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.

168,742 papers · 148 categories

Trend · papers per month

240480720960 · Jun 202019922001200920172026
48 results for orthogonal networks

Orthogonal initialization does not speed up training in ultra-wide neural networks.

problem Exploring the effect of orthogonal initialization on training speed in deep neural networks.
method Study of neural tangent kernel dynamics in FCNs and CNNs with orthogonal initialization.
result The NTK of orthogonally-initialized networks remains constant during training, suggesting no speedup in the NTK regime.

OPT framework improves neural network generalization by learning an orthogonal transformation.

problem Improving neural network generalization.
method Orthogonal over-parameterized training (OPT) framework that minimizes hyperspherical energy.
result OPT framework provably minimizes hyperspherical energy and improves empirical generalization.

Method estimates heterogeneous causal effects on networks using orthogonal learning.

problem Challenges in estimating causal effects on networks due to treatment effects on both treated and neighbors, and network homophily.
method Two-stage orthogonal learning framework: first stage uses graph neural networks for nuisance components, second stage residualizes and interpretable attention-based model for causal effects.
result Improves heterogeneous effect estimation and supports interpretable analyses.

Orthogonal initialization speeds up convergence in deep linear networks.

problem The impact of initialization on convergence speed and model performance in deep neural networks.
method Analysis of orthogonal initialization in deep linear networks, proving its superiority over Gaussian initialization.
result Orthogonal initialization speeds up convergence relative to Gaussian initialization in deep networks.

Deep networks with orthogonal weights show stable fluctuations, improving generalization and training speed.

problem Fluctuations in deep networks with Gaussian weights can impair training, especially in networks with depth comparable to width.
method Analytical and numerical studies of fully-connected networks with orthogonal weight initialization and tanh activations.
result Rectangular networks with orthogonal weights have stable fluctuations independent of network depth, leading to better generalization and training speed.

New method enforces orthogonality in convolutional layers for improved robustness.

problem Improving adversarial robustness in deep learning models.
method Applying the Cayley transform to skew-symmetric convolutions in the Fourier domain.
result The proposed method preserves orthogonality and enhances adversarial robustness compared to existing techniques.

An algorithm for efficient computation of equivariant neural network layers.

problem Efficiently computing with Brauer's group equivariant neural network layers.
method Category theoretic constructions and Kronecker product matrices.
result Significant reduction in computational cost compared to naive implementation.

SpecNet2 improves spectral embedding without orthogonalization, achieving better performance and efficiency.

problem Improving spectral embedding methods for better performance and efficiency.
method Optimizes an equivalent objective of the eigen-problem without orthogonalization, allowing separate row and column sampling.
result Local and global convergence of the new objective using batch-based gradient descent is proven, and improved performance and efficiency are demonstrated on simulated and image datasets.

Develops a new framework to measure network connectedness across and within markets.

problem Lack of flexible methods to measure network connectedness and its evolution.
method Allows network nodes to be connected in clusters, with shocks orthogonal across clusters and correlated within clusters.
result Demonstrates the effectiveness of the new framework in a detailed empirical analysis of equity markets.

Random convolutional networks can be fooled with adversarial examples.

problem Existence of adversarial examples for random convolutional networks.
method Utilizing isoperimetric inequalities on the special orthogonal group so(d)\mathbb{so}(d).
result Adversarial examples exist for various random convolutional networks.

Gradient descent biases towards stable rank networks for nearly-orthogonal data.

problem Understanding implicit bias in non-smooth neural networks trained by gradient descent.
method Analysis of two-layer ReLU and leaky ReLU networks trained by gradient descent on nearly-orthogonal data.
result Gradient descent biases towards networks with stable rank and uniform margin for nearly-orthogonal data.

Characterizes group-equivariant neural networks for three groups.

problem Understanding equivariant neural networks for orthogonal, special orthogonal, and symplectic groups.
method Characterized all possible group-equivariant neural networks for three groups.
result Found spanning sets of matrices for learnable, linear equivariant layer functions.

DONUT improves treatment effect estimation by enforcing orthogonality constraints.

problem Estimating treatment effects from observational data is challenging due to unobserved outcomes.
method DONUT uses a regularization framework that formalizes unconfoundedness as orthogonality, leading to deep orthogonal networks.
result DONUT outperforms state-of-the-art methods in estimating average treatment effects.

This work proves the asymptotic freeness of layerwise Jacobians in MLPs with Haar orthogonal matrices.

problem Proving the asymptotic freeness of layerwise Jacobians in multilayer perceptrons (MLPs).
method Replacing each layer's parameter matrix with itself multiplied by a Haar orthogonal matrix, and using the invariance of the MLP.
result Proves the asymptotic freeness of layerwise Jacobians in MLPs with Haar orthogonal matrices.

New algorithm speeds up group equivariant neural networks computations.

problem Challenging computations in group equivariant neural networks.
method Diagrammatic framework based on category theory for matrix multiplication.
result Exponential improvement in time complexity for matrix multiplication.

