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.

169,181 papers · 148 categories

Trend · papers per month

4328651,2971,729 · Jun 202019922001200920182026
48 results for CNN structure learning

Graph CNNs adapt to varying graph structures for better performance.

problem Fixed graph structures limit the performance of Graph CNNs on real data.
method Adaptive graph learning and distance metric learning for efficient graph construction.
result Adaptive Graph CNNs improve convergence speed and predictive accuracy on various graph datasets.

A new framework combines CNN and GRU for better structural damage detection.

problem Improving damage detection in structural engineering using machine learning.
method Hierarchical CNN and Gated Recurrent Unit (GRU) framework to model spatial and temporal relations.
result The proposed HCG framework significantly outperforms existing methods for structural damage detection.

Transform classical network structures to graph CNN for better graph recognition.

problem Transforming classical network structures to graph CNN for better graph recognition.
method Review and introduce ResNet, Inception, and DenseNet into graph CNN, constructing G_ResNet, G_Inception, G_DenseNet.
result Demonstrated how different network structures work on graph CNN in the graph recognition task.

Deep structured output learning shows great promise in tasks like semantic image segmentation. We proffer a new, efficient deep structured model learning scheme, in which we show how deep Convolutional Neural Networks (CNNs) can be used to estimate the messages in message passing inference for structured prediction wit…

2015-06-06abs ↗pdf ↗

CNNs reveal retinal ganglion cell features, linking visual processing to neuroscience.

problem Understanding what CNNs learn about retinal neuronal circuits.
method Trained CNNs on white noise images to predict neural responses from salamander retinas.
result CNN filters resemble biological retinal components and ganglion cell receptive fields.

ResNet-type CNNs achieve optimal error rates in function classes with block-sparse structures.

problem Optimal approximation and estimation in function classes with sparse constraints.
method Developed ResNet-type CNNs that can approximate and estimate functions with block-sparse structures.
result ResNet-type CNNs attain minimax optimal error rates in Hölder and Barron classes.

The paper analyzes deep ReLU CNNs' approximation properties in 2D space.

problem Establishing L2L^2 approximation properties for deep ReLU CNNs.
method Analysis based on decomposition theorem for convolutional kernels, properties of ReLU activation, and connections with one-hidden-layer ReLU NNs.
result Universal approximation theorem for deep ReLU CNNs with classic structure.

Combines CNN and RNN for hierarchical image classification.

problem Hierarchical relations between image categories are not captured by flat classifiers.
method Uses a CNN for feature extraction and an RNN for capturing hierarchical class relations. Incorporates residual learning.
result Hierarchical networks outperform state-of-the-art CNNs on a real-world dataset.

CNNs use a bottleneck structure to focus on a few frequencies, affecting function representation.

problem Understanding how CNNs focus on specific frequencies in their feature learning.
method Defined Convolution Bottleneck (CBN) structure, measured CBN rank, and analyzed parameter norms.
result Parameter norm scales with depth and CBN rank, and networks with optimal parameters exhibit this structure.

3D Steerable CNNs learn equivariant features for 3D data.

problem Learning rotationally equivariant features in volumetric data.
method SE(3)-equivariant convolutions using steerable kernel basis.
result 3D Steerable CNNs are effective for protein structure classification and amino acid propensity prediction.

Paper proposes a new technique to compress CNNs while maintaining accuracy.

problem CNNs struggle with traditional low-rank approximation methods, leading to degraded accuracy.
method Introduces a training technique that finds a flat minimum in low-rank approximation without a decomposed structure.
result CNN models can be compressed with higher accuracy and lower computation than conventional methods.

New CNN method improves brain MR segmentation across scanners and protocols.

problem Degradation of CNN accuracy on images from different scanners and protocols.
method Lifelong multi-domain learning with shared filters and domain-specific batch normalization.
result Significantly closes the gap to benchmark performance.

Ego-CNN detects critical structures in graphs efficiently.

problem Lack of precise detection of critical structures in existing graph embedding models.
method Ego-CNN uses ego-convolutions at each layer and stacks them in an ego-centric way.
result Ego-CNN achieves comparable task performance to state-of-the-art models and can incorporate scale-free priors.

