Search Results

This introductory chapter presents an overview of the subsequent chapters in this book which will discuss topics such as epistemological theory, Statistical Prediction Rules, Strategic Reliabilism, and Standard Analytic Epistemology.

Functional magnetic resonance imaging (fMRI) has rapidly emerged as an important tool for understanding the neural correlates of myriad cognitive and emotional processes. By offering a direct window into brain activity, fMRI promised to usher in a new era of psychiatry, in which clinical decisions would be based on objective neurobiological measures rather than often-contentious symptom classifications. However, research conducted over the past two decades has failed to realise this goal, and fMRI remains an unutilized tool in the psychiatric clinic. In this chapter, we consider reasons why fMRI has failed to fulfill its promise as a clinically useful tool for psychiatrists and consider recent progress towards this goal in three key clinical domains: diagnosis, prediction and treatment monitoring. In particular, we focus on moving away from using traditional case-control comparisons of mean differences in brain activation to more sophisticated statistical techniques using brain activity patterns to classify individual participants to key clinical outcome variables, such as diagnosis or treatment response.

Jakob Hohwy seeks to recover an approach to consciousness from a general theory of brain function, namely the prediction error minimization theory. The way this theory applies to mental and developmental disorder demonstrates its relevance to consciousness. The resulting view is discussed in relation to a contemporary theory of consciousness, namely, the idea that conscious perception depends on Bayesian metacognition which is also supported by considerations of psychopathology. This Bayesian theory is first disconnected from the higher-order thought theory, and then, via a prediction error conception of action, connected instead to the global workspace theory. Considerations of mental and developmental disorder therefore show that a very general theory of brain function is relevant to explaining the structure of conscious perception. Furthermore, Hohwy argues that this theory can unify two contemporary approaches to consciousness in a move that seeks to elucidate the fundamental mechanism for the selection of representational content into consciousness.

The topics of language and emotion cannot be ignored in their roles in environmental history. Nor can that of religions, whose powers are not everywhere and all the time eclipsed by the worldviews of the sciences. Thus myth, symbol and value are all bound up together in any account of developments over time. Out of this complexity it can be discerned that a tension has always existed between a fragmentation of society which is parallelled in the non-human world, and coalescences which likewise are found in both spheres. This tension was lowest in forager times but has increased with each step in access to energy and materials. In society, phenomena like individualism have grown and in step, sequestration of nature into dedicated zones has taken place. At the other extreme, like-minded people are in instant contact with each other across the world and emission gases from one country affect the whole stratosphere. Each step in such processes has brought about its unique conditions and so history provides little insight into what the future may bring.

The electronic monitoring (EM) of offenders bestowed on community corrections a hitherto unattainable capacity for remote, real-time oversight (“immediacy”) over offenders’ movements and schedules. Emerging forms of algorithmic prediction may shift EM’s focus to “imminence detection” via smartphones and the vastly increased datasets they generate. Some EM vendors anticipate this, and even the National Institute of Justice is considering Artificial Intelligence (AI) as a means of managing community correctional processes. AI and smartphone monitoring could have a transformational impact on community corrections, and the cultural allure of predictivity will play a part in establishing that, but the perceived need to punish will impede the full organizational adoption of predictivity. Tangible immediacy will still matter more than nebulous imminence.

This chapter describes how the term “world” becomes a category of racial meaning. Tracking the new coining of phrases such as “colored world,” “yellow world,” and “brown world” during and after the Russo-Japanese War, it shows the ways in which this discourse constructs new ways of seeing that are linked less to scientific racism or social Darwinism and more to modes of futurology that animate race as a historical tendency in the world. It analyzes popular and influential histories of the future such as B.L. Putnam Weale’s Conflict of Colour (1909), Charles Pearson’s National Life and Character: A Forecast (1893), and Brooks Adams’s Law of Civilization and Decay: An Essay on History in relation to newspaper discourse on the yellow peril in order to see how race is used as forecast.

The third chapter introduces prediction as a Bayesian process that ensures survival. It then focuses on social prediction—predicting other people’s behavior—exploring consistently low correlations between traits (a form of social cognition) and behavior. It then describes the brain correlates of prediction violation, social reward, and social punishment, before arguing that social cognition may not be necessary to predict other people’s behavior; instead people may rely on other heuristics such as social norms. Therefore, it discounts the importance of social cognition for this aspect of human survival.

Following the outline of basic theory and evidence in chapters 7 and 8, this chapter sets the challenge of attempting to predict evolutionary dynamics in realistically diverse communities. Many challenges and opportunities facing human populations rely on being able to predict living systems. Even when a single focal species such as a pest or disease agent is of particular concern, its dynamics and responses to control measures always depend on interactions with a diverse set of other species. Even when the focus is on whole-ecosystem functioning, that depends on trait responses of constituent species. The chapter outlines several case studies where a multispecies evolutionary approach is required, including managing marine fisheries, controlling crop pests, and managing human microbiomes for improved health. To illustrate possible ways forwards, a model of evolution in a microbial community is presented, and possible methods for tracking evolution in diverse communities are discussed.

