.8 in classification, comparable to scalar pressure field inputs only, but with 39% lower epistemic uncertainty in favor of utilizing the vector field component. Moreover, we verify that a similar conclusion applies to the active learning scenario, where we demonstrate the ability to utilize vector field information and epistemic uncertainty estimates to train a model that achieves an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$F_{1}$</tex-math></inline-formula> score of 0.8 while using only 23% of the overall data. The ability to decompose uncertainty into an aleatoric and epistemic component leads to additional model explainability.

Concurrent Awareness Desire Satisfactionism

AbstractDesire satisfactionists are united by their belief that what makes someone well-off is the satisfaction of their desires. But this commitment obscures a number of underlying differences, since there are several theoretical choice points on the way to making this commitment precise. This article is about two of the most important choice points. The first concerns an epistemic requirement on well-being. Suppose that one's desire that P is satisfied. Must one also know (or believe, or justifiably believe) that one's desire that P is satisfied in order to benefit from P? If so, there is an epistemic requirement on well-being. The second concerns the time at which one benefits. Well-being is a temporal phenomenon: given that one benefits from the satisfaction of one's desire that P, when does one benefit? Perhaps one benefits at the times at which one desires P, or the times at which P obtains, or both. I defend a view I call “concurrent awareness desire satisfactionism”: one benefits only at times at which both one desires P and P obtains (concurrence) and one benefits only if one is aware that one's desire is satisfied (awareness). I motivate this view by showing how it gives us solutions to many of the canonical problems facing desire satisfactionism. Then I put the two parts of the view together and explore some of its further implications. Ultimately, I conclude that well-being is an organic unity composed of a desiderative component, an epistemic component, and a worldly component, none of which are valuable on their own, but which are valuable when they are related in the right way.

Practice-Crosscutting Concept pairs of the NGSS and ways of knowing in the Earth & Space Sciences

The epistemic components of science are included in instruction through the three-dimensional learning of the Next Generation Science Standards (NGSS). Students use the Science & Engineering Practices to engage the Disciplinary Core Ideas through the perspective of the Crosscutting Concepts. The NGSS focus on the epistemic aspects that all sciences share, but the domain-specific epistemic aspects, referred to here as ways of knowing, are still important to include in science education because they require different pedagogical methods. In this paper, I show how curriculum developers can incorporate the ways of knowing of the Earth & Space Sciences into their units through the choice of Practices and Crosscutting Concepts in the lesson-specific learning performance statements. I start by identifying the most common Practice-Crosscutting Concept pairs of the Earth & Space Sciences performance expectations and comparing them to the literature on the ways of knowing in the geosciences. The performance expectations represent the multiple modes of inquiry and systems thinking characteristic of the Earth & Space Sciences. Yet, they don’t effectively represent other aspects of the Earth & Space Sciences, like visualization, spatial reasoning, deep time, and large size and spatial scales. I then encourage curriculum developers to move beyond the example of the performance expectations by providing suggested Practice and Crosscutting Concept pairs and groupings that can be used by curriculum developers to communicate these other aspects. I also include example unit outlines to show what the suggestions might look like in application.

Forecasting thermoacoustic instabilities in liquid propellant rocket engines using multimodal Bayesian deep learning

We present a method that combines multiple sensory modalities in a rocket thrust chamber to predict impending thermoacoustic instabilities with uncertainties. This is accomplished by training an autoregressive Bayesian neural network model that forecasts the future amplitude of the dynamic pressure time series, using multiple sensor measurements (injector pressure/ temperature measurements, static chamber pressure, high-frequency dynamic pressure measurements, high-frequency OH* chemiluminescence measurements) and future flow rate control signals as input. The method is validated using experimental data from a representative cryogenic research thrust chamber. The Bayesian nature of our algorithms allows us to work with a dataset whose size is restricted by the expense of each experimental run, without making overconfident extrapolations. We find that the networks are able to accurately forecast the evolution of the pressure amplitude and anticipate instability events on unseen experimental runs 500 milliseconds in advance. We compare the predictive accuracy of multiple models using different combinations of sensor inputs. We find that the high-frequency dynamic pressure signal is particularly informative. We also use the technique of integrated gradients to interpret the influence of different sensor inputs on the model prediction. The negative log-likelihood of data points in the test dataset indicates that prediction uncertainties are well-characterized by our model and simulating a sensor failure event results in a dramatic increase in the epistemic component of the uncertainty, as would be expected when a Bayesian method encounters unfamiliar, out-of-distribution inputs.

Two comparative studies of computer simulations and experiments as learning tools in school and out-of-school education

Interactive computer simulations and hands-on experiments are important teaching methods in modern science education. Especially for the communication of complex current topics with social relevance (socioscientific issues), suitable methods in science education are of great importance. However, previous studies could not sufficiently clarify the educational advantages and disadvantages of both methods and often lack adequate comparability. This paper presents two studies of direct comparisons of hands-on experiments and interactive computer simulations as learning tools in science education for secondary school students in two different learning locations (Study I: school; Study II: student laboratory). Using a simple experimental research design with type of learning location as between-subjects factor (NStudy I = 443, NStudy II = 367), these studies compare working on computer simulations versus experiments in terms of knowledge achievement, development of situational interest and cognitive load. Independent of the learning location, the results showed higher learning success for students working on computer simulations than while working on experiments, despite higher cognitive load. However, working on experiments promoted situational interest more than computer simulations (especially the epistemic and value-related component). We stated that simulations might be particularly suitable for teaching complex topics. The findings reviewed in this paper moreover imply that working with one method may complement and supplement the weaknesses of the other. We conclude that that the most effective way to communicate complex current research topics might be a combination of both methods. These conclusions derive a contribution to successful modern science education in school and out-of-school learning contexts.