DGCNN improves graph CNNs by handling irregular graphs.

problem Handling structural information loss and redundancy in graph CNNs.
method Proposes DGCNN using DGCL with mixed Gaussian model to handle irregular graphs.
result DGCNN outperforms state-of-the-art methods in graph classification and retrieval.

Study on infinitely-wide CNNs and their adaptability to function spatial scales.

problem Understanding how CNNs efficiently learn high-dimensional functions and their adaptability to function spatial scales.
method Study infinitely-wide deep CNNs in the kernel regime, characterizing their spectrum and using generalisation bounds to prove adaptability.
result Deep CNNs adapt to the spatial scale of the target function, with error decay controlled by the effective dimensionality of function subsets.

Computational approaches to drug discovery can reduce the time and cost associated with experimental assays and enable the screening of novel chemotypes. Structure-based drug design methods rely on scoring functions to rank and predict binding affinities and poses. The ever-expanding amount of protein-ligand binding an…

2016-12-08abs ↗pdf ↗

CNN predicts airfoil lift coefficients across various conditions.

problem Developing a CNN for predicting airfoil lift coefficients under different conditions.
method Trained multiple CNN architectures on airfoil lift coefficients with varying flow conditions and object geometries.
result CNN model shows competitive prediction accuracy with minimal geometric constraints.

CNN improves causal inference by controlling time-structured covariates.

problem Estimating the effect of early retirement on health outcomes while controlling for time-structured covariates.
method Used CNN to fit nuisance models explaining treatment and outcome, combining them into an augmented inverse probability weighting estimator.
result Uniformly valid inference achieved through CNN, providing rates of convergence and uniformly valid inference guarantees.

ResRep prunes CNNs without losing accuracy by separating remembering and forgetting.

problem Pruning CNNs to reduce FLOPs without sacrificing accuracy.
method Decoupling remembering and forgetting in CNNs, using SGD for remembering and a novel update rule for forgetting.
result Achieved lossless pruning with high compression ratio (76.15% accuracy on ImageNet with 45% FLOPs reduction).

BCD-Net uses identical CNN structures for image recovery in undersampled imaging.

problem Challenges in obtaining accurate images from undersampled or noisy measurements.
method Incorporates image mapping CNN into BCD signal recovery method using alternating direction method of multipliers.
result Significantly more accurate image recovery compared to existing methods.

Symmetric CNNs improve sequential recommendation and protein structure prediction.

problem Improving prediction accuracy in sequential recommendation and protein structure inference.
method Developed a CNN architecture that preserves symmetry in convolutional layers, using parameterized convolutional kernels.
result Symmetric structured CNNs achieve better performance with fewer parameters.

New bounds for CNNs show better generalization than previous models.

problem Improving understanding of CNNs' generalization ability.
method Proposed tighter generalization bounds for CNNs by exploiting the sparse and permutation structure of weight matrices and spectral norms of convolution operations.
result Theoretical and experimental results show tighter bounds for CNNs than existing bounds.

The paper analyzes deep neural networks using information theory to improve classification accuracy.

problem Improving classification accuracy in deep neural networks.
method Modeling the output of convolutional filters as a random variable conditioned on class and network structure, computing conditional entropy as a compact code.
result The conditional entropy feature analysis leads to higher classification accuracy than the original CNN.

New CNN approach detects cell nuclei with prior information.

problem Challenges in detecting cell nuclei due to image quality and morphology diversity.
method Develops SP-CNN with trainable shape prior layer to guide CNN learning.
result TSP-CNN outperforms state-of-the-art alternatives on challenging datasets.

New CNN model for fast unsupervised document embedding.

problem Efficient unsupervised document embedding with parallelizable architecture.
method Convolutional Neural Network (CNN) for parallel inference and stochastic forward prediction for unsupervised learning.
result Comparable accuracy to state-of-the-art at significantly reduced computational cost.

NTS-NOTEARS learns DBNs from time-series data with prior knowledge.

problem Learning dynamic Bayesian networks from time-series data with nonlinear and lagged relationships.
method Uses 1D CNNs to model DBNs, incorporating prior knowledge as constraints.
result Achieves state-of-the-art DAG structure quality compared to parametric and nonparametric methods.