PGD-trained models have a preferential direction in their gradients, which improves robustness.
problem Mathematical lack of clarity in the direction of preferential gradient alignment after adversarial training.
method Proposed a novel definition of preferential direction and evaluated it using a metric based on GANs.
result PGD-trained models have higher alignment with the proposed preferential direction than baseline models.
Direct optimization of binary latent VAEs achieves competitive results without sampling.
problem Training VAEs with discrete latent variables using standard methods is challenging.
method Applied evolutionary algorithms to directly optimize discrete latent distributions.
result Direct optimization is efficient and competitive in zero-shot learning.
MixUp explained as a special adversarial training technique.
problem Improving model robustness against adversarial attacks.
method Introducing directional adversarial training (DAT) and proving MixUp's equivalence to a specific subclass of DAT.
result The family of Untied MixUp schemes is equivalent to the entire class of DAT schemes and can improve upon MixUp.
DIP-FAT improves adversarial training by diversifying perturbations.
problem Adversarial examples fool deep neural networks, leading to overfitting and poor performance.
method DIP-FAT uses random directions to diversify perturbations in adversarial training.
result DIP-FAT reduces overfitting and improves clean data accuracy.
RNNs trained on head direction task mimic brain's compass and shifter neurons.
problem Modeling brain's head direction system using neural networks.
method Optimized recurrent neural networks trained on angular velocity integration.
result RNNs naturally emerge with compass and shifter neuron-like properties.
Predicts short-term futures contract direction using neural networks and order flow data.
problem Challenges in predicting short-term directional movement of futures contracts.
method Engineering features from technical analysis, order flow, and order-book data; training a Tabnet neural network.
result Achieved an accuracy of 0.601 in predicting directional change on the Silver Futures Contract.
Stochastic Gradient Descent (SGD) based training of neural networks with a large learning rate or a small batch-size typically ends in well-generalizing, flat regions of the weight space, as indicated by small eigenvalues of the Hessian of the training loss. However, the curvature along the SGD trajectory is poorly und…
Paper presents a neural network for estimating wavefronts in direction of arrival scenarios.
problem Estimating the number of wavefronts in direction of arrival scenarios.
method Cross-entropy trained multilayer neural network for online adaptation of antenna array imperfections.
result The method outperforms classical model order selection schemes in accuracy, especially at low signal-to-noise-ratios.
Photonic co-processor speeds up training of large neural networks.
problem Training large neural networks with backpropagation is inefficient and communication is a bottleneck.
method Direct Feedback Alignment (DFA) with a photonic accelerator.
result Photonic accelerator can compute random projections with trillions of parameters.
The study analyzes optimization trajectories in neural networks to reveal redundancy and redundancy-reducing strategies.
problem Understanding the directional structure and redundancy in neural network optimization.
method Introducing natural notions of complexity for optimization trajectories and analyzing their directional nature.
result Training only scalar batchnorm parameters can match the performance of training the entire network, indicating potential for hybrid optimization schemes.
Improved CNN training with BDFA reduces computational cost and improves accuracy.
problem Low training performance of DFA in CNN.
method Combining DFA with BP, introducing feedback weight initialization, and proposing BDFA.
result BDFA shows better performance than conventional BP, especially in small datasets.
Simulation of high-speed train aerodynamics using RANS and machine learning.
problem Aerodynamic analysis of high-speed trains under turbulent flow conditions.
method RANS equations with turbulence model, machine learning (GEP, GPR, RF) for predictions.
result Random Forest (RF) provides the most accurate predictions for aerodynamic coefficients.
A new pruning method reduces neural network computation without retraining.
problem Efficiently reduce neural network computation while maintaining accuracy.
method Structured directional pruning via perturbation orthogonal projection.
result Achieves state-of-the-art pruned accuracy without retraining.
