Research
On-device research index

arXiv research

A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.

169,051 papers · 148 categories

Trend · papers per month

70141211281 · Jun 202019922001200920172026
48 results for intrinsic weights

A new classifier uses weighted orthogonal regression for robust classification with limited data.

problem Challenges in classification with insufficient training data.
method Exploits intrinsic structure of data through Eigen components with specific weights determined by eigenvalues.
result Robust learning in classification problems with limited data.

New method detects intrinsic cross-correlations in non-stationary time series affected by common factors.

problem Bias in cross-correlation analysis due to common external factors.
method Multifractal temporally weighted detrended partial cross-correlation analysis (MF-TWDPCCA).
result MF-TWDPCCA accurately detects intrinsic cross-correlations between non-stationary time series.

Deep multi-task learning benefits from low intrinsic dimensionality, leading to better generalization.

problem Improving generalization in deep multi-task learning with high-dimensional models.
method Parametrizing multi-task networks in a low-dimensional space using random expansions and weight compression.
result First non-vacuous generalization bounds for deep multi-task networks are derived.

The paper classifies hypersurfaces with constant weighted mean curvature.

problem Characterizing and classifying hypersurfaces with specific curvature properties.
method Using intrinsic properties of the second fundamental form and analyzing weighted volume and growth.
result Characterization of hyperplanes and generalized round cylinders.

Empirical study shows SGD's random seed impacts model weights more than training examples, suggesting intrinsic privacy.

problem Understanding and leveraging the intrinsic randomness of SGD for improved privacy and utility.
method Large-scale empirical study on 120,000 models across four datasets, focusing on convex and non-convex objectives.
result Intrinsic randomness of SGD can reduce the need for additional noise to achieve privacy guarantees, with estimated εi(D)ε_i(\mathcal{D}) as low as 6.3.

New regularizer improves neural network robustness and generalization.

problem Ineffective weight decay for networks with homogeneous activation functions.
method Proposes an invariant regularizer to penalize intrinsic weight norms.
result Improves generalization and adversarial robustness on various datasets.

Shallow nonlinear networks can separate classes linearly with polynomially scaling width.

problem Understanding the linear separability of deep networks' features.
method Modeling inputs as a union of low-dimensional subspaces and using random weights and quadratic activations.
result Shallow nonlinear networks can achieve linear separation with polynomially scaling width.

The paper aims at proving global height estimates for Killing graphs defined over a complete manifold with nonempty boundary. To this end, we first point out how the geometric analysis on a Killing graph is naturally related to a weighted manifold structure, where the weight is defined in terms of the length of the Kil…

2016-12-05abs ↗pdf ↗

The paper compares isoperimetric quotients and capacities in weighted manifolds.

problem Comparing isoperimetric quotients and capacities in weighted manifolds.
method Analysis of weighted Laplacian of the distance function and techniques for non-compact submanifolds.
result Parabolicity and hyperbolicity criteria for weighted manifolds.

In this work we study the intrinsic geometry of the space of Kahler metrics under various Riemannian metrics. The first part is on the Dirichlet metric. We motivate its study, we compute its curvature, and we make links with the Calabi metric, the K-energy, the degenerate complex Hessian equation. The second part is on…

2012-02-29abs ↗pdf ↗

GeLoRA optimizes LoRA fine-tuning by dynamically adjusting ranks based on intrinsic dimensionality.

problem Efficient fine-tuning of large language models with limited computational resources.
method GeLoRA computes intrinsic dimensionality to adaptively select LoRA ranks, balancing expressivity and efficiency.
result GeLoRA consistently outperforms recent baselines within the same parameter budget on multiple tasks.

A well known question in differential geometry is to control the constant in isoperimetric inequality by intrinsic curvature conditions. In dimension 2, the constant can be controlled by the integral of the positive part of the Gaussian curvature. In this paper, we showed that on simply connected conformal flat manifol…

2013-06-07abs ↗pdf ↗

Novel defense algorithm improves SVMs against data poisoning attacks.

problem Vulnerability of SVMs to targeted training data manipulations like poisoning attacks.
method Developed a weighted SVM using K-LID to de-emphasize suspicious data samples.
result Significant reduction in classification error rates (10% on average) with the proposed defense.

