Random convex analysis tackles problems in random environments.
problem Dealing with problems in random environments like conditional convex risk measures.
method Developing random convex analysis over random locally convex modules, establishing inferior limit behavior, continuity, subdifferentiability, and approximating ε-subdifferentials.
result Established relationships among subdifferentiability, Gâteaux-differentiability, and Fréchet-differentiability for proper L0-convex functions. To provide a solid analytic foundation for the module approach to conditional risk measures, this paper establishes a complete random convex analysis over random locally convex modules by simultaneously considering the two kinds of topologies (namely the (ε,λ)--topology and the locally L0-- convex topolo…
The purpose of this paper is to give a selective survey on recent progress in random metric theory and its applications to conditional risk measures. This paper includes eight sections. Section 1 is a longer introduction, which gives a brief introduction to random metric theory, risk measures and conditional risk measu…
We extend to the framework of locally L0-convex modules some results from classical convex analysis. Namely, randomized versions of Mazur lemma and Krein-Smulian theorem under mild stability properties are provided.
Develops a new deep learning framework for privacy-preserving text representations.
problem Privacy concerns in deep learning frameworks requiring data pooling to a trusted server.
method Three modules: embedding, randomization, and classifier. Novel LDP protocol reduces privacy impact on accuracy.
result Framework delivers comparable or better performance than non-private and existing LDP protocols.
We study optimal investment strategies that maximize expected utility from consumption and terminal wealth in a pure-jump asset price model with Markov-modulated (regime switching) jump-size distributions. We give sufficient conditions for existence of optimal policies and find closed-form expressions for the optimal v…
Extends Chern character theory to dg algebras, proving index theorems and constructing path integrals.
problem Constructing Chern character for θ-summable Fredholm modules over dg algebras.
method Introduced θ-summable Fredholm modules, constructed Chern character as a cocycle, proved index theorem.
result Rigorous construction of path integral for N=1/2 supersymmetry satisfying localization formula.
New algorithm tackles non-convex matrix completion in semi-random settings.
problem Matrix completion in semi-random environments with varying observation probabilities.
method Proposes a pre-processing step to re-weight semi-random input, followed by a nearly-linear time algorithm.
result Recovering ground-truth matrix using non-convex local minima after pre-processing.
FedCONST adapts update magnitudes to enhance feature generalization in FL.
problem Heterogeneous client data in FL leads to overfitting and distorted transferable features.
method FedCONST uses linear convex constraints to stabilize training and preserve generalization.
result FedCONST enhances feature transferability and robustness, achieving state-of-the-art performance.
End-to-end method for robot localization in simulated and real environments.
problem Generating robot actions to maximize pose disambiguation in a reference map.
method Differentiable learning of perception and planning modules, using convolutional neural networks and deep reinforcement learning.
result The system outperforms traditional approaches for perception or planning.
Proposes a new neural head for asymmetric representation learning.
problem Asymmetric representation learning in directed relations.
method Role-aware neural convex divergence head.
result Role-aware projections improve directional accuracy over plain ICNN-Bregman heads.
Matrix completion algorithms work well with random initialization.
problem Matrix completion with positive semidefinite constraints.
method Proved the absence of spurious local minima for non-convex optimization.
result Non-convex optimization algorithms can find global minima with arbitrary initialization.
An important part of the classical theory of real or complex manifolds is the theory of (smooth, real analytic or complex analytic) vector bundles. With any vector bundle over a manifold (M,F) the sheaf of its (smooth, real analytic or complex analytic) sections is associated which is a locally free sheaf of F-modules,…
Proposes a novel network-based neighborhood regression for biological systems.
problem Lack of comprehensive analysis on biological modules using both global and local network data.
method Develops a community-wise least square optimization approach to analyze gene modules and their regulatory strength.
result Achieves exact minimax optimality and linear consistency in identifying gene module associations.
Unified analysis of multi-attribute graph learning with non-convex penalties.
problem Graph inference from multi-attribute data.
method Penalized log-likelihood objective function with ADMM and local linear approximation.
result Local consistency in support recovery and precision matrix estimation for non-convex penalties.
Gradient ascent solves tensor decomposition efficiently, proving all local maxima are global.
problem Optimizing tensor decomposition problems in machine learning.
method Gradient ascent, Kac-Rice formula, random matrix theory.
result Gradient ascent guarantees solving tensor decomposition problems efficiently.
