Towards generalizing deep-audio fake detection networks

arXiv:2305.13033v3 Announce Type: replace-cross
Abstract: Today's generative neural networks allow the creation of high-quality synthetic speech at scale. While we welcome the creative use of this new technology, we must also recognize the risks. As synthetic speech is abused for monetary and identity theft, we require a broad set of deepfake identification tools. Furthermore, previous work reported a limited ability of deep classifiers to generalize to unseen audio generators. We study the frequency domain fingerprints of current audio generators. Building …

abstract arxiv audio creative cs.lg cs.sd deepfake detection eess.as fake generative identification identity identity theft networks neural networks new technology quality risks scale set speech synthetic synthetic speech technology theft tools type