New method relaxes spatial invariance in locally connected layers, improving accuracy.
problem Improving classification accuracy with locally connected layers.
method Designing a low-rank locally connected layer with varying spatially varying combining weights.
result Relaxing spatial invariance improves classification accuracy over convolution and locally connected layers.
LSNN improves CNN by smoothing local receptive fields.
problem Limited capturing of local receptive fields in CNN.
method LSNN represents kernel and smoother to capture local fields' importance and relations.
result LSNN outperforms CNN and locally connected layer on MNIST variants.
Study analyzes error in ReLU networks with local connections.
problem Improving neural network performance and understanding approximation errors.
method Analyzed approximation error of ReLU networks with local connections.
result Error estimate depends on depth and width of hidden layers.
Shortcut connections in ResNet help avoid local optima, leading to efficient training.
problem Understanding why shortcut connections in ResNet lead to efficient training.
method Two-layer non-overlapping convolutional ResNet, gradient descent with proper normalization.
result Gradient descent avoids spurious local optima, converging to a global optimum.
New neuron model learns and adapts its receptive field.
problem Learning and focusing on informative inputs.
method Adaptive locally connected neuron model using backpropagation.
result Focusing neurons outperform dense layers in classification tasks.
Deep networks trained with Hebbian updates perform similarly to back-propagation on image datasets.
problem Training deep networks with realistic asymmetric connections and updates.
method Use Hebbian updates with separate feedforward and feedback weights, and local rule for updates.
result Similar performance to back-propagation achieved with Hebbian updates on challenging image datasets.
The paper proves skip connections help neural networks avoid shallow local minima.
problem Understanding how skip connections affect the loss landscape of deep neural networks.
method Theoretical analysis of the topology of loss landscapes of deep ReLU neural networks with skip connections.
result Skip connections help control the connectedness of sub-level sets, avoiding shallow local minima.
Proposes LC-ST-FCN for better ride-sourcing demand forecasting.
problem Local statistical differences in ride-sourcing demand across a city.
method LC-ST-FCN framework combining 3D and 2D convolutions, locally connected layers.
result Significant improvements in demand forecasting compared to baselines.
The paper calculates bounds on the local Lipschitz constants of neural network layers.
problem Understanding the Lipschitz constants of neural network layers for robustness analysis.
method Analytical approach to determine upper bounds on local Lipschitz constants of affine-ReLU functions.
result The method produces tighter bounds than the standard conservative bound, especially for small perturbations.
This paper presents two unsupervised learning layers (UL layers) for label-free video analysis: one for fully connected layers, and the other for convolutional ones. The proposed UL layers can play two roles: they can be the cost function layer for providing global training signal; meanwhile they can be added to any re…
Adding one neuron fixes neural network's bad local minima.
problem Non-convexity of loss function leading to many bad local minima in neural networks.
method Adding one special neuron with a skip connection to the output or one per layer.
result Every local minimum becomes a global minimum after adding one neuron.
Graph neural network simplifies to linear model with attention mechanisms.
problem Semi-supervised learning with limited labeled data.
method Proposes a graph neural network removing fully-connected layers and replacing them with attention mechanisms.
result Attention-based graph neural network outperforms existing methods on benchmark datasets.
Shallow networks with local learning rules can match deep learning performance.
problem Training deep neural networks is biologically implausible; the goal is to achieve similar performance with shallow networks.
method Investigated shallow networks with one hidden layer and a single readout layer, using various local learning rules for the hidden layer and supervised learning for the readout layer.
result Shallow networks can achieve test accuracy comparable to deep learning models, suggesting the use of different datasets for testing.
A new channel locality block improves CNN performance.
problem Improving the performance of convolutional neural networks.
method Proposed a variant of Squeeze-and-Excitation block using convolutional layers to learn nearby channel correlation.
result Our C-Local block achieved higher accuracy than the standard SE block on the cifar-10 dataset.
Proves exponential sample complexity separations in local differential privacy.
problem Sample complexity in locally private protocols.
method Connection between communication complexity and sample complexity, using specific lower bounds for two problems.
result Exponential separations between differentially private protocols.
While the optimization problem behind deep neural networks is highly non-convex, it is frequently observed in practice that training deep networks seems possible without getting stuck in suboptimal points. It has been argued that this is the case as all local minima are close to being globally optimal. We show that thi…
Convolutional layers can be mathematically equated to fully connected layers.
problem Understanding the equivalence between convolutional and fully connected layers for neural networks.
method Demonstrated that convolutional operations can be converted to matrix multiplication, showing equivalence.
result Convolutional layers and fully connected layers are mathematically equivalent in linear cases.
