While psychologists, neuroscientists, and philosophers continue to debate about thenature of human emotions, machine learning researchers hurry to develop artificial systems capable of recognizing and synthesizing (i.e., artificially generating) emotions. Such efforts have been primarily motivated by the vast space of potential applications of emotion recognition and generation systems. Applications like personalized advertising, machine-assisted education, machine-assisted psychotherapy, employee assessment, elder care robots, machine-assisted mental health diagnosis, and emotion responsive gaming, are just a couple of examples. In this context, I aim to accomplish the following objectives: (1) to review the literature on emotion from a conceptual perspective, this is, how the concept of emotion has been understood and operationalized in the fields of psychology and machine learning; (2) to critically examine the machine learning literature regarding the conceptualization and measurement of emotion; (3) to identify areas of improvement and innovation in the conceptualization and measurement of emotion for basic and applied research, with special attention to the needs of machine learning researchers.