Deep Nets: What have They Ever Done for Vision?
- PDF / 1,868,496 Bytes
- 22 Pages / 595.276 x 790.866 pts Page_size
- 103 Downloads / 186 Views
Deep Nets: What have They Ever Done for Vision? Alan L. Yuille1 · Chenxi Liu1 Received: 10 January 2019 / Accepted: 9 November 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract This is an opinion paper about the strengths and weaknesses of Deep Nets for vision. They are at the heart of the enormous recent progress in artificial intelligence and are of growing importance in cognitive science and neuroscience. They have had many successes but also have several limitations and there is limited understanding of their inner workings. At present Deep Nets perform very well on specific visual tasks with benchmark datasets but they are much less general purpose, flexible, and adaptive than the human visual system. We argue that Deep Nets in their current form are unlikely to be able to overcome the fundamental problem of computer vision, namely how to deal with the combinatorial explosion, caused by the enormous complexity of natural images, and obtain the rich understanding of visual scenes that the human visual achieves. We argue that this combinatorial explosion takes us into a regime where “big data is not enough” and where we need to rethink our methods for benchmarking performance and evaluating vision algorithms. We stress that, as vision algorithms are increasingly used in real world applications, that performance evaluation is not merely an academic exercise but has important consequences in the real world. It is impractical to review the entire Deep Net literature so we restrict ourselves to a limited range of topics and references which are intended as entry points into the literature. The views expressed in this paper are our own and do not necessarily represent those of anybody else in the computer vision community. Keywords Deep neural networks · Computer vision · Success · Limitation · Cognitive science · Neuroscience
1 Introduction In the last few years Deep Nets have enabled enormous advances in computer vision and the study of biological visual systems. But as researchers in these areas, we have mixed feelings about them. On the one hand, we marvel at their successes and how they have led to amazing results on some real world tasks and, in academic settings, they almost always outperform alternative approaches on benchmark datasets. But, on the other hand, we are conscious of their current limitations, aware of papers (Darwiche 2018; Marcus 2018) which criticize them from the perspectives Communicated by Ivan Laptev. For those readers unfamiliar with Monty Python see: https://youtu.be/ Qc7HmhrgTuQ.
B
Chenxi Liu [email protected] Alan L. Yuille [email protected]
1
of machine reasoning and cognitive science respectively, and are concerned about the hype that sometimes surrounds them. The nature of our research means that we interact with research faculty in many disciplines (cognitive science, computer science, applied mathematics, neuroscience, engineering, physics, and radiology) and Deep Nets are a frequent topic of conversation. We find ourselves spending half th
Data Loading...