South Korea (hereinafter referred to as Korea) lies in the middle latitudes of the Northern Hemisphere. Until the 1960s, Korea was a typical agrarian country, with agriculture generating roughly half of its gross national product (GNP) and employing more than half of the labor force. Agriculture still plays an important role in the Korean national economy, but it accounts for a relatively much lower share of the GNP (5.3% in 1997) and engages much less of the population (11.0%). The agricultural share of the national economy is declining continuously. Farms in Korea, as in many other Asian countries, have traditionally been small. Average farm size has been growing slowly from 0.86 ha in 1960 to 1.39 ha in 2001, despite a significant reduction in the average number of persons per household engaged in farming—from 6.20 persons to 2.91 persons. As a result, agriculture has become more intensive. The country has four distinct seasons: summer, fall, winter, and spring. Summer and winter have a longer duration than spring or fall. The summer rainy season (Changma) in the Korean Peninsula includes the period from late June to late July. About three quarters of the annual precipitation falls during the summer season. The average annual precipitation in Korea is 1,274 mm, which is about 1.3 times the world average (973 mm). The variation in annual precipitation is larger, with an annual minimum of 784 mm and an annual maximum of 2675 mm in Seoul. Heavy rains fall during the Changma season, which is influenced by monsoons. The National Institute of Agricultural Science and Technology (NIAST) classified Korea (except Jeju Island) into 19 climate zones to efficiently use agricultural resources for wetland rice production. Among the 19 zones, zone 14, which is the Southern Charyeong Plain, yields the best harvest and the most stable rice production. Zones 11 (Yeongnam Basin), 17 (the northeastern coast), and 18 (the mid-eastern coast) are categorized as drought-risk areas at transplanting stage based on the ratio of evaporation to precipitation (Choi and Yun, 1989).

Sorell focuses on two state (police) uses of big data that have elicited concern: the creation of DNA databases and the use of past data to predict future crimes and criminals. In response to the former, Sorell argues that there is nothing intrinsically wrong with large-scale, indiscriminate databases of DNA profiles. These do not constitute an invasion of privacy, and nor do they necessarily render an entire population suspect, although he accepts that in the current climate they may be interpreted that way. As regards predictive policing, Sorell’s argument is that these uses are more concerning, basing future decisions on past information that may no longer be pertinent and could well be discriminatory.

Words are, if not exactly semantic ‘atoms’, at least assumed to be core elements of meaning. Their high degree of context dependence poses challenges to classic views of systematicity, however. In this chapter, I suggest an alternative view in which words are understood as cues to meaning rather than having intrinsic meaning themselves. This proposal is explored through experimental tests, and a computational model is developed that illustrates the properties of the ‘words-as-cues’ hypothesis.

In this chapter we propose precise operational criteria of systematicity that reveal a connection between the notion of systematicity and causal roles for category membership. We argue that neural network approaches that build on the assumption that grammatical knowledge is encoded implicitly, such as Elman's SRN, fall short of demonstrating systematic behavior precisely because such implicit knowledge plays no causal role in the network dynamics. On the other hand neural networks that employ explicit, encapsulated representations (i.e., representations that encapsulate contextual details) do enable categories to play causal roles. We draw upon insights from neurobiology to show how the hierarchical, columnar organization of the cortex in fact provides a basis for encapsulated representations that are invariant. We then sketch a novel approach to neural network modeling that illustrates how encapsulated representations can be operated on and dynamically bound into complex representations, producing rule-like, systematic behavior capable of dealing with hierarchical syntax.

This chapter details the progress toward operational numerical weather prediction. All three weather services—the Navy, the Air Force, and the Weather Bureau—which were members of the Joint Meteorological Committee of the Joint Chiefs of Staff, coordinated the details of the new Joint Numerical Weather Prediction Unit (JNWPU) to be located in Suitland, Maryland. Despite challenges, they produced the first “operational” weather map in May 1955—almost three years after deciding to move numerical weather prediction from the realm of research to operations. In so doing, the three weather services advanced numerical techniques more quickly than would have been possible in the less time-critical research environment.

The opening of the new Joint Numerical Weather Prediction Unit marked the end of the preliminary research period, but it was just the beginning of the worldwide spread of numerical weather prediction. This chapter briefly extends the story to the current time. As their very different meteorological missions exacerbated cultural differences, the Navy and the Air Force removed their personnel and formed their own operational prediction units, leaving the Weather Bureau to fund and staff its own center. As computer availability, processing speed, and memory capacity increased, universities began their own modeling and research projects. The modeling and prediction efforts of individual European nations joined forces to create the European Center for Mid-Range Weather Forecasting (ECMWF), which would provide formidable competition to US-based efforts. In time, modelers would attempt to forecast for longer and longer periods, until long-range forecasts took the first steps to becoming climate models.

We present a theory of dialogue as a form of cooperative joint activity. Dialogue is treated as a system involving two interlocutors and a shared workspace that contains their contributions and relevant non-linguistic context. The interlocutors construct shared plans and use them to “post” contributions to the workspace, to comprehend joint contributions, and to distribute control of the dialogue between them. A fundamental part of this process is to simulate their partner’s contributions and to use it to predict the upcoming state of the shared workspace. As a consequence, they align their linguistic representations and their representations of the situation and of the “games” underlying successful communication. The shared workspace is a highly limited resource, and the interlocutors use their aligned representations to say just enough and to speak in good time. We describe how dialogue and cooperative joint activity can be brought together in situated communication, and then sketch some implications of our account for human-like artificial systems.

This chapter focuses on two specific steps in the machine learning process, called model validation and model evaluation. Specifically, model validation is the step used to tune the hyperparameters of the model. Here, we often integrate a cross-validation process, which we discuss in detail in this chapter. Model evaluation, on the other hand, is the process of testing the performance of the model using unseen data, the test dataset. These processes are used to ensure that the model we developed through the algorithms discussed in Chapter 6 are reliable, given our data. The chapter will include labs to give you a practical introduction to these steps, given the modeling techniques discussed in the last chapter.