Filters








283 Hits in 8.4 sec

Adversarial Disentanglement of Speaker Representation for Attribute-Driven Privacy Preservation [article]

Paul-Gauthier Noé, Mohammad Mohammadamini, Driss Matrouf, Titouan Parcollet, Andreas Nautsch, Jean-François Bonastre
2021 arXiv   pre-print
In order to allow the user to choose which information to protect, we introduce in this paper the concept of attribute-driven privacy preservation in speaker voice representation.  ...  As a first solution to this concept, we propose to use an adversarial autoencoding method that disentangles in the voice representation a given speaker attribute thus allowing its concealment.  ...  Section 2 explains the attribute-driven privacy preservation concept; Section 3 presents the adversarial disentangling autoencoder 2 for hiding a binary attribute in x-vectors; Section 4 presents the experimental  ... 
arXiv:2012.04454v3 fatcat:axtcjx7qxngxlhuculnbt43gi4

An Attribute-Aligned Strategy for Learning Speech Representation [article]

Yu-Lin Huang, Bo-Hao Su, Y.-W. Peter Hong, Chi-Chun Lee
2021 arXiv   pre-print
speech emotion recognition (SER), and an emotionless representation for speaker verification (SV).  ...  Also, our proposed learning strategy reduces the model and training process needed to achieve multiple privacy-preserving tasks.  ...  Notice that PP stands for privacy-preserving, where columns of origin stand for original representation without privacy protection, PP-SER stands for identity-free SER, and PP-SV stands for emotionless  ... 
arXiv:2106.02810v1 fatcat:vrzlgjle2fe7lpfpddacm4xsfe

A bridge between features and evidence for binary attribute-driven perfect privacy [article]

Paul-Gauthier Noé and Andreas Nautsch and Driss Matrouf and Pierre-Michel Bousquet and Jean-François Bonastre
2022 arXiv   pre-print
We show the applicability of the approach on an attribute-driven privacy task where the sex information is removed from speaker embeddings.  ...  Results on VoxCeleb2 dataset show the efficiency of the method that outperforms in terms of privacy and utility our previous experiments based on adversarial disentanglement.  ...  INTRODUCTION Attribute-driven privacy aims to make the data independent of an attribute a user wants to keep secret [1, 2] .  ... 
arXiv:2110.05840v2 fatcat:22g5r7ajgnaprhlqnrnyiuptly

A Tandem Framework Balancing Privacy and Security for Voice User Interfaces [article]

Ranya Aloufi, Hamed Haddadi, David Boyle
2021 arXiv   pre-print
Adversaries may use advanced transformation tools to trigger a spoofing attack using fraudulent biometrics for a legitimate speaker.  ...  These techniques transform one or more elements of a speech signal, e.g., identity and emotion, while preserving linguistic information.  ...  Likewise, recent applications have suggested the implementation of disentanglement [7] in learning speech representations can enhance the robustness of speech representations and overcome common speaker  ... 
arXiv:2107.10045v1 fatcat:vgw7mmseurb3bfr6s2xg4c3enq

Disentanglement for audio-visual emotion recognition using multitask setup [article]

Raghuveer Peri, Srinivas Parthasarathy, Charles Bradshaw, Shiva Sundaram
2021 arXiv   pre-print
This work explores the disentanglement of multimodal signal representations for the primary task of emotion recognition and a secondary person identification task.  ...  We evaluate three different techniques for disentanglement and report results of up to 13% disentanglement while maintaining emotion recognition performance.  ...  Jaiswal and Provost [20] explored privacy-preserving multimodal emotion representations, where audio and text modalities were utilized.  ... 
arXiv:2102.06269v1 fatcat:r6jkribihrcglm4jzmavkostxe

Deep Representation Learning in Speech Processing: Challenges, Recent Advances, and Future Trends [article]

Siddique Latif, Rajib Rana, Sara Khalifa, Raja Jurdak, Junaid Qadir, Björn W. Schuller
2021 arXiv   pre-print
Recent reviews in speech have been conducted for ASR, SR, and SER, however, none of these has focused on the representation learning from speech -- a gap that our survey aims to bridge.  ...  This has motivated the adoption of a recent trend in speech community towards utilisation of representation learning techniques, which can learn an intermediate representation of the input signal automatically  ...  Privacy preserved representation learning is a relatively unexplored research topic.  ... 
arXiv:2001.00378v2 fatcat:ysvljxylwnajrbowd3kfc7l6ve

To train or not to train adversarially: A study of bias mitigation strategies for speaker recognition [article]

Raghuveer Peri, Krishna Somandepalli, Shrikanth Narayanan
2022 arXiv   pre-print
Moreover, there are only a handful of bias mitigation strategies developed for speaker recognition systems.  ...  Speaker recognition is increasingly used in several everyday applications including smart speakers, customer care centers and other speech-driven analytics.  ...  When the false rejects are more important, our methods preserve the fairness of existing speaker representations.  ... 
arXiv:2203.09122v1 fatcat:fecp74scajfzpevqukjk45ibda

Deep Learning for Text Style Transfer: A Survey [article]

Di Jin, Zhijing Jin, Zhiting Hu, Olga Vechtomova, Rada Mihalcea
2021 arXiv   pre-print
Text style transfer is an important task in natural language generation, which aims to control certain attributes in the generated text, such as politeness, emotion, humor, and many others.  ...  We also provide discussions on a variety of important topics regarding the future development of this task. Our curated paper list is at https://github.com/zhijing-jin/Text_Style_Transfer_Survey  ...  The third approach, Latent Representation Splitting (LRS), as illustrated in Figure 2c, first disentangles the input text into two parts: the latent attribute representation a, and semantic representation  ... 
arXiv:2011.00416v5 fatcat:wfw3jfh2mjfupbzrmnztsqy4ny

