To send content items to your account,
please confirm that you agree to abide by our usage policies.
If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account.
Find out more about sending content to .
To send content items to your Kindle, first ensure firstname.lastname@example.org
is added to your Approved Personal Document E-mail List under your Personal Document Settings
on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part
of your Kindle email address below.
Find out more about sending to your Kindle.
Note you can select to send to either the @free.kindle.com or @kindle.com variations.
‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi.
‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.
The preconception, pregnancy and immediate postpartum and newborn periods are times for mothers and their offspring when they are especially vulnerable to major stressors – those that are sudden and unexpected and those that are chronic. Their adverse effects can transcend generations. Stressors can include natural disasters or political stressors such as conflict and/or migration. Considerable evidence has accumulated demonstrating the adverse effects of natural disasters on pregnancy outcomes and developmental trajectories. However, beyond tracking outcomes, the time has arrived for gathering more information related to identifying mechanisms, predicting risk and developing stress-reducing and resilience-building interventions to improve outcomes. Further, we need to learn how to encapsulate both the quantitative and qualitative information available and share it with communities and authorities to mitigate the adverse developmental effects of future disasters, conflicts and migrations. This article briefly reviews prenatal maternal stress and identifies three contemporary situations (wildfire in Fort McMurray, Alberta, Canada; hurricane Harvey in Houston, USA and transgenerational and migrant stress in Pforzheim, Germany) where current studies are being established by Canadian investigators to test an intervention. The experiences from these efforts are related along with attempts to involve communities in the studies and share the new knowledge to plan for future disasters or tragedies.
Antibodies at gastrointestinal mucosal membranes play a vital role in immunological protection against a range of pathogens, including helminths. Gastrointestinal health is central to efficient livestock production, and such infections cause significant losses. Fecal samples were taken from 114 cattle, across three beef farms, with matched blood samples taken from 22 of those animals. To achieve fecal antibody detection, a novel fecal supernatant was extracted. Fecal supernatant and serum samples were then analysed, using adapted enzyme-linked immunosorbent assay protocols, for levels of total immunoglobulin (Ig)A, IgG, IgM, and Teladorsagia circumcincta-specific IgA, IgG, IgM and IgE (in the absence of reagents for cattle-specific nematode species). Fecal nematode egg counts were conducted on all fecal samples. Assays performed successfully and showed that IgA was the predominant antibody in fecal samples, whereas IgG was predominant in serum. Total IgA in feces and serum correlated within individuals (0.581, P = 0.005), but other Ig types did not. Results support the hypothesis that the tested protocols are an effective method for the non-invasive assessment of cattle immunology. The method could be used as part of animal health assessments, although further work is required to interpret the relationship between results and levels of infection and immunity.
Good bone quality in breeding ewes is important for the mineralisation of foetal skeletons and to sustain maternal dentition, as tooth loss is the main reason for culling sheep in the UK. Among other functions, bone is a storage depot for calcium and other key minerals that are mobilised to meet major demands such as during lactation. As other studies in humans and poultry have shown, there is substantial genetic variation (h2 between 0.5 and 0.8) for bone properties, suggesting a similar situation in ewes. These properties, e.g. bone density, are key to successful production and nurturing of healthy lambs, which can be used in selective breeding strategies to extend breeding ewes’ productive lives. CT has been shown to be a useful method of assessing bone properties in sheep (Rubin et al., 2001). This study quantifies the main bone types in Scottish Blackface ewes and investigates environmental factors affecting bone quality.
In Chapter 6 we mirror closely the exposition given in the previous chapter on regression, beginning with the approximation of the underlying data generating function itself by bases of features, and going on to finally describing cross-validation in the context of classification. In short we will see that all of the tools from the previous chapter can be applied to the automatic design of features for the problem of classification as well.
Automatic feature design for the ideal classification scenario
In Fig. 6.1 we illustrate a prototypical dataset on which we perform the general task of two class classification, where the two classes can be effectively separated using a nonlinear boundary. In contrast to those examples given in Section 4.5, where visualization or scientific knowledge guided the fashioning of a feature transformation to capture this nonlinearity, in this chapter we suppose that this cannot be done due to the complexity and/or high dimensionality of the data. At the heart of the two class classification framework is the tacit assumption that the data we receive are in fact noisy samples of some underlying indicator function, a nonlinear generalization of the step function briefly discussed in Section 4.5, like the one shown in the right panel of Fig. 6.1. Akin to regression, our goal with classification is then to approximate this data-generating indicator function as well as we can using the data at our disposal.
In this section we will assume the impossible: that we have clean and complete access to every data point in the space of a two class classification environment, whose labels take on values in ﹛-1, 1﹜, and hence access to its associated indicator function y (x). Although an indicator function is not continuous, the same bases of continuous features discussed in the previous chapter can be used to represent it (near) perfectly.
Approximation of piecewise continuous functions
In Section 5.1 we saw how fixed and adjustable neural network bases of features can be used to approximate continuous functions. These bases can also be used to effectively approximate the broader class of piecewise continuous functions, composed of fragments of continuous functions with gaps or jumps between the various pieces.
