Developed a BP algorithm for training neural networks with 2nd order neurons.
problem Training neural networks with nonlinear quadratic operations.
method Created a general backpropagation algorithm.
result Validated the generalized BP algorithm through numerical studies.
Recently, we proposed to transform the outputs of each hidden neuron in a multi-layer perceptron network to have zero output and zero slope on average, and use separate shortcut connections to model the linear dependencies instead. We continue the work by firstly introducing a third transformation to normalize the scal…
A new method for growing neural networks by splitting neurons, improving efficiency.
problem Optimizing neural network structures, especially for lightweight architectures.
method A progressive training approach using steepest descent to adaptively grow and split neurons.
result The method provides a computationally efficient way to optimize neural network structures.
Quadratic neurons enhance deep networks' approximation capabilities.
problem Improving deep networks' expressive power and efficiency.
method Introduced quadratic neurons and analyzed their performance in deep quadratic networks.
result Quadratic networks can approximate functions more efficiently and uniquely than conventional networks.
Estimates neuronal connectivity from spike times using flexible Hawkes processes.
problem Learning latent network structure from multivariate point process data.
method Proposes a new nonstationary Hawkes process and uses sparse least squares estimation.
result Establishes non-asymptotic error bounds and selection consistency for estimated parameters.
ISAAC Newton uses input-based curvature for efficient training.
problem Efficient training in small-batch stochastic regimes.
method ISAAC Newton conditions gradients using selected second-order information based on input.
result Effective training even in small-batch stochastic regimes, competitive to first-order and second-order methods.
Dropout has been witnessed with great success in training deep neural networks by independently zeroing out the outputs of neurons at random. It has also received a surge of interest for shallow learning, e.g., logistic regression. However, the independent sampling for dropout could be suboptimal for the sake of conver…
Neural networks (NN) have achieved state-of-the-art performance in various applications. Unfortunately in applications where training data is insufficient, they are often prone to overfitting. One effective way to alleviate this problem is to exploit the Bayesian approach by using Bayesian neural networks (BNN). Anothe…
A new method identifies a stable subnetwork in overparameterized student models.
problem Identifying a stable subnetwork in overparameterized student models.
method Spectral representation of linear transfer of information, focusing on eigenvalues and eigenvectors.
result A stable student substructure is isolated that mirrors the true complexity of the teacher.
This work studies fluctuation in multilayer neural networks using mean field theory.
problem Understanding fluctuation in multilayer neural networks with mean field training.
method Developed a second-order mean field limit to capture fluctuation, demonstrating stability of gradient descent training.
result Gradient descent training in multilayer networks biases towards minimal fluctuation, even after convergence.
NAST generalizes scattering transform for non-stationary time series analysis.
problem Analyzing non-stationary time series data.
method Neural activation of scattering transform with various activation functions and high pass filters.
result Central and non-central limit theorems for NAST of Gaussian processes.
Neuron Shapley identifies key neurons in deep networks, improving model accuracy and fairness.
problem Identifying responsible neurons in deep networks for better model performance and fairness.
method Neuron Shapley framework quantifies neuron contributions, accounting for interactions.
result Removing just 30 critical filters can destroy model accuracy, revealing network function.
New q-neurons improve neural network performance.
problem Improving neural network activation functions.
method Introducing q-neurons based on Jackson's q-derivatives with stochastic parameters. result Consistently improved performance over state-of-the-art activation functions.
Describes explaining neurons in deep representations using compositional logical concepts.
problem Interpreting neuron behavior in deep neural networks.
method Identifying compositional logical concepts that closely approximate neuron behavior.
result Compositional explanations provide insights into model performance and allow for adversarial example creation.
SeReNe prunes neurons with low sensitivity to reduce network size.
problem Large neural networks consume too many resources on resource-constrained devices.
method Exploits neural sensitivity as a regularizer to prune neurons with low sensitivity.
result Pruning neurons with low sensitivity achieves competitive compression ratios.
Under-parameterized networks can either copy or average teacher weights, leading to universal optimal solutions.
problem Approximating a teacher network with an under-parameterized student network.
method Analyzing shallow neural networks with erf activation function and unitary teacher weights, proving copy-average configurations are critical points and finding the optimal solution.
result The optimal solution for under-parameterized networks has a universal structure, whether copying or averaging teacher neurons.
