Intriguing properties of neural networks
Why is this work in the frame?
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Abstract
Deep neural networks are highly expressive models that have recently achieved state of the art performance on speech and visual recognition tasks. While their expressiveness is the reason they succeed, it also causes them to learn uninterpretable solutions that could have counter-intuitive properties. In this paper we report two such properties. First, we find that there is no distinction between individual high level units and random linear combinations of high level units, according to various methods of unit analysis. It suggests that it is the space, rather than the individual units, that contains of the semantic information in the high layers of neural networks. Second, we find that deep neural networks learn input-output mappings that are fairly discontinuous to a significant extend. We can cause the network to misclassify an image by applying a certain imperceptible perturbation, which is found by maximizing the network's prediction error. In addition, the specific nature of these perturbations is not a random artifact of learning: the same perturbation can cause a different network, that was trained on a different subset of the dataset, to misclassify the same input.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
The record
- Venue
- arXiv (Cornell University)
- Topic
- Neural Networks and Applications
- Field
- Computer Science
- Canadian institutions
- Université de Montréal
- Funders
- —
- Keywords
- Computer scienceArtificial neural networkArtifact (error)Perturbation (astronomy)Artificial intelligenceDeep neural networksMachine learning
- Has abstract in OpenAlex
- yes