Global optimization for low-rank matrix recovery from noisy measurements.
problem Low-rank matrix recovery from noisy measurements.
method Factorized parametrization, curvature bound, stochastic gradient descent.
result Global convergence guarantee for stochastic gradient descent from random initialization.
The hermitian analog of Aleksandrov's area measures of convex bodies is investigated. A characterization of those area measures which arise as the first variation of unitarily invariant valuations is established. General smooth area measures are shown to form a module over smooth valuations and the module of unitarily …
Proposes ContSup to boost local learning by supplying context between isolated modules.
problem Local learning's performance degrades with more isolated modules.
method Theoretical analysis and ContSup scheme to supply context between modules.
result Significant performance improvement with minimal overhead.
We study the convex duality method for robust utility maximization in the presence of a random endowment. When the underlying price process is a locally bounded semimartingale, we show that the fundamental duality relation holds true for a wide class of utility functions on the whole real line and unbounded random endo…
Paper extends Schur's theorem to spherical curves via monotonicity.
problem Comparing chord lengths of convex and spherical curves.
method Monotonicity and expansion module approach.
result Schur's Theorem extended to spherical curves.
Model captures external influences through random parameters and regime switching.
problem Capturing external influences in asset dynamics with uncertainty and regime changes.
method Developed a stochastic model with random parameters and regime switching, mathematically consistent and interpretable.
result Demonstrated the model's versatility through local volatility models and characteristic functions.
Study on skein module dimensions at irreducible representations.
problem Dimension of skein module at irreducible representations.
method Localization of skein module at maximal ideal corresponding to irreducible representation.
result Localization forms a one-dimensional free module over the unreduced coordinate ring.
GridPyM handles grid diagrams for knot theory.
problem Handling grid diagrams for knot theory.
method Generates and simplifies grids, models local transformations.
result Models local transformations between grid diagrams.
New research shows parallel optimization is ineffective for convex problems.
problem The inefficiency of parallel optimization methods for convex problems.
method Lower bounds analysis in the local oracle model of computation.
result Parallel and randomized algorithms cannot speed up convex optimization in various geometries and objective functions.
A fast method for decentralized non-convex optimization over networks.
problem Decentralized non-convex optimization problems over a network of nodes.
method GT-SAGA, a randomized incremental gradient method that evaluates one component gradient per node per iteration.
result GT-SAGA achieves almost sure and mean-squared convergence to a first-order stationary point for general smooth non-convex problems.
If M is an oriented 3-manifold, let S(M) denote the Homflypt skein module of M. We show that S(M_1 connect sum M_2) is isomorphic to S(M_1) tensor S(M_2) modulo torsion. In fact, we show that S(M_1 connect sum M_2) is isomorphic to S(M_1) tensot S(M_2) if we are working over a certain localized ring. We show the simila…
Novel GP-modulated Cox process framework with linear inequality constraints.
problem Modeling point patterns with positiveness and inequality constraints.
method Directly impose positiveness and inequality constraints on the Gaussian process without restrictions on covariance functions.
result Accurate inference of intensity functions with improved results for monotonic processes.
The paper constructs quasi-isomorphisms for cyclic homology of group actions.
problem Computing cyclic homology and periodic cyclic homology of group actions.
method Explicit quasi-isomorphisms for crossed-product algebras and locally convex algebras.
result New spectral sequences for cyclic homology of group actions.
New methods for federated learning reduce communication costs.
problem Efficiently solving optimization problems in a distributed setting.
method Developed two strategies for achieving consensus in federated learning: fixed number of local steps and randomized computations.
result Convergence analysis and experiments show benefits of the proposed methods.
The paper develops residue currents for cohesive modules and proves a generalized Poincaré-Lelong formula.
problem Analyzing coherent sheaves on complex manifolds using global analytic methods.
method Developing residue currents for cohesive modules and proving their properties.
result Proves a generalized Poincaré-Lelong formula for cohesive modules.
Skein lasagna module calculates 4-manifold invariants using handle decompositions.
problem Calculating invariants of 4-manifolds and links.
method Express skein lasagna module in terms of handle decompositions.
result Skein lasagna module can be locally infinite dimensional.
