Study on tensor nuclear norm's decomposability and subdifferential.
problem Understanding tensor nuclear norm in higher-order tensors.
method Showed decomposability over specific subspaces, derived subdifferential inclusions, and studied subgradients.
result Established the statistical performance of tensor robust principal component analysis.
Unified approach for error bounds in group sparse compressed sensing.
problem Error bounds for group sparse vectors in compressed sensing.
method Unified approach using decomposable and γ-decomposable norms.
result Bounds for various group sparse norms derived.
Tensor rank and low-rank tensor decompositions have many applications in learning and complexity theory. Most known algorithms use unfoldings of tensors and can only handle rank up to n⌊p/2⌋ for a p-th order tensor in Rnp. Previously no efficient algorithm can decompose 3rd order ten…
DoRA improves adaptation efficiency for large models by factoring norms and fusing kernels.
problem High-rank DoRA is computationally expensive and infeasible on common GPUs.
method Factored norms and fused Triton kernels to reduce memory and speed up computation.
result Fused implementation is up to 2.0x faster for inference and 1.9x faster for gradient computation.
Algorithm for factorizing financial network data into groups.
problem Modeling groups within heterogeneous financial networks.
method Non-negative factorization of an occurrence tensor, l0 norm for sparsity, efficient splitting method.
result Effective factorization of financial documents into embedded groups.
For supervised and unsupervised learning, positive definite kernels allow to use large and potentially infinite dimensional feature spaces with a computational cost that only depends on the number of observations. This is usually done through the penalization of predictor functions by Euclidean or Hilbertian norms. In …
The paper examines torsions in Minkowskian product of Finsler metrics.
problem Investigating Cartan torsion and mean Cartan torsion in Minkowskian product of Finsler metrics.
method Deriving explicit formulas for Cartan torsion and mean Cartan torsion, analyzing their geometric behavior, and providing conditions for bounded norm.
result Both Cartan torsion and mean Cartan torsion decompose additively if and only if the Minkowskian product is Euclidean, and a necessary and sufficient condition for the norm of the mean Cartan torsion to remain bounded in the Euclidean case is provided.
In a recent paper, it is shown that the LASSO algorithm exhibits "near-ideal behavior," in the following sense: Suppose y=Az+η where A satisfies the restricted isometry property (RIP) with a sufficiently small constant, and ∥η∥2≤ε. Then minimizing ∥z∥1 subject to $\Vert y - Az \Ver…
New method solves rank-1 L1-norm TUCKER2 decomposition efficiently.
problem Exact solution to rank-1 L1-norm TUCKER2 decomposition of tensors.
method Proved equivalent to combinatorial optimization, derived two algorithms.
result L1-TUCKER2 outperforms other methods in tensor approximation for outlier-corrupted data.
We define a torsion invariant T for every balanced sutured manifold (M,g), and show that it agrees with the Euler characteristic of sutured Floer homology SFH. The invariant T is easily computed using Fox calculus. With the help of T, we prove that if (M,g) is complementary to a Seifert surface of an alternating knot, …
Suppose a given observation matrix can be decomposed as the sum of a low-rank matrix and a sparse matrix (outliers), and the goal is to recover these individual components from the observed sum. Such additive decompositions have applications in a variety of numerical problems including system identification, latent var…
New optimization problem for graph and hypergraph learning tasks.
problem Learning tasks on graphs and hypergraphs.
method Quadratic decomposable submodular function minimization (QDSFM) via dual strategy and double-loop algorithms.
result Linear convergence rates for outer-loop optimization and effective hypergraph-based PageRank algorithm.
The paper characterizes functions of shallow ReLU NN denoisers under minimal norm constraints.
problem Understanding the theoretical success of neural network denoisers.
method Characterization of functions realized by shallow ReLU NN denoisers under minimal norm constraints.
result The functions realized by shallow ReLU NN denoisers are contractive toward clean data points and generalize better than the empirical MMSE estimator at low noise levels.
Efficient surrogate losses and regularization methods for structured prediction.
problem Efficiency and performance in structured prediction with rich label structures.
method Development of bi-criteria surrogate losses and shared Frobenius norm for regularization.
result Improved efficiency and performance in inference and optimization for structured prediction.
New method infers graph from dependent matrix data.
problem Inferring graph from dependent matrix data.
method Sparse-group lasso-based frequency-domain formulation with ADMM approach.
result Local convergence of inverse PSD estimators to true value.
New method shows stochastic momentum can converge quickly on optimization problems.
problem Improving convergence of stochastic optimization methods.
method Stochastic heavy ball momentum with minibatching.
result Stochastic heavy ball momentum retains fast linear rate on quadratic problems.