Artificial neural networks are most commonly trained with the back-propagation algorithm, where the gradient for learning is provided by back-propagating the error, layer by layer, from the output layer to the hidden layers. A recently discovered method called feedback-alignment shows that the weights used for propagat…
This paper introduces a hierarchical associative memory model with multiple layers.
problem Limitations of traditional associative memory models with only one hidden layer.
method Develops a fully recurrent model with arbitrary layers, including locally connected ones, and a corresponding energy function.
result The model can dynamically assemble memories using weights from lower layers and higher layers' rules.
The study shows removing fully connected output layers improves efficiency without sacrificing performance.
problem Large number of parameters in fully connected layers for high-category datasets.
method Examined architectures replacing fully connected output layers with fixed layers and compared performance.
result Fixed classifiers offer no additional benefit over removing the output layer and its parameters.
DropCluster clusters features in convolutional layers to prevent overfitting.
problem Dropout's effectiveness in convolutional layers is limited due to feature spatial correlation.
method DropCluster clusters features in convolutional layers and randomly drops clusters during training.
result DropCluster controls overfitting better than other approaches on various datasets.
Layered graphical models improve discriminative learning efficiency.
problem Improving discriminative learning efficiency in graphical models.
method Designing layered graphical models (LGMs) in analogy to neural networks, using tensorized truncated variational inference and backpropagation.
result LGMs achieve competitive results in image classification, comparable to neural networks.
Improved loss functions adapt to weight-space anisotropy, outperforming isotropic counterparts.
problem Adapting to the anisotropic nature of deep weight spaces for better performance.
method Refined local entropic loss functions restricted to a subset of weights, exploiting anisotropy.
result Partial local entropies outperform isotropic counterparts on image classification tasks.
Vision Transformers show different internal representations compared to CNNs.
problem Understanding how Vision Transformers solve image classification tasks.
method Comparative analysis of ViT and CNN architectures on image classification benchmarks.
result ViT has more uniform representations across all layers, while CNNs have more varied representations.
Wide neural networks have non-attracting local minima.
problem Understanding the impact of suboptimal local minima in deep and wide neural networks.
method Construction of non-attracting local minima and saddle points in wide neural networks.
result Wide neural networks have non-attracting local minima, even though they are not negatively impacted by suboptimal local optima.
Proposes a balanced multi-component and multi-layer neural network for efficient function approximation.
problem Accurately and efficiently approximating complex functions with high degrees of freedom and computational cost.
method Inspired by a multi-component approach, MMNN combines single-layer networks with a multi-layer decomposition strategy.
result Significant reduction in training parameters, more efficient training process, and improved accuracy compared to FCNNs or MLPs.
Piecewise linear activations create many spurious local minima in neural networks.
problem Understanding the loss surface of neural networks with piecewise linear activations.
method Proved the existence of infinite spurious local minima and partitioned the loss surface into smooth cells.
result Piecewise linear activations create many spurious local minima that are invariant under a continuous path.
DRew dynamically rewires message passing to improve long-range tasks.
problem Over-squashing in MPNNs due to local message passing.
method Layer-dependent rewiring and delay mechanism for gradual densification.
result DRew outperforms graph Transformers and multi-hop MPNNs on long-range tasks.
Deep ResNets can have better local minima than linear predictors.
problem Understanding the optimization landscape of deep ResNets compared to linear predictors.
method Analyzing the optimization landscape of ResNets with multiple residual blocks, showing geometric conditions under which ResNets have better local minima.
result Theorem showing that any critical point in the optimization landscape of deep ResNets is either at least as good as the best linear predictor or has a strictly negative eigenvalue in its Hessian.
GCN and GPCA are mathematically connected, leading to improved node classification performance.
problem Improving node classification performance in semi-supervised settings.
method Established a mathematical connection between GCN and GPCA, demonstrating their equivalence and using this to design an effective initialization strategy.
result GPCA paired with a simple MLP achieves similar or better performance than GCN on semi-supervised node classification tasks.
DenseNets improve accuracy and efficiency in convolutional networks.
problem Improving accuracy and efficiency in deep convolutional networks.
method Introducing Dense Convolutional Networks (DenseNet) with direct connections between all layers.
result DenseNets achieve significant improvements over state-of-the-art networks on object recognition benchmarks.
New method trains neural networks with local error signals, outperforming global methods.
problem Training neural networks with global error signals.
method Layer-wise training with local error signals.
result Layer-wise training with local error signals can approach state-of-the-art performance.
