Privacy-Preserving Deep Learning Techniques for Audio Data: Challenges and Advances

Dear Colleagues,

The exponential growth of deep learning applications in audio processing, ranging from speech recognition and speaker verification to emotion detection and biometric authentication, has unlocked unprecedented possibilities across industries. However, audio data inherently carry sensitive personal information, raising significant concerns about privacy, data security, and ethical AI deployment.

This Special Issue focuses on privacy-preserving deep learning techniques for audio data, inviting contributions that address the critical balance between leveraging advanced AI models and safeguarding user privacy. As audio becomes a key modality in AI-driven systems, ensuring confidentiality, preventing unauthorized data exploitation, and complying with regulatory frameworks (such as GDPR) are paramount.

We welcome innovative research and comprehensive reviews that explore methods such as federated learning, differential privacy, secure computation, adversarial defense mechanisms, and explainable AI, all within the context of audio data. Submissions may also address real-world applications, privacy risk assessments, and frameworks that enhance trustworthiness in AI-powered audio systems.

The aim is to gather state-of-the-art advancements and foster interdisciplinary dialogue, contributing to the development of secure, ethical, and high-performing audio AI solutions.

Call for paper:

We are pleased to invite you to contribute to this Special Issue on "Privacy-Preserving Deep Learning Techniques for Audio Data: Challenges and Advances".

In recent years, the rapid advancement of deep learning has significantly enhanced the analysis and understanding of audio data, enabling breakthroughs in speech recognition, speaker identification, emotion detection, and audio-based biometric systems. However, these advancements come with critical privacy concerns, as audio data often contain sensitive personal information, including identity, emotional state, health status, and other private attributes. The potential misuse, unauthorized access, or inadvertent leakage of such information poses serious ethical, legal, and security challenges.

Addressing privacy in audio-related AI systems is therefore a crucial and timely research area, demanding innovative solutions that balance performance with robust privacy protection.

Aim of the Special Issue and how the subject relates to the journal scope

This Special Issue aims to bring together cutting-edge research focused on privacy-aware deep learning methodologies for audio data. We seek contributions that explore novel frameworks, algorithms, and applications where privacy is preserved without compromising the effectiveness of AI models. Topics of interest include federated learning, differential privacy, adversarial training for privacy, and explainable AI in audio processing.

The scope aligns with the journal’s mission to advance AI by addressing contemporary challenges at the intersection of technology, ethics, and human-centric applications. The goal is to foster a comprehensive collection of high-quality works that push forward both theoretical and practical aspects of privacy-preserving audio AI systems.

Suggested themes and article types for submissions

In this Special Issue, original research articles and comprehensive reviews are welcome. Research areas may include (but are not limited to) the following:

• Privacy-preserving machine learning techniques applied to audio data;
• Federated learning for speech and speaker recognition;
• Differential privacy in audio signal processing;
• Adversarial methods for protecting sensitive information in audio datasets;
• Deepfake detection and anti-spoofing techniques in voice data;
• Explainable and interpretable AI for privacy in audio-based applications;
• Continual learning and privacy in evolving audio systems;
• Ethical, legal, and social implications of AI-driven audio analysis;
• Privacy risks and mitigation strategies in voice assistants and smart devices.

Dr. Carmelo Pino
Dr. Massimo Orazio Spata
Dr. Francesco Rundo
Dr. Angelo Alberto Messina
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. AI is an international peer-reviewed open access monthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 1600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• privacy
• audio processing
• deep learning
• federated learning
• differential privacy
• speech recognition
• speaker identification
• secure AI
• adversarial privacy
• explainable AI