Proposes an algorithm for infinite-dimensional sparse learning in system identification.
problem System identification without known model structures.
method Atomic norm regularization and greedy algorithm for solving an infinite-dimensional group lasso problem.
result The proposed algorithm outperforms benchmark methods in impulse response fitting and pole location estimation.
Unified framework for coupled tensor completion improves recovery accuracy.
problem Improving recovery accuracy in coupled tensor completion.
method Unified framework using tensor ring (TR) decomposition with shared latent factors and novel optimization model.
result The proposed method achieves superior recovery accuracy on real-world data compared to state-of-the-art methods.
Currents represent generalized surfaces studied in geometric measure theory. They range from relatively tame integral currents representing oriented compact manifolds with boundary and integer multiplicities, to arbitrary elements of the dual space of differential forms. The flat norm provides a natural distance in the…
Modern geometric measure theory, developed largely to solve the Plateau problem, has generated a great deal of technical machinery which is unfortunately regarded as inaccessible by outsiders. Some of its tools (e.g., flat norm distance and decomposition in generalized surface space) hold interest from a theoretical pe…
Paper proposes ADMM algorithms for non-smooth optimization under RDP.
problem Optimizing composite functions with non-smooth penalties under privacy constraints.
method Developed ssADMM and mpADMM algorithms for non-smooth optimization problems with RDP guarantees.
result Both ssADMM and mpADMM outperform baseline methods in high privacy settings.
New framework for optimizing machine learning risks.
problem Optimizing non-decomposable machine learning objectives.
method Empirical X-risk minimization (EXM) framework with algorithmic techniques.
result Developed algorithms for solving EXM with smooth non-convex objectives.
Automated robust solvers for arbitrary operators using game theory and Gaussian fields.
problem Developing scalable numerical solvers for any bounded linear operator.
method Formulating the problem as a game theory problem and using Gaussian fields to find optimal strategies.
result Introducing the Fast Gamblet Transform (FGT) for efficient linear system solving and eigenspace analysis.
Regular sliceness implies once-stably decomposable sliceness in symplectizations.
problem Relationship between regular and decomposable Lagrangian cobordisms in symplectizations.
method Stabilization-free strategy and satellite operations.
result Regular sliceness implies once-stably decomposable sliceness.
Histogram transform ensembles improve density estimation accuracy.
problem Improving density estimation accuracy for various distributions.
method Histogram transform ensembles (HTE) with theoretical analysis and experimental validation.
result HTE outperforms single histogram transforms and offers almost optimal convergence rates in Hölder space C0,α. New tensor formulation reveals gradient flow's bias in linear neural networks.
problem Understanding implicit bias in linear neural network training.
method Tensor formulation of neural networks, including fully-connected, diagonal, and convolutional networks.
result Gradient flow on linear tensor networks converges to solutions of specific optimization problems.
New method improves density estimation accuracy and resistance to dimensionality issues.
problem Improving density estimation accuracy and resistance to dimensionality issues.
method Best-scored random forest density estimation method.
result The method achieves better estimation results than other state-of-the-art approaches.
Study shows rigidity in cusp-decomposable manifolds' geometry.
problem Understanding the geometry of cusp-decomposable manifolds.
method Examined large scale geometry and quasi-isometries.
result Proved quasi-isometric rigidity for fundamental groups.
This paper investigates Shampoo's heuristics and decouples preconditioner updates.
problem Improving Shampoo's heuristics for training neural networks.
method Decomposing preconditioner updates, correcting eigenvalues, and adapting eigenbasis computation frequency.
result Principled techniques to remove Shampoo's heuristics and improve training algorithms.
This paper proposes an online MTL framework that improves scalability and robustness.
problem Efficient online multi-task learning with correlated and personalized task structures.
method Decomposes weight matrix into low-rank common structure and personalized patterns using nuclear norm and group lasso.
result Achieves sub-linear regret and improved performance with log-determinant function.
The paper defines new types of positivity and proves properties of Schur forms for vector bundles.
problem Defining and characterizing new types of positivity for vector bundles.
method Introducing and characterizing two types of strongly decomposable positivity, proving properties of Schur forms.
result Schur forms of strongly decomposable positive vector bundles are positive or weakly positive, answering a question of Griffiths.
Researchers establish bounds and continuity of decomposed Möbius energies using cosine formula.
problem Estimating the bounds and continuity of decomposed Möbius energies.
method Using the cosine formula to evaluate upper and lower bounds and modulus of continuity of decomposed energies.
result Affirmative answer to the question of estimating decomposed energies using the cosine formula.