New summary measures reveal geometric structure in weighted measures on manifolds.

problem Lack of geometric information in standard weight-only summaries.
method Heat-kernel entropy profiles, tracking nonuniformity across scales.
result Geometric effective sample size discounts nearby or duplicate particles.

We construct an explicit scheme to associate to any potential symbol an operator acting between sections of natural bundles (associated to irreducible representations) for a so-called AHS-structure. Outside of a finite set of critical (or resonant) weights, this procedure gives rise to a quantization, which is intrinsi…

2009-04-21abs ↗pdf ↗

Introduces Exponentially Weighted Signature for better path representation.

problem Uniform treatment of historical information in signatures.
method Generalizes EFM signature to bounded linear operators, enabling contextualised temporal weighting.
result EWS is the unique solution to a linear controlled differential equation and generalizes state-space models.

Bayesian deep learning with heavy-tailed weights achieves near-optimal performance.

problem Deep neural networks with heavy-tailed weights achieve near-optimal performance in various contexts.
method Introduced a Bayesian deep learning prior based on heavy-tailed weights and ReLU activation, showing near-optimal minimax contraction rates.
result Posterior distribution achieves near-optimal minimax contraction rates, adaptive to smoothness and intrinsic dimension.

We define a discrete Laplace-Beltrami operator for simplicial surfaces. It depends only on the intrinsic geometry of the surface and its edge weights are positive. Our Laplace operator is similar to the well known finite-elements Laplacian (the so called ``cotan formula'') except that it is based on the intrinsic Delau…

2005-03-11abs ↗pdf ↗

We endow each closed, orientable Alexandrov space (X,d)(X, d) with an integral current TT of weight equal to 1, T=0and{(}T)=X\partial T = 0 and \set(T) = X, in other words, we prove that (X,d,T)(X, d, T) is an integral current space with no boundary. Combining this result with a result of Li and Perales, we show that non-collapsing sequ…

2017-03-23abs ↗pdf ↗

Introduces intrinsic Hopf-Lax semigroup linking to intrinsic slope.

problem Understanding intrinsic Hopf-Lax semigroup and its relation to intrinsic slope.
method Introduces and proves the link between intrinsic Hopf-Lax semigroup and intrinsic slope.
result Intrinsic Hopf-Lax semigroup is a subsolution of Hamilton-Jacobi type equality.

This study uncovers how neural architectures and weights interact in classification tasks.

problem Understanding the role of neural architecture and weights in classification performance.
method Developed a novel method to find optimal task-specific architectures as binary networks with {0, 1}-valued weights, using approximate gradient descent.
result Well-trained architectures may not require fine-tuning of weights, highlighting the importance of structure over weights.

This work improves sampling efficiency on complex spaces using determinantal processes.

problem Efficient sampling from large-scale datasets with general spaces.
method Determinantal point processes on general spaces and diffusion geometry.
result Improved sampling rates for determinantal processes on Riemannian manifolds and networks.

This paper develops a new theory for ensemble learning beyond variance reduction.

problem Ensemble learning's effectiveness for stable estimators is not fully explained by variance reduction.
method Develops a general weighting theory for ensemble learning, formalizing ensembles as linear operators and introducing geometric and spectral constraints.
result Structured weights can outperform uniform averaging by reshaping approximation geometry and redistributing spectral complexity.

We study sequences of integral current spaces (Xj,dj,Tj)(X_j,d_j,T_j) such that the integral current structure TjT_j has weight 11 and no boundary and, all (Xj,dj)(X_j,d_j) are closed Alexandrov spaces with curvature uniformly bounded from below and diameter uniformly bounded from above. We prove that for such sequences either the…

2014-11-25abs ↗pdf ↗

Develop a framework for barycentric projections of optimal transport plans on Riemannian manifolds.

problem Optimal transport couplings are probabilistic objects, while many learning pipelines require deterministic maps.
method Develop a framework for barycentric projections of transport couplings on Riemannian manifolds.
result The intrinsic projection maps each source point to the conditional Fréchet mean of its destination law and is shown to be the best deterministic representative under squared geodesic loss.