The VoicePrivacy 2020 Challenge: Results and findings [article]

Natalia Tomashenko, Xin Wang, Emmanuel Vincent, Jose Patino, Brij Mohan Lal Srivastava, Paul-Gauthier Noé, Andreas Nautsch, Nicholas Evans, Junichi Yamagishi, Benjamin O'Brien, Anaïs Chanclu, Jean-François Bonastre (+2 others)
2021 arXiv   pre-print
In addition, we present experimental results for alternative privacy metrics and attack models developed as a part of the post-evaluation analysis.  ...  Finally, we summarize our insights and observations that will influence the design of the next VoicePrivacy challenge edition and some directions for future voice anonymization research.  ...  Acknowledgment VoicePrivacy was born at the crossroads of projects VoicePersonae, COM-PRISE (https://www.compriseh2020.eu/), and DEEP-PRIVACY.  ... 
arXiv:2109.00648v3 fatcat:oyu4fa32xjfnvhno7h5mlcxr3i

2021 Index IEEE Transactions on Multimedia Vol. 23

2021 IEEE transactions on multimedia  
The Author Index contains the primary entry for each item, listed under the first author's name.  ...  The primary entry includes the coauthors' names, the title of the paper or other item, and its location, specified by the publication abbreviation, year, month, and inclusive pagination.  ...  ., +, TMM 2021 4541-4554 Data privacy Privacy-Preserving In-Home Fall Detection Using Visual Shielding Sensing and Private Information-Embedding.  ... 
doi:10.1109/tmm.2022.3141947 fatcat:lil2nf3vd5ehbfgtslulu7y3lq

Deep Learning for Text Style Transfer: A Survey

Di Jin, Zhijing Jin, Zhiting Hu, Olga Vechtomova, Rada Mihalcea
2021 Computational Linguistics  
Text style transfer is an important task in natural language generation, which aims to control certain attributes in the generated text, such as politeness, emotion, humor, and many others.  ...  We also provide discussions on a variety of important topics regarding the future development of this task.  ...  Adversarial decomposition of text Manning. 2017. Get to the point: representation.  ... 
doi:10.1162/coli_a_00426 fatcat:v7vmb62ckfcu5k5mpu2pydnrxy

2020 Index IEEE Transactions on Multimedia Vol. 22

2020 IEEE transactions on multimedia  
., +, TMM June 2020 1385-1394 Adversarial Attribute-Text Embedding for Person Search With Natural Lan- guage Query.  ...  ., +, TMM Nov. 2020 2858-2872 Data protection Exploiting Vulnerabilities of Deep Neural Networks for Privacy Protection.  ...  Image watermarking Blind Watermarking for 3-D Printed Objects by Locally Modifying Layer Thickness. 2780 -2791 Low-Light Image Enhancement With Semi-Decoupled Decomposition.  ... 
doi:10.1109/tmm.2020.3047236 fatcat:llha6qbaandfvkhrzpe5gek6mq

ICASSP 2020 Table of Contents

2020 ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
TOPOLOGY OPTIMIZATION FOR IMAGE DENOISING Wengtai Su, National Tsing Hua University, Taiwan; Gene Cheung, Richard P.  ...  .......................................... 3057 LEARNING DISENTANGLED REPRESENTATION OF IMPULSIVE NOISE ON STEERING GEAR Seok-Jun Bu, Namu Park, Yonsei University, Korea (South); Gue-Hwan Nam, Jae-Yong  ...  Stokes, Microsoft Corporation, United States IFS-L2.2: PRIVACY-PRESERVING IMAGE SHARING VIA SPARSIFYING LAYERS Onur Günlü, Rafael F.  ... 
doi:10.1109/icassp40776.2020.9054406 fatcat:6h7hh2hxhne4pbmphharu2et2m

From Theories on Styles to their Transfer in Text: Bridging the Gap with a Hierarchical Survey [article]

Enrica Troiano and Aswathy Velutharambath and Roman Klinger
2021 arXiv   pre-print
They can, for instance, rephrase a formal letter in an informal way, convey a literal message with the use of figures of speech, edit a novel mimicking the style of some well-known authors.  ...  We organize them into a hierarchy, highlighting the challenges for the definition of each of them, and pointing out gaps in the current research landscape. The hierarchy comprises two main groups.  ...  Adversarial Learning. Implicit disentanglement has been instantiated by adversarial learning in several ways.  ... 
arXiv:2110.15871v2 fatcat:ddpowdm6pbazzd5mwl65nrge5q

The Association for the Advancement of Artificial Intelligence 2020 Workshop Program

Grace Bang, Guy Barash, Ryan Bea, Jacques Cali, Mauricio Castillo-Effen, Xin Chen, Niyati Chhaya, Rachel Cummings, Rohan Dhoopar, Sebastijan Dumanci, Huáscar Espinoza, Eitan Farchi (+29 others)
2020 The AI Magazine  
The Association for the Advancement of Artificial Intelligence 2020 Workshop Program included twenty-three workshops covering a wide range of topics in artificial intelligence.  ...  This report contains the required reports, which were submitted by most, but not all, of the workshop chairs.  ...  Privacy-Preserving Artificial Intelligence (W18) The goal of the AI Privacy-Preserving Artificial Intelligence workshop was to provide a platform for researchers to discuss problems and present solutions  ... 
doi:10.1609/aimag.v41i4.7398 fatcat:r6bw77vy4zgmrbgyuvsjs5knta
« Previous Showing results 1 — 15 out of 283 results