This research investigates selectively pruning hyper and hypo neurons to improve neural network generalization.
problem Improving neural network generalization to unseen data.
method Investigates pruning hyper and hypo neurons selectively in fully connected layers of CNNs.
result Selective pruning of hyper and hypo neurons improves model performance on out-of-domain data.
Modeling hidden neurons in SNNs using mesoscopic approximations.
problem Underconstrained problem of modeling unobserved neurons in SNNs.
method Coarse-graining and mean-field approximations to derive neuLVM.
result neuLVM can efficiently model large SNNs and recover connectivity parameters.
Topological methods improve neuron analysis and tracer injection summary.
problem Traditional methods fail to capture the tree-like structure of neurons.
method Discrete Morse (DM) Theory for neuron skeletonization and consensus tree summarization.
result Significant performance improvements over non-topological methods.
Quadratic autoencoder improves low-dose CT image denoising.
problem Low-dose CT image denoising.
method Quadratic autoencoder architecture applied to CT denoising.
result Quadratic autoencoder achieves superior denoising performance and efficiency.
Solves internal covariate shift and dying neurons with linked neurons.
problem Internal covariate shift and dying neurons in deep learning.
method Defining linked neurons with two constraints: shared operating point and non-zero gradient.
result Linked neurons effectively solve internal covariate shift and improve training efficiency.
Neural network robustness evaluated in the presence of failing neurons.
problem Evaluating robustness of neural networks in the face of neuron failures.
method Leveraging Lipschitz continuity of activation functions, calculating Forward Error Propagation.
result Tight bounds on the number of failing neurons before network accuracy is compromised.
We developed a faster method for calculating neuron importance in neural networks.
problem Assigning importance to individual neurons in deep learning models.
method We developed Neuron Integrated Gradients, a scalable implementation of Total Conductance.
result Neuron Integrated Gradients is faster and empirically stronger than DeepLIFT.
BEAN models neuronal correlations to create interpretable representations.
problem Hard interpretation of dense-layer representations in DNNs.
method Inspired by neuroscience, BEAN models neuronal correlations and dependencies.
result BEAN enables formation of interpretable neuronal clusters without sacrificing model performance.
SpaRCe optimizes reservoir computing by learning neuron thresholds to improve performance and prevent forgetting.
problem Improving performance and preventing forgetting in reservoir computing networks.
method Integrates neuron-specific learnable thresholds to optimize sparsity without altering dynamics, learning read-out weights and thresholds via gradient rule.
result Threshold learning improves performance and alleviates catastrophic forgetting.
It will be shown that according to theorems of K. Menger, every neuron grid if identified with a curve is able to preserve the adopted qualitative structure of a data space. Furthermore, if this identification is made, the neuron grid structure can always be mapped to a subset of a universal neuron grid which is constr…
This work improves DNN interpretability by reducing neuron ambiguity.
problem Lack of interpretability in DNNs, especially in healthcare applications.
method Developed a metric to evaluate neuron consistency, used adversarial examples to identify ambiguous features, and proposed adversarial training to improve consistency.
result Reduced ambiguity of neurons in DNNs, improving interpretability.
A new method to understand neural networks by sampling the 'inverse set' of a neuron.
problem Understanding the internal representation of neurons in neural networks.
method Optimization-based sampling approach to characterize the input space that excites a neuron.
result Inspection of samples reveals regularities that help understand the neuron's representation.
Neural networks learn to mimic brain neurons with two-input activation functions, improving performance and robustness.
problem Training neural networks to mimic the complex interactions of brain neurons.
method Developed a network-in-network architecture with two-input activation functions, optimized hyperparameters, and compared to conventional ReLU networks.
result Two-input activation functions can learn soft XOR functions, improving network performance and robustness.
CHANI learns classification tasks with local transformations inspired by biology.
problem Proving neural networks can learn classification tasks with local transformations.
method CHANI uses spiking neurons modeled by Hawkes processes with expert aggregation for local learning.
result CHANI can learn and encode multiple classes, forming assemblies of neurons.
BinaryGAN uses binary neurons for GAN training.
problem Training GANs with binary outputs.
method End-to-end backpropagation with sigmoid-adjusted straight-through estimators.
result BinaryGAN can generate binary-valued predictions.