Orthogonium offers unified, efficient layers for robust deep learning.

problem Fragmented and computationally demanding implementations of orthogonal and 1-Lipschitz layers.
method Unified, efficient PyTorch library providing orthogonal and 1-Lipschitz layers.
result Reduced overhead and standardized tools for robust experimentation.

This study explains gradient flow dynamics in neural networks for small initialisation.

problem Understanding the training dynamics of neural networks for small initialisation.
method Analysis of gradient flow dynamics for one-hidden layer ReLU networks with orthogonal inputs.
result Gradient flow converges to zero loss and characterizes implicit bias towards minimum variation norm.

We consider the problem of sampling from posterior distributions for Bayesian models where some parameters are restricted to be orthogonal matrices. Such matrices are sometimes used in neural networks models for reasons of regularization and stabilization of training procedures, and also can parameterize matrices of bo…

2019-01-23abs ↗pdf ↗

Recurrent Neural Networks (RNNs) are designed to handle sequential data but suffer from vanishing or exploding gradients. Recent work on Unitary Recurrent Neural Networks (uRNNs) have been used to address this issue and in some cases, exceed the capabilities of Long Short-Term Memory networks (LSTMs). We propose a simp…

2017-07-29abs ↗pdf ↗

Neural networks learn incrementally from orthogonal data, interpolating with minimal complexity.

problem Understanding the learning dynamics and implicit bias in ReLU networks with orthogonal data.
method Gradient flow analysis of two-layer ReLU networks from small initialization with orthogonal training data.
result The learned interpolator has a squared 2\ell_2-norm scaling as n\sqrt{n}, close to the minimal interpolator's complexity.

Proposes a novel method for detecting novelty in multi-modal data.

problem Challenges in detecting novelty in high-dimensional, multi-modal data.
method Orthogonalized latent space for disentangling features and defining novelty score.
result Proposed method outperforms state-of-the-art algorithms in novelty detection.

Recently mean field theory has been successfully used to analyze properties of wide, random neural networks. It gave rise to a prescriptive theory for initializing feed-forward neural networks with orthogonal weights, which ensures that both the forward propagated activations and the backpropagated gradients are near $…

2018-10-09abs ↗pdf ↗

Different neural networks trained on the same dataset often learn similar input-output mappings with very different weights. Is there some correspondence between these neural network solutions? For linear networks, it has been shown that different instances of the same network architecture encode the same representatio…

2018-11-28abs ↗pdf ↗

New algorithms reduce orthogonality constraint enforcement time in machine learning.

problem Efficiently solving orthogonality constraints in machine learning.
method Extending the landing algorithm to Stiefel manifold, incorporating stochastic and variance reduction techniques.
result All proposed methods achieve the same convergence rate as Riemannian counterparts enforcing constraints.

Initialization of parameters in deep neural networks has been shown to have a big impact on the performance of the networks (Mishkin & Matas, 2015). The initialization scheme devised by He et al, allowed convolution activations to carry a constrained mean which allowed deep networks to be trained effectively (He et al.…

2017-02-21abs ↗pdf ↗

Anti-transfer learning prevents misleading representations for speech tasks.

problem Misleading representations learned from orthogonal tasks in speech processing.
method Penalizes similarity between activations of a network and another trained on an orthogonal task.
result Improves classification accuracy and invariance to the orthogonal task.

Enhances Gaussian processes with spherical features for better scalability and flexibility.

problem Lack of representation learning in Gaussian processes compared to deep neural networks.
method Introduces spherical inter-domain features to improve GP approximation and scalability.
result The method alleviates limitations and improves scalability compared to alternative strategies.

Recurrent neural networks (RNNs) have been successfully used on a wide range of sequential data problems. A well known difficulty in using RNNs is the \textit{vanishing or exploding gradient} problem. Recently, there have been several different RNN architectures that try to mitigate this issue by maintaining an orthogo…

2018-11-09abs ↗pdf ↗

Paper improves Monte Carlo sampling with new theoretical insights and methods.

problem Improving Monte Carlo sampling for variance reduction.
method Theoretical analysis of negatively dependent random variables and novel extensions using number theory and particle algorithms.
result Near-Orthogonal Monte Carlo (NOMC) consistently outperforms Orthogonal Monte Carlo (OMC) in various applications.

Many modern big data applications feature large scale in both numbers of responses and predictors. Better statistical efficiency and scientific insights can be enabled by understanding the large-scale response-predictor association network structures via layers of sparse latent factors ranked by importance. Yet sparsit…

2017-04-26abs ↗pdf ↗

Two-layer networks trained on low-dimensional subspaces are vulnerable to adversarial examples.

problem Vulnerability of two-layer neural networks to adversarial examples on low-dimensional subspaces.
method Analysis of gradient behavior and effect of initialization scale and regularization.
result Decreasing initialization scale or adding L2 regularization can improve robustness to adversarial perturbations orthogonal to the data.

GAIT-prop derives a biologically plausible learning rule from backpropagation.

problem Biological implausibility in traditional backpropagation for neural networks.
method GAIT-prop uses a top-down model to convert output error into plausible targets for weight updates.
result GAIT-prop and backpropagation give identical weight updates under certain conditions.