SOPHON: Non-Fine-Tunable Learning to Restrain Task Transferability For Pre-trained Models

1 code implementation • 19 Apr 2024 • Jiangyi Deng, Shengyuan Pang, Yanjiao Chen, Liangming Xia, Yijie Bai, Haiqin Weng, Wenyuan Xu

In addition, we carefully design the optimization process to entrap the pre-trained model within a hard-to-escape local optimum regarding restricted domains.

Meta-Learning

SafeGen: Mitigating Unsafe Content Generation in Text-to-Image Models

1 code implementation • 10 Apr 2024 • Xinfeng Li, Yuchen Yang, Jiangyi Deng, Chen Yan, Yanjiao Chen, Xiaoyu Ji, Wenyuan Xu

The key idea is to eliminate unsafe visual representations from the model regardless of the text input.

Catch You and I Can: Revealing Source Voiceprint Against Voice Conversion

no code implementations • 24 Feb 2023 • Jiangyi Deng, Yanjiao Chen, Yinan Zhong, Qianhao Miao, Xueluan Gong, Wenyuan Xu

Voice conversion (VC) techniques can be abused by malicious parties to transform their audios to sound like a target speaker, making it hard for a human being or a speaker verification/identification system to trace the source speaker.

Representation Learning Speaker Verification +1

V-Cloak: Intelligibility-, Naturalness- & Timbre-Preserving Real-Time Voice Anonymization

no code implementations • 27 Oct 2022 • Jiangyi Deng, Fei Teng, Yanjiao Chen, Xiaofu Chen, Zhaohui Wang, Wenyuan Xu

Apart from the anonymity loss, we further incorporate the intelligibility loss and the psychoacoustics-based naturalness loss.

Automatic Speech Recognition Automatic Speech Recognition (ASR) +3