Researchers develop methods to learn neuron dynamics from colored noise.
problem Learning nonlocal stochastic neuron dynamics from colored noise.
method Proposed two methods for closing Fokker-Planck equations: nonlocal large-eddy-diffusivity closure and data-driven sparse regression.
result Mutual information and total correlation between stimulus and neuron states calculated for FHN neuron.
Single neuron with ADA learns XOR and outperforms other functions.
problem Classifying linearly non-separable data.
method Proposed a new artificial neuron with apical dendrite activation.
result ADA function achieves 100% accuracy on XOR and superior performance on benchmark datasets.
Despite our extensive knowledge of biophysical properties of neurons, there is no commonly accepted algorithmic theory of neuronal function. Here we explore the hypothesis that single-layer neuronal networks perform online symmetric nonnegative matrix factorization (SNMF) of the similarity matrix of the streamed data. …
Stable unactivated neurons reduce expressiveness in ReLU networks.
problem Reducing expressiveness in ReLU neural networks due to stably unactivated neurons.
method Investigated the probability of neurons being stably unactivated in ReLU networks with symmetric weight and bias distributions.
result Proved the probability of a neuron being stably unactivated in the second hidden layer of a ReLU network.
Novel chaotic neurons improve AI with minimal training data.
problem Limited training data for AI algorithms.
method Intrinsically chaotic neurons inspired by chaos theory.
result Classification accuracy up to 95.8% with just 2 training samples per class.
Single-spike neurons can approximate as well as multi-spike neurons.
problem Limitation of single-spike neurons in spiking neural networks.
method Comparison of single-spike and multi-spike neural networks.
result Single-spike and multi-spike neural networks are equivalent in approximation capabilities.
Model improves information transfer from visual streams.
problem Challenges in unsupervised learning from continuous visual data.
method Inspired by physics, maximizes mutual information through temporal process.
result Focus of attention enhances information transfer from input stream.
Optimal neuron activation functions improve neural network performance.
problem Limited expressive power of standard neuron activation functions in neural networks.
method Additive Gaussian process regression to construct individual neuron activation functions.
result Optimal neuron activation functions lead to better performance and reduced overfitting.
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.
Novel 'strong neuron' improves deep learning efficiency and robustness.
problem Improving deep learning efficiency and robustness against adversarial attacks.
method Introducing a novel 'strong neuron' model and a constructive training algorithm.
result Achieved 10x-100x reduction in operations count and hardware requirements.
Deep CNN model predicts neuronal cell health from images.
problem Predicting the biological activity of chemical compounds on neuronal cells.
method Deep convolutional neural network (CNN) with residual connections.
result Achieved 99.6% accuracy in distinguishing treated from untreated cells.
New artificial neuron doubles weight for improved deep learning accuracy.
problem Improving deep learning models' accuracy.
method Introducing a double-weight neuron, tested on MNIST and CIFAR-10 datasets.
result Significant improvement in classification accuracy for MNIST and CIFAR-10 datasets.
Optimizes deep neural networks using splines and adaptive knots.
problem Improving the optimization of deep neural networks.
method Integrates a second-order total-variation criterion to optimize activation functions, deriving a representer theorem.
result Optimal network configurations can be achieved with nonuniform linear splines with adaptive knots.
Single ReLU neuron's gradient dynamics reveal support vectors as key to generalization.
problem Understanding the generalization capability of ReLU networks.
method Examined gradient flow dynamics and support vectors in single ReLU neuron training.
result Support vectors play a crucial role in the generalization of ReLU networks.
Framework for explaining CNN predictions using input resampling.
problem Limited model interpretability in neural networks.
method Select neurons by two metrics over perturbed input images.
result Identifies neurons that influence and generalize network output.
A new RNN model based on coupled oscillators mitigates gradient issues.
problem Gradient vanishing and exploding issues in RNNs.
method Time-discretization of a system of second-order ODEs modeling coupled oscillators.
result The model maintains bounded gradients, leading to stable learning of long-term dependencies.
Novel method decorrelates neurons for better deep learning model generalization.
problem High correlations between neurons limit deep learning model generalization.
method Regularization terms from minimum spanning tree of neuron cliques, using correlation dissimilarities.
result Our regularizers outperform existing methods and minimize neuron redundancies.