Geografía histórica y tiempo geográfico, concepto y superación de dicotomías

The space-related dichotomies of time/space, society/nature and space/society are still a complex methodological challenge for Geography. Simultaneously, Historical Geography is a sub-area capable of reading and producing a dialectic representation of space, by discourse, cartography and images. Thus, this article proposes the concept of geographical time, epistemic component and analytic entity of trans-scalar and trans-temporal spatial issues, operated by historical cartography. Methodologically, it correlates: i) Historical Geography and the scalar representations of space; (ii) cartographical representation as means of overcoming the time/space dichotomy; and (iii) a cartographical praxis of geographical time, in support of society↔nature and space↔society dialectics, substantiated in the eighteenth century Colonial Brazil.

Computation of epistemic uncertainty due to limited data samples in small field dosimetry using Fuzzy Set Theory.

To estimate the epistemic (or fuzzy) uncertainty, arising due to limited data samples in the measurement of the output factors (OFs) of the small fields using Fuzzy Set Theory (FST). EBT3 film samples of size 50 × 50 mm2 were used for the measurement of the OF of stereotactic radiosurgery (SRS) cones of size 4, 6, 7.5, 10, 12.5 and 15 mm diameter, normalized with respect to the output of 100 × 100 mm2 open field size. Three measurements were done per cone/field size. Red color channel was chosen for the dosimetry purpose, net optical density (NOD) was converted to the dose using non-linear relation. To estimate the epistemic uncertainty associated with the measured OFs due to limited number of data samples, a triangular fuzzy number (TFN) was assumed as the fuzziness in the dose delivered by the individual SRS cone/field. Uncertainty in the OF was estimated by applying the Fuzzy Vertex Method (FVM). The membership functions of the OF were constructed for each cone size and the nature of the uncertainty in the OF of the cones was expressed in the terms of its fuzziness. For the sake of completeness of the study, the statistical uncertainty involved in the procedure has also been calculated. The statistical and fuzzy uncertainties in the measurement of OF of cones range from 3.28 to 6.25% and 2.58 to 5.44% respectively. The smallest cone of 4 mm has the largest values of statistical and fuzzy uncertainties. The membership functions of the OF for the studied cones were triangular in nature. The epistemic uncertainty arising due to limited number of data samples holds a significant fraction of the prescribed dose, and therefore, should not be ignored in the total uncertainty estimation. This study highlights the significance of epistemic component of measurement uncertainty arising out due to the insufficient/limited number of measurements of a quantity.

Whence the Expected Free Energy?

The expected free energy (EFE) is a central quantity in the theory of active inference. It is the quantity that all active inference agents are mandated to minimize through action, and its decomposition into extrinsic and intrinsic value terms is key to the balance of exploration and exploitation that active inference agents evince. Despite its importance, the mathematical origins of this quantity and its relation to the variational free energy (VFE) remain unclear. In this letter, we investigate the origins of the EFE in detail and show that it is not simply "the free energy in the future." We present a functional that we argue is the natural extension of the VFE but actively discourages exploratory behavior, thus demonstrating that exploration does not directly follow from free energy minimization into the future. We then develop a novel objective, the free energy of the expected future (FEEF), which possesses both the epistemic component of the EFE and an intuitive mathematical grounding as the divergence between predicted and desired futures.

Who Is a Wise Person? Zhuangzi and Epistemological Discussions of Wisdom

This essay articulates the contribution that the Zhuangzi can make to contemporary epistemological discussions of wisdom. It suggests that wisdom in the Zhuangzi involves, in part, correctly distinguishing the "heavenly" (or the naturally given) from human artifice. It is important for humanity to understand naturally given conditions (e.g., seasons, climate, forces, mortality) to grasp what is within, and what beyond, our initiatives. To enable this, we need to be openly engaged with the world, rather than approach it with rigid convictions about outcomes or goals. We characterize such openness and readiness to engage as an attitude, that of "epistemic humility." In building up our position, we assess two themes in analytic philosophical debates on wisdom: the epistemic component of wisdom and the domain of wisdom. This helps establish what we believe is an important contribution of the Zhuangzi to the discussion on wisdom, that is, the wise person's epistemic humility.

What Makes a Reporter Human?

The objective of this article is to lay out a research agenda, named “Augmented Journalism”, to explore the future of news journalism in the context of Artificial Intelligence (AI). The article will determine what the critical epistemic components of this novel paradigm would be. “Augmented journalism” puts the focus on human-machine interaction, and the main question that arises from this context is “What makes a reporter human?” A broad intradisciplinary approach to human capabilities is needed, not just in the social sciences, in particular sociology, journalism studies, and psychology, but also in technical skills (computer science and user design), as well as in the operational context (media economics and management studies). The vision presented here stems from years of research into journalism practice, AI tools for news, news media business models and the Silicon Valley mindset. The underlying logic is that the future of journalism lies in its capacity to create value for the various sectors of stakeholders, from consumers to policy makers. The project will explore what makes the work of journalists valuable, what part consists of “human” capabilities, and how news work can be augmented with AI in order to stay relevant, but efficient.

Reportatives "sollen"+Infinitiv in Interrogativsätzen

In this paper, the reportive construction sollen+infinitive is treated as merely agnostic, with the negative epistemic component emerging qua conversational implicature. The paper aims to test whether the interrogative sentence type can be seen as a contextual trigger for this implicature. It concludes that the negative epistemic implicature can be triggered or strengthened by proper questions and mirative questions, but not by rhetorical or unresolvable questions as defined by Celle (2018).