In the last decade the user base of machine learning has grown dramatically. From a relatively small circle in computer science, engineering, and mathematics departments the users of machine learning now include students and researchers from every corner of the academic universe, as well as members of industry, data scientists, entrepreneurs, and machine learning enthusiasts. The book before you is the result of a complete tearing down of the standard curriculum of machine learning into its most basic components, and a curated reassembly of those pieces (painstakingly polished and organized) that we feel will most benefit this broadening audience of learners. It contains fresh and intuitive yet rigorous descriptions of the most fundamental concepts necessary to conduct research, build products, tinker, and play.
Intended audience and book pedagogy
This book was written for readers interested in understanding the core concepts of machine learning from first principles to practical implementation. To make full use of the text one only needs a basic understanding of linear algebra and calculus (i.e., vector and matrix operations as well as the ability to compute the gradient and Hessian of a multivariate function), plus some prior exposure to fundamental concepts of computer programming (i.e., conditional and looping structures). It was written for first time learners of the subject, as well as for more knowledgeable readers who yearn for a more intuitive and serviceable treatment than what is currently available today.
To this end, throughout the text, in describing the fundamentals of each concept, we defer the use of probabilistic, statistical, and neurological views of the material in favor of a fresh and consistent geometric perspective. We believe that this not only permits a more intuitive understanding of many core concepts, but helps establish revealing connections between ideas often regarded as fundamentally distinct (e.g., the logistic regression and support vector machine classifiers, kernels, and feed-forward neural networks). We also place significant emphasis on the design and implementation of algorithms, and include many coding exercises for the reader to practice at the end of each chapter. This is because we strongly believe that the bulk of learning this subject takes place when learners “get their hands dirty” and code things up for themselves.
In Sections 3.2 and 4.5 we have discussed how understanding of regression and classification datasets can be used to forge useful features in particular instances. With regression we saw that by visualizing low-dimensional data we could form excellent features for particular datasets like e.g., data from Galileo's classic ramp experiment. Later, when discussing classification, we also saw how basic features can be designed for e.g., image data using our understanding of natural signals and the mammalian visual processing system. Unfortunately, due to our general ignorance regarding most types of phenomena in the universe, instances such as these are rare and we often have no knowledge on which to construct reasonable features at all. However, we can, as described in the next three chapters, automate the process of feature design itself by leveraging what we know strong features should accomplish for regression/classification tasks.
As discussed in the end of Section 3.2, rarely can we design perfect or even strongly performing features for the general regression problem by completely relying on our understanding of a given dataset. In this chapter we describe tools for automatically designing proper features for the general regression problem, without the explicit incorporation of human knowledge gained from e.g., visualization of the data, philosophical reflection, or domain expertise.
We begin by introducing the tools used to perform regression in the ideal but extremely unrealistic scenario where we have complete and noiseless access to all possible input feature/output pairs of a regression phenomenon, i.e., a continuous function (as first discussed in Section 3.2). Here we will see how, in the case where we have such unfettered access to regression data, perfect features can be designed automatically by combining elements from a set of basic feature transformations. We then see how this process for building features translates, albeit imperfectly, to the general instance of regression where we have access to only noisy samples of a regression relationship. Following this we describe cross-validation, a crucial procedure to employing automatic feature design in practice. Finally we discuss several issues pertaining to the best choice of primary features for automatic feature design in practice.
Automatic feature design for the ideal regression scenario
In Fig. 5.1 we illustrate a prototypical dataset on which we perform regression, where our input feature and output have some sort of clear nonlinear relationship. Recall from Section 3.2 that at the heart of feature design for regression is the tacit assumption that the data we receive are in fact noisy samples of some underlying continuous function (shown in dashed black in Fig. 5.1). Our goal in solving the general regression problem is then, using the data at our disposal (which we may think of as noisy glimpses of the underlying function), to approximate this data-generating function as well as we can.
In this section we will assume the impossible: that we have complete access to a clean version of every input feature/output pair of a regression phenomenon, or in other words that our data completely traces out a continuous function y (x).
Machine learning is a rapidly growing field of study whose primary concern is the design and analysis of algorithms which enable computers to learn. While still a young discipline, with much more awaiting to be discovered than is currently known, today machine learning can be used to teach computers to perform a wide array of useful tasks. This includes tasks like the automatic detection of objects in images (a crucial component of driver-assisted and self-driving cars), speech recognition (which powers voice command technology), knowledge discovery in the medical sciences (used to improve our understanding of complex diseases), and predictive analytics (leveraged for sales and economic forecasting). In this chapter we give a high level introduction to the field of machine learning and the contents of this textbook. To get a big picture sense of how machine learning works we begin by discussing a simple toy machine learning problem: teaching a computer how to distinguish between pictures of cats from those with dogs. This will allow us to informally describe the procedures used to solve machine learning problems in general.
Teaching a computer to distinguish cats from dogs
To teach a child the difference between “cat” versus “dog”, parents (almost!) never give their children some kind of formal scientific definition to distinguish the two; i.e., that a dog is a member of Canis Familiaris species from the broader class of Mammalia, and that a cat while being from the same class belongs to another species known as Felis Catus. No, instead the child is naturally presented with many images of what they are told are either “dogs” or “cats” until they fully grasp the two concepts. How do we know when a child can successfully distinguish between cats and dogs? Intuitively, when they encounter new (images of) cats and dogs, and can correctly identify each new example. Like human beings, computers can be taught how to perform this sort of task in a similar manner. This kind of task, where we aim to teach a computer to distinguish between different types of things, is referred to as a classification problem in machine learning.