New geometric analysis of PWSPDs balances density and geometry in high-dimensional data.

problem Balancing density and geometry in high-dimensional data.
method Power-weighted shortest-path distances (PWSPDs) and their geometric and computational analyses.
result High probability guarantees on the equivalence of PWSPDs on complete and nearest neighbor graphs.

Cieliebak, Mundet i Riera and Salamon recently formulated a definition of branched submanifold of Euclidean space in connection with their discussion of multivalued sections and the Euler class. This note proposes an intrinsic definition of a weighted branched manifold Z that is obtained from the usual definition of or…

2005-09-28abs ↗pdf ↗

We introduce new sufficient conditions for intrinsic knotting and linking. A graph on n vertices with at least 4n-9 edges is intrinsically linked. A graph on n vertices with at least 5n-14 edges is intrinsically knotted. We also classify graphs that are 0, 1, or 2 edges short of being complete partite graphs with respe…

2007-01-15abs ↗pdf ↗

A new metric framework for weighted projective spaces improves clustering and analysis.

problem Proximity measurement in weighted projective spaces with intrinsic scaling and topology.
method Hierarchical clustering framework based on Finsler geometry, quotienting weighted scaling action.
result The constructed metric dFd_F satisfies the triangle inequality, making it a genuine metric.

This paper develops a method to learn lower-dimensional submanifolds of brain connectomes.

problem Learning lower-dimensional representations of manifold-valued data, especially brain connectomes.
method Riemannian variational autoencoder with intrinsic generative model.
result The method can learn weighted submanifolds of manifold-valued data.

New risk measure and quadrangle improve financial decision-making.

problem Heterogeneous risk assessments among analysts.
method Established analytical characterizations of WGRM and incorporated FRQ into WRQ.
result WGRM and WRQ framework improves risk-adjusted performance and downside resilience.

SWIFT learns intrinsic rewards from LLM hidden states for efficient best-of-N sampling.

problem Efficiency and scalability of reward models for LLMs.
method SWIFT (Simple Weighted Intrinsic Feedback Technique) learns a reward function directly from LLM hidden states.
result SWIFT outperforms existing baselines by 12.7% on MATH dataset while using less than 0.005% of their parameters.

New graph shows edge deletion/contraction doesn't always result in intrinsically linked graphs.

problem Edge operations in intrinsically knotted graphs don't always produce intrinsically linked graphs.
method Presented a new intrinsically knotted graph.
result Edge operations in intrinsically knotted graphs don't always result in intrinsically linked graphs.

The study defines and characterizes extrinsic catenaries in hyperbolic space.

problem Understanding catenaries in hyperbolic geometry.
method Defined extrinsic catenaries in hyperbolic plane, characterized them, and proved their relation to minimal surfaces.
result Extrinsic catenaries in hyperbolic space are critical points of a potential functional and generating curves of minimal surfaces.

New method estimates intrinsic dimensionality using angles, not distances.

problem Estimating local intrinsic dimensionality accurately.
method Introduces a new estimator using the distribution of angles between neighbor points.
result New estimator behaves similarly but complementarily to existing measures of intrinsic dimensionality.

A new method integrates forms on Riemann surfaces, leading to modular forms.

problem Integrating differential forms with poles on Riemann surfaces.
method Simple procedure to integrate differential forms with arbitrary holomorphic poles, establishing an analytic theory for integrals over configuration spaces.
result Regularized graph integrals on elliptic curves are almost-holomorphic modular forms.

A directed graph GG is intrinsically linked\textit{intrinsically linked} if every embedding of that graph contains a non-split link LL, where each component of LL is a consistently oriented cycle in GG. A tournament\textit{tournament} is a directed graph where each pair of vertices is connected by exactly one directed edge. We consider intr…

2019-01-11abs ↗pdf ↗