Research map / Learning theory and algorithms
Matrix and tensor methods: research map
1,467 papers on Matrix and tensor methods in Learning theory and algorithms, from ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR (2016–2026), grouped into 3 clusters and 13 approaches. The busiest year so far is 2017.
What it is
Methods for data organised as matrices or multi-dimensional arrays (tensors): completing missing entries (as in recommendation), finding low-rank structure, sparse recovery and robust PCA, with guarantees on when they work.
Terms you will see on this page:
- Low-rank
- a matrix that can be described by a few underlying factors.
- Matrix completion
- filling in missing entries of a low-rank matrix.
- Tensor decomposition
- factorising a multi-way array into simple components.
Start reading with the most cited papers here:
- Tensor Robust Principal Component Analysis: Exact Recovery of Corrupted Low-Rank Tensors via Convex Optimization (CVPR 2016 · 556 citations)
- Polynomial Codes: an Optimal Design for High-Dimensional Coded Matrix Multiplication (NeurIPS 2017 · 382 citations)
Within Learning theory and algorithms, its share held steady from 7.8% in 2023–24 to 7.7% in 2025–26 (173 → 186 papers at ICML, NeurIPS, ICLR, CVPR and AAAI, the venues with data for all four years).
Explore Matrix and tensor methods in the interactive map
Working on something in this topic? Describe your idea in scime atlas to see which approach it falls under, the closest papers by meaning and how crowded the spot has become.
What changed recently
Approaches whose share of the topic moved most between 2023–24 and 2025–26 (ICML, NeurIPS, ICLR, CVPR and AAAI):
- Geometry of low-rank and SPD structure rising: 6% → 15% of the topic (11 → 27 papers)
- Structured low-rank learning rising: 7% → 11% of the topic (12 → 21 papers)
- Tensor network decompositions rising: 2% → 6% of the topic (4 → 11 papers)
- Matrix completion and factorization falling: 12% → 8% of the topic (21 → 14 papers)
- Sparse regression with guarantees falling: 8% → 4% of the topic (14 → 7 papers)
Papers from 2025–2026 that are already cited most:
- Learning to Normalize on the SPD Manifold under Bures-Wasserstein Geometry (CVPR 2025 · 6 citations)
- OTLRM: Orthogonal Learning-based Low-Rank Metric for Multi-Dimensional Inverse Problems (AAAI 2025 · 3 citations)
- Neural Networks Perform Sufficient Dimension Reduction (AAAI 2025 · 3 citations)
- Non-Convex Tensor Recovery from Local Measurements (AAAI 2025 · 2 citations)
Approaches and key papers
Low-rank matrices · 700 papers
Keywords: matrices, matrix completion, rank matrix
Approaches in this cluster:
- Low-rank approximation theory (203 papers)
Analyzes error and perturbation bounds for SVD and low-rank estimation of matrices. - Matrix completion and factorization (162 papers)
Recovers low-rank matrices from partial entries via alternating minimization and sparse regularization. - Randomized sketching for matrices (136 papers)
Uses random projections and frequent-directions sketches for fast approximate linear algebra. - Structured low-rank learning (113 papers)
Applies low-rank structure to ranking, fairness and probability matrix problems with guarantees. - Geometry of low-rank and SPD structure (86 papers)
Studies rank, random matrix geometry and positive-definite manifolds in network training and learning.
Most cited and most cited since 2024:
- Polynomial Codes: an Optimal Design for High-Dimensional Coded Matrix Multiplication (NeurIPS 2017 · 382 citations)
- Mutual information for symmetric rank-one matrix estimation: A proof of the replica formula (NeurIPS 2016 · 125 citations)
- HyperAttention: Long-context Attention in Near-Linear Time (ICLR 2024 · 9 citations)
- Learning to Normalize on the SPD Manifold under Bures-Wasserstein Geometry (CVPR 2025 · 6 citations)
Sparse recovery and PCA · 595 papers
Keywords: sparse, recovery, pca
Approaches in this cluster:
- Sparse regression with guarantees (151 papers)
Recovers sparse signals through hard thresholding, group sparsity and robust estimation with statistical guarantees. - High-dimensional estimation theory (129 papers)
Analyzes spectral and message-passing estimators in high-dimensional regression and multi-index models. - Signal recovery and phase retrieval (90 papers)
Recovers signals from measurements through convex programming and generative or structured priors. - Sparse and robust PCA (89 papers)
Designs provable algorithms for sparse PCA and subspace recovery under noise, missing data and outliers. - Nonconvex low-rank and sparse optimization (136 papers)
Gives global convergence and landscape results for matrix factorization and primal-dual coordinate methods.
Most cited and most cited since 2024:
- Theoretical Linear Convergence of Unfolded ISTA and Its Practical Weights and Thresholds (NeurIPS 2018 · 151 citations)
- PD-Sparse : A Primal and Dual Sparse Approach to Extreme Multiclass and Multilabel Classification (ICML 2016 · 130 citations)
- Secure Distributed Sparse Gaussian Process Models Using Multi-Key Homomorphic Encryption (AAAI 2024 · 4 citations)
- Precise asymptotics of reweighted least-squares algorithms for linear diagonal networks (NeurIPS 2024 · 3 citations)
Tensor decompositions · 172 papers
Keywords: tensors, decomposition, rank tensor
Approaches in this cluster:
- Tensor network decompositions (54 papers)
Represents high-order tensors with tensor networks and tensor trains, with theory on their learning capacity. - Randomized tensor decomposition algorithms (48 papers)
Speeds up CP, Tucker and ring decompositions through sketching and sampling with guarantees. - Low-rank tensor completion (70 papers)
Recovers missing tensor entries with nuclear norm, nonconvex and approximate iterative methods.
Most cited and most cited since 2024:
- Tensor Robust Principal Component Analysis: Exact Recovery of Corrupted Low-Rank Tensors via Convex Optimization (CVPR 2016 · 556 citations)
- Supervised Learning with Tensor Networks (NeurIPS 2016 · 196 citations)
- Towards Multi-Mode Outlier Robust Tensor Ring Decomposition (AAAI 2024 · 9 citations)
- HDQMF: Holographic Feature Decomposition Using Quantum Algorithms (CVPR 2024 · 4 citations)
Related topics in Learning theory and algorithms
- Sample complexity (1,553)
- Combinatorial and search optimization (1,576)
- Fairness (550)
- Submodular and game algorithms (1,064)
- Prediction and decision losses (2,400)
- Kernels and regression theory (1,272)
- Clustering (998)
- Optimal transport (462)
