Model-Based Deep Learning: Integrating Signal Processing and Machine Learning

Dear Colleagues,

This Special Issue (SI) aims to consolidate and advance the frontier of research at the intersection of model-based signal processing and data-driven deep learning. Traditional signal processing relies on mathematically rigorous models derived from physics, statistics, and domain knowledge, offering interpretability and efficiency but often struggling with complex, real-world dynamics. In contrast, purely data-driven deep learning excels at learning intricate patterns from large datasets but acts as a black box requiring massive data and computational resources while lacking interpretability.

This SI will focus on Model-Based Deep Learning (MBDL) as a unifying paradigm that hybridizes these two approaches. We seek contributions that leverage partial domain knowledge to structure, guide, and enhance the robustness of deep learning systems, while simultaneously using data-driven methods to complete or accelerate traditional models. The goal is to foster the development of systems that are more data-efficient, interpretable, reliable, and high-performing than purely data-driven or purely model-based approaches alone. The central emphasis of this SI will be on novel methodologies and applications in which deep learning architectures and training paradigms are explicitly designed around, or integrated with, established signal processing models. This SI will cover a wide range of topics at the nexus of signal processing and machine learning.

This Special Issue will not operate in a vacuum but will strategically supplement and extend the current body of literature in the following crucial ways:

Moving Beyond Purely Data-Driven Paradigms: While there are numerous publications and SIs on AI for signal processing or deep learning and its applications, they often emphasize replacing traditional models with black-box DNNs. This SI will specifically highlight the synergy between models and data, addressing the limitations of purely data-driven methods (data hunger, lack of interpretability) that are becoming increasingly apparent. It positions MBDL not as a mere alternative, but as a necessary evolution.

Catalyzing Cross-Disciplinary Collaboration: By framing the topic around the integration of two fundamental paradigms—model-based and data-driven—this issue will attract contributions from diverse fields (e.g., optimization, information theory, statistical signal processing, and various engineering domains), fostering a cross-pollination of ideas that is essential for tackling complex signal processing challenges.

Dr. Haijian Zhang
Dr. Xing Tang
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. Electronics is an international peer-reviewed open access semimonthly 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 2400 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.