LayerNorm transformers have dead directions that can be read from their parameters alone.
problem Locating dead directions in LayerNorm transformers
method Using the inverse-scale direction of LayerNorm affine parameters
result Predicted dead direction matches measured bottom singular direction
Improves deep transfer learning by preventing performance degradation.
problem Deep transfer learning can degrade performance when using inappropriate pre-trained weights.
method Proposes a novel strategy to compute new descent directions that preserve regularization effects.
result DTNH strategy improves performance of deep transfer learning tasks by 0.1%--7%.
A new pruning method finds sparse minimizers in flat regions of deep neural networks.
problem Finding sparse minimizers in flat regions of deep neural networks.
method Directional pruning using an ℓ1 proximal gradient algorithm. result Empirical results show promising sparsity (92%) and comparable minima to SGD.
New methods validate a hypothesis explaining how neural nets generalize well.
problem Why over-parameterized nets generalize well despite memorizing training data.
method Developed new algorithms to suppress weak gradient directions without per-example gradients.
result Validated a hypothesis about gradient directions and their role in generalization.
New methods improve neural directed link prediction across all sub-tasks.
problem Directed link prediction requires handling edge directionality and bidirectionality, not just edge existence.
method Proposes three strategies: Multi-Class Framework, Multi-Objective, and Scalarized approaches.
result Improved performance across all three sub-tasks of directed link prediction.
New method prevents deep learning models from forgetting past tasks.
problem Catastrophic forgetting in continual learning.
method Direction-constrained optimization (DCO) with autoencoders.
result Model performance is improved without forgetting past tasks.
Forward gradients improve neural network training without backpropagation issues.
problem Training neural networks without backpropagation's locking and memorization problems.
method Using directional derivatives in forward differentiation mode, with biased guesses based on feedback from small auxiliary networks.
result Using gradients from a local loss as a candidate direction improves Forward Gradient methods.
Early training of deep neural networks leads to small, directionally converging weights.
problem Training dynamics of deep homogeneous neural networks with small initializations.
method Gradient flow analysis and study of KKT points for neural correlation function.
result Weights converge in direction to KKT points during early training stages.
Develops a new neural spike train decoding framework using topological data.
problem Decoding neural spike trains from head direction and grid cells.
method Combines simplicial complex discovery with deep learning to capture higher-order connectivity.
result Demonstrates effectiveness on head direction and trajectory prediction datasets.
Gradient descent-based adversarial training converges to robust classifiers on linearly separable data.
problem Understanding the inductive bias of adversarial training for robustness.
method Gradient descent on binary classification tasks with linearly separable data, focusing on inductive bias and convergence rates.
result Gradient descent-based adversarial training converges to the maximum margin classifier at a faster rate than clean data training.
Paper finds conditions for benign overfitting in neural networks.
problem Benign overfitting in leaky ReLU two-layer neural networks.
method Established directional convergence and classification error bounds.
result Benign overfitting occurs with high probability on mixture data.
Improves information cascade models using contrastive training and DSTs.
problem Improving models of information cascades using limited labeled data.
method Proposes a contrastive training procedure for models of information cascades as directed spanning trees (DSTs).
result Unsupervised training with additional content features achieves significantly better results, reaching half the accuracy of a fully supervised model.
Proposes WPGD for cost-sensitive adversarial training in deep learning.
problem Balancing robustness and accuracy in adversarial training.
method WPGD solves an optimal transport problem on the output space of the network.
result WPGD provides finer control over robustness-accuracy trade-off.
Deep RL policies are vulnerable to adversarial perturbations, but vanilla training yields more robust policies.
problem Vulnerability of deep reinforcement learning policies to adversarial perturbations.
method Analysis of deep reinforcement learning policy landscape and comparison of vanilla vs. adversarial training.
result Vanilla training yields more robust policies compared to adversarial training.
New method accelerates neural network training by focusing on flat directions.
problem Improving neural network training speed and stability.
method Bulk-SGD, interpolated gradient methods.
result Updates along the Dominant subspace can accelerate convergence but compromise stability.