Chia and Nakano (2009) introduced the concept of M-decomposability of probability densities in one-dimension. In this paper, we generalize M-decomposability to any dimension. We prove that all elliptical unimodal densities are M-undecomposable. We also derive an inequality to show that it is better to represent an M-de…
The paper tackles tensor factorization and completion from noisy data.
problem Sparse nonnegative tensor factorization and completion from partial and noisy observations.
method Minimizes the sum of maximum likelihood estimation and tensor ℓ0 norm with nonnegativity constraints. result Error bounds and minimax lower bounds are established for the proposed model.
Characterizes decomposable torus fiber bundles in low dimensions.
problem Decomposability of torus fiber bundles over a circle.
method Characterization through properties of matrices.
result Complete characterization of decomposable bundles with fiber dimension less than 4.
New method for faster graph parameter inference from large random Kronecker graphs.
problem Efficiently infer graph parameters from large random Kronecker graphs.
method Decompose adjacency matrix into signal and noise components, then use denoising and solving approach.
result Proposed method achieves comparable or better performance than existing methods at lower computational cost.
Study shows Khovanov homology's relation to decomposable Lagrangian cobordisms.
problem Understanding the relationship between Khovanov homology and decomposable Lagrangian cobordisms.
method Utilized previously defined filtered invariants to give obstructions.
result Partial answer to Ekholm, Honda, and Kálmán's question about Khovanov homology and decomposable Lagrangian cobordisms.
Paper proposes a new method to minimize submodular functions with fewer calls to simpler oracles.
problem Minimizing the sum of submodular set functions with limited information.
method Introduces a modified convex problem requiring constrained total variation oracles that can be solved with fewer calls to minimization oracles.
result Shows significant reduction in the number of calls to minimization oracles.
The study shows how to embed cusp-decomposable manifolds quasi-isometrically.
problem Embedding cusp-decomposable manifolds quasi-isometrically.
method Using properties of the electric space of the universal cover, we show quasi-isometric embeddings.
result Isomorphisms between fundamental groups of higher graph manifolds preserve the decomposition into pieces.
New manifold types defined on quotient spaces.
problem Characterizing specific manifold structures.
method Introducing decomposable LCP manifolds and studying their properties.
result Characterized LCP manifolds on quotient spaces.
A Seifert surface F for a knot K is disk decomposable if there is a taut sutured manifold heirarchy for the complement of F, whose decomposing surfaces are all disks. It follows that F has minimal genus for the knot K, and has handlebody complement, i.e., F is free. We show that these necessary conditions for disk deco…
Differentially private algorithms for submodular maximization under various constraints.
problem Maximizing decomposable submodular functions under constraints while preserving privacy.
method Designing differentially private algorithms for both monotone and non-monotone decomposable submodular maximization under general matroid constraints.
result Improved utility guarantees and competitive performance compared to non-private algorithms.
We analyze a class of estimators based on convex relaxation for solving high-dimensional matrix decomposition problems. The observations are noisy realizations of a linear transformation X of the sum of an approximately) low rank matrix Θ⋆ with a second matrix Γ⋆ endowed with a complementary …
XPINNs improve generalization by decomposing PDEs but may overfit.
problem Understanding when XPINNs outperform PINNs in generalization.
method Theoretical bounds and empirical validation.
result XPINNs improve generalization by decomposing complex PDEs but may overfit.
New method improves robust low-rank matrix completion for computer vision.
problem Robust low-rank matrix completion for partially observed data.
method Formulated as a nonsmooth Riemannian optimization problem over Grassmann manifold, solved with an alternating manifold proximal gradient continuation method.
result Demonstrated advantages over existing approaches in background extraction from surveillance videos.
Aitchison and Rubinstein constructed two knot complements that can be decomposed into two regular ideal dodecahedra. This paper shows that these knot complements are the only knot complements that decompose into n regular ideal dodecahedra, providing a partial solution to a conjecture of Neumann and Reid.
The classes of Monge-Ampère systems, decomposable and bi-decomposable Monge-Ampère systems, including equations for improper affine spheres and hypersurfaces of constant Gauss-Kronecker curvature are introduced. They are studied by the clear geometric setting of Lagrangian contact structures, based on the existence of …
New space of tensorial bodies defined, properties and representatives studied.
problem Characterizing convex bodies in tensor norms.
method Introduced a new space of tensorial bodies, defined a Banach-Mazur distance, and proved existence of a compact type compactum.
result Topological representatives for the space of tensorial bodies and the Banach-Mazur type compactum are given.