The basic idea behind the “accelerated discovery” of this book’s title is that by using the ability of a computer to process large quantities of data, one can find interconnections that lead to new hypotheses. To take one example as typical, consider the case of drug repurposing in which one seeks to find new applications for a drug that is already approved for use on humans. By scanning information from medical abstracts on drugs and on medical conditions, the system seeks links between the drug and conditions that suggest a mechanism by which the drug might affect the condition. This suggests hypotheses for future research.

The chapters of the book fall into two sections. In the first, a number of techniques for hypothesis discovery are described. These involve scanning data from online sources to construct tables and networks that encapsulate the data.

The scanning of the data makes use of natural language techniques to extract relationships between entities. An example might be that a certain drug is effective against malaria so the drug and malaria become entities of interest. These relationships can then be organized and illuminated in several ways. An important technique is that of a co-occurrence table, which shows how frequently each pair of entities occurs together in the data. Relationships can also be illustrated through networks and graphs. An important aspect of these organizational structures is that they can be displayed in a way that allows the scientist to see the data in a meaningful way that may suggest hypotheses for future research. So color is used to highlight possible areas of interest in the graphical representations.

In the second half of the book, a large number of applications are described in greater detail. Most of these applications are in the area of biology/medicine. The author mentions that there are other areas in which the ideas have been used; however, because these are covered by nondisclosure agreements, they do not appear in the book. Different applications use different techniques, so these chapters offer an interesting series of snapshots of the accelerated discovery process. One aspect of the process that is not discussed, which perhaps should have been, is the kind of computing power that the process requires.

The presentation is clear with a practitioner’s viewpoint. The book would be an excellent starting place for someone wishing to learn more about the techniques. A student could easily read this book and try her hand at accelerated discovery. A final chapter suggests areas for future work both in applications and on the computing structures required for the process.

It is regrettable that the color illustrations are placed in a central section of the book; surely modern typesetting tools make it easy to incorporate color illustrations anywhere in the book. One side effect of this is that sometimes the book refers to color highlighting in an illustration for which only the grayscale picture is present.