FCC-GAN combines fully connected and convolutional layers for improved GAN performance.
problem Lack of understanding in choosing GAN network architectures.
method Proposes FCC-GAN, a hybrid architecture combining fully connected and convolutional layers.
result FCC-GAN outperforms traditional GAN architectures in terms of learning speed and sample quality.
Deep linear networks with arbitrary loss have all local minima as global minima.
problem Finding optimal solutions in deep linear networks with arbitrary convex losses.
method Provided a short and elementary proof for all local minima being global minima under specific conditions.
result All local minima are global minima for deep linear networks with certain layer widths.
Proposes a new learning mechanism for deep networks using local errors.
problem Learning using delayed and non-local errors is hard for biological neural networks.
method Generates local errors in each layer using fixed, random auxiliary classifiers.
result Approaches standard backpropagation performance on MNIST, CIFAR10, and SVHN datasets.
Stanza separates convolutional and fully connected layers for faster deep learning training.
problem Heavy data transfer between workers and servers in distributed deep learning.
method Layer separation: most nodes train convolutional layers, others train fully connected layers only.
result Significant acceleration of training time (1.34x--13.9x) over current systems.
CRITS improves time series classification with interpretable local explanations.
problem Lack of detailed explanations in time series classification models.
method CRITS uses convolutional kernels, max-pooling, and rectified linear units to extract feature weights.
result CRITS provides intrinsically interpretable local explanations without requiring gradients or random perturbations.
Backpropagation-free RL method trains layers using local signals.
problem Vanishing or exploding gradients in backpropagation-based RL.
method Local pairwise distance matching for layer-wise training without backpropagation.
result Backpropagation-free method achieves competitive performance and stability.
AdaComp compresses gradients adaptively for efficient distributed training.
problem Communication constraints in distributed training of deep neural networks.
method Adaptive Residual Gradient Compression (AdaComp) that selects gradient residues and tunes compression rate.
result Excellent compression rates (200X-40X) without accuracy loss.
We present the Video Ladder Network (VLN) for efficiently generating future video frames. VLN is a neural encoder-decoder model augmented at all layers by both recurrent and feedforward lateral connections. At each layer, these connections form a lateral recurrent residual block, where the feedforward connection repres…
A new histogram layer improves texture analysis performance.
problem Extracting features for texture analysis from local spatial regions.
method Directly computes local spatial distribution of features during backpropagation.
result Improves performance on three material/texture datasets.
A new unpooling layer enhances graph generation in molecular models.
problem Efficient graph generation for complex models like molecules.
method Trainable unpooling layer that enlarges and restructures graphs.
result The unpooling layer improves graph generation in molecular models.
Layer normalization improves federated learning with skewed labels.
problem Label skewness in federated learning datasets.
method Identified feature normalization as key mechanism; applied to latent features before classifier.
result Normalization accelerates global training and improves convergence under extreme label shift.
HaKAN uses Hahn-KAN blocks to forecast multivariate time series.
problem Long-term time series forecasting challenges with high complexity and spectral bias.
method HaKAN integrates channel independence, patching, and a stack of Hahn-KAN blocks with residual connections. It uses Hahn polynomial-based learnable activation functions.
result HaKAN consistently outperforms state-of-the-art methods on various forecasting benchmarks.
New algorithms for deep learning mimic brain's learning but struggle with complex images.
problem Evaluating biologically inspired deep learning algorithms on complex image datasets.
method Implemented and compared various biologically inspired algorithms (TP, FA, DTP) on MNIST, CIFAR-10, and ImageNet.
result Biologically inspired algorithms perform well on MNIST but poorly on CIFAR and ImageNet, suggesting new architectures or algorithms are needed.
Deep networks with a wide layer ensure sublevel set connectivity.
problem Ensuring connectivity of sublevel sets in deep learning.
method Analyzing the connectivity of sublevel sets in deep neural networks with a specific layer width.
result A single wide layer of width N+1 suffices to prove connectivity of sublevel sets. CLAPS improves conformal regression by adaptively scaling interval widths based on last-layer Laplace uncertainty.
problem Lack of adaptive interval width scaling in conformal regression for heterogeneous inputs.
method CLAPS uses heteroscedastic last-layer Laplace uncertainty to adaptively scale interval widths, combining aleatoric and epistemic uncertainties.
result CLAPS provides competitive interval efficiency with nominal-level coverage, reducing to aleatoric scaling as epistemic uncertainty decreases.
New proof shows deep neural nets can have sub-optimal local minima.
problem Can over-parameterization eliminate sub-optimal local minima in deep neural networks?
method Counter-example with generic input data and non-linear activation functions.
result Sub-optimal local minima exist in deep neural networks regardless of width.