Objective Functions, Deep Learning and Random Forests

Introduction: Science A computer scientist seems an odd choice to speak either about science in the forest, or science in the past. Computer science is more often located in cities and offices than in forests, and is concerned with the challenges of the future rather than the past. ‘Science’ appears to be a point of enquiry shared with this symposium, but even this word is open to debate. It is often observed that a discipline including the word ‘science’ in its name introduces doubt as to why the claim is necessary. While Cambridge has long admired natural philosophy and the ‘natural sciences’, computer science fails to qualify as one of them. We computer scientists do not study nature, but only what we make ourselves. A computer scientist is perhaps more akin to a novelist, sculptor or carpenter than to an astronomer or entomologist. This may be why computer scientists are unusually sensitive to the question of whether their work is objective.