New maps help understand deformations of modules over Lie algebroids.
problem Understanding deformations of modules over Lie algebroids.
method Introduce semiregularity maps and use DG-Lie algebra control.
result Semiregularity maps annihilate obstructions under certain conditions.
Gradient descent with random init solves 1HL NNs in under-param regime.
problem Learning a one-hidden-layer neural network with quadratic activations.
method Provable gradient-based method with random initialization.
result Gradient descent iterates converge to globally optimal model with linear rate.
This paper improves image super-resolution by integrating cross-scale non-local attention.
problem Improving image super-resolution by leveraging long-range and cross-scale feature correlations.
method Proposes a Cross-Scale Non-Local (CS-NL) attention module integrated into a recurrent neural network.
result Significantly improved performance on SISR benchmarks.
New schemes improve error estimates for sampling from non-log-concave distributions.
problem Improving sampling from non-log-concave distributions with super-linear drift growth.
method Developed tamed Euler and randomized Euler schemes with error estimates.
result Near-optimal error bounds for sampling and optimization problems.
Previous work (Pradines, 1966, Aof and Brown, 1992) has given a setting for a holonomy Lie groupoid of a locally Lie groupoid. Here we develop analogous 2-dimensional notions starting from a locally Lie crossed module of groupoids. This involves replacing the Ehresmann notion of a local smooth coadmissible section of a…
We use recoupling theory to study the Kauffman bracket skein module of the quaternionic manifold over Z[A,A^{-1}] localized by inverting all the cyclotomic polynomials. We prove that the skein module is spanned by five elements. Using the quantum invariants of these skein elements and the Z_2 homology of the manifold, …
New protocols for locally private learning of linear models with reduced data leakage.
problem Locally private learning of linear models with minimal data leakage.
method Noninteractive LDP convex optimization protocols for generalized linear losses and Euclidean median problems.
result First algorithms with sub-exponential dependence on dimensionality for nonsmooth losses.
Diagrammatic method calculates knot pairings in 3-sphere.
problem Computing knot pairings in 3-sphere.
method Diagrammatic computation of bilinear forms.
result Constructs bilinear forms on twisted Alexander modules.
Consider the middle perversity intersection cohomology groups of various compactifications of a Hermitian locally symmetric space. Rapoport and independently Goresky and MacPherson have conjectured that these groups coincide for the reductive Borel-Serre compactification and the Baily-Borel-Satake compactification. Thi…
Study on the dimension of skein modules of Dehn fillings for specific knots.
problem Determining the dimension of skein modules for Dehn fillings of knots.
method Analyzing Kauffman bracket skein modules and using character varieties.
result Dimension of skein modules for almost all primitive roots of unity and slopes.
This paper re-visits the spectral method for learning latent variable models defined in terms of observable operators. We give a new perspective on the method, showing that operators can be recovered by minimizing a loss defined on a finite subset of the domain. A non-convex optimization similar to the spectral method …
Cochran defined the nth-order integral Alexander module of a knot in the three sphere as the first homology group of the knot's (n+1)th-iterated abelian cover. The case n=0 gives the classical Alexander module (and polynomial). After a localization, one can get a finitely presented module over a principal ideal domain,…
For a ring R, we denote by R[L] the free R-module spanned by the isotopy classes of singular links in S3. Given two invertible elements x,t∈R, the HOMFLY-PT skein module of singular links in S3 (relative to the triple (R,t,x)) is the quotient of R[L] by local rela…
Overlap-Local-SGD improves distributed SGD by overlapping communication and computation.
problem High communication delay and node slowdown in distributed SGD.
method Adding an anchor model to synchronize local updates and pull them towards the anchor model.
result Overlap-Local-SGD speeds up distributed training and mitigates straggler effects.
This paper proposes an alternative to E2E training for deep networks, reducing memory footprint.
problem High GPUs memory footprint in end-to-end training of deep networks.
method Locally supervised learning with information propagation loss to avoid information collapse.
result The proposed method achieves competitive performance with less than 40% memory footprint compared to E2E training.
L-modules are a combinatorial analogue of constructible sheaves on the reductive Borel-Serre compactification of a locally symmetric space. We define the micro-support of an L-module; it is a set of irreducible modules for the Levi quotients of the parabolic Q-subgroups associated to the strata. We prove a vanishing th…