Applied and Computational Harmonic Analysis 2025

11 papers
In 2025, Applied and Computational Harmonic Analysis featured 11 papers centered on the intersection of mathematical theory and computational applications. A prominent research theme involved complex-valued spectral representations for directed graphs. Significant focus was placed on the mathematical design of deep learning systems, particularly regarding algorithmic transparency and societal constraints. Investigations into neural network architectures included the identification of wide shallow networks, controlled learning of pointwise nonlinearities, and unified stochastic frameworks for quantization and pruning. Additional research addressed signal processing and estimation challenges, such as robust phase retrieval under adversarial sparse outliers and the development of parameter-free two-bit covariance estimators with improved error rates. The year’s contributions also explored graph theory through the sparsification of the regularized magnetic Laplacian using multi-type spanning forests. Prolific authors included Junren Chen, Michael Ng, Gitta Kutyniok, and Holger Boche. These works collectively advanced optimization techniques, including randomized Kaczmarz methods with tail averaging, and refined the theoretical understanding of high-dimensional data structures.

Top Research Directions

The 12 research directions with Applied and Computational Harmonic Analysis 2025 papers, ranked by paper count, next to 2024. Share is the fraction of the edition’s papers in the direction.

#Direction20252024Share 2025Share 20242024 rank
1Adaptive Neighborhood Selection for Non-Uniform Data109.1%0.0%—
2Algorithmic Stability Analysis for Generalization Bounds109.1%0.0%—
3Beyond the Rank One Assumption in Spiked Models109.1%0.0%—
4Dual Certificates for Exact High-Dimensional Recovery109.1%0.0%—
5Hermitian Representations for Directed Graph Neural Networks109.1%0.0%—
6Laplacian Spectral Property Preservation in Graph Sparsification109.1%0.0%—
7Phase Retrieval from Intensity Measurements109.1%0.0%—
8Preconditioning Strategies for Ill-Conditioned Linear Systems109.1%0.0%—
9Quantifying Computational Complexity in Algorithmic Design109.1%0.0%—
10Scaling Generalization Bounds Gracefully with Network Depth109.1%0.0%—
11The Mechanics of Benign Overfitting in High Dimensions109.1%0.0%—
12Why Large-Batch Training Degrades Generalization109.1%0.0%—

Download all directions as CSV

Institutions

Institutions credited by Applied and Computational Harmonic Analysis 2025 papers: a paper credits every institution named in its affiliations, once. 25 institutions share 10 credited papers.

#InstitutionPapersShare20242024 rank
1University of California San Diego 220.0%21
2University of Hong Kong 220.0%0—
302y22ws83 110.0%0—
4Athena Research and Innovation Center In Information Communication & Knowledge Technologies 110.0%0—
5California Institute of Technology 110.0%137
6Centre de Recherche en Informatique, Signal et Automatique de Lille 110.0%0—
7Centre National de la Recherche Scientifique 110.0%0—
8George Mason University 110.0%0—
9Hong Kong Baptist University 110.0%0—
10King Abdullah University of Science and Technology 110.0%0—
11Ludwig-Maximilians-Universität München 110.0%135
12Munich Center for Machine Learning 110.0%121
13Munich Center for Quantum Science and Technology 110.0%134
14National and Kapodistrian University of Athens 110.0%0—
15National University of Defense Technology 110.0%0—
16Ruhr University Bochum 110.0%133
17Shanghai Jiao Tong University 110.0%0—
18Technical University of Munich 110.0%120
19University of California, Berkeley 110.0%114
20University of Cambridge 110.0%0—
21University of Oxford 110.0%0—
22Université de Lille 110.0%0—
23Zhejiang University 110.0%0—
24École Centrale de Lille 110.0%0—
25École Polytechnique Fédérale de Lausanne 110.0%123

Download all institutions as CSV

About this data

Papers. Applied and Computational Harmonic Analysis 2025 lists 11 papers on Lacuna, duplicate versions counted once; rejected and withdrawn submissions are left out.

Directions. Directions are Lacuna’s clustering of the concept elements extracted from each paper, so a paper can belong to several and direction counts add up to more than the paper count. A direction’s share is its papers divided by the edition’s papers. Comparisons with 2024 use shares, since the edition went from 19 to 11 papers.

Institutions. A paper credits each institution named in its affiliations, resolved to a ROR record, once. 10 of 11 papers (90.9%) credit at least one institution and are the base for institution shares; the rest have no resolved affiliation yet.

Data. Counts come from Lacuna’s data build of 2026-11-19 and change as papers are added. Tables as CSV: directions · institutions. Cite as: Lacuna, “Applied and Computational Harmonic Analysis 2025”, https://lacuna.tiptreesystems.com/venue/applied-and-computational-harmonic-analysis/2025.

Top Authors

AuthorPapers
Adalbert Fono1
Ethan N. Epperly1
Gitta Kutyniok1
Haoyu Zhang1
Holger Boche1
Junren Chen1
Michael Ng1
Michael Unser1
Rémi Bardenet1
Vicky Kouni1

Papers

Title
A parameter-free two-bit covariance estimator with improved operator norm error rate
Controlled learning of pointwise nonlinearities in neural-network-like architectures
Efficient identification of wide shallow neural networks with biases
Framelet message passing
Generalization analysis of an unfolding network for analysis-based compressed sensing
Mathematical algorithm design for deep learning under societal and judicial constraints: The algorithmic transparency requirement
Minibatch and local SGD: Algorithmic stability and linear speedup in generalization
Randomized Kaczmarz with tail averaging
Robust outlier bound condition to phase retrieval with adversarial sparse outliers
Sparsification of the regularized magnetic Laplacian with multi-type spanning forests
Unified stochastic framework for neural network quantization and pruning