Title: Hard to Cheat: A Turing Test based on Answering Questions about Images

Abstract: Progress in language and image understanding by machines has sparkled the interest of the research community in more open-ended, holistic tasks, and refueled an old AI dream of building intelligent machines. We discuss a few prominent challenges that characterize such holistic tasks and argue for "question answering about images" as a particular appealing instance of such a holistic task. In particular, we point out that it is a version of a Turing Test that is likely to be more robust to over-interpretations and contrast it with tasks like grounding and generation of descriptions. Finally, we discuss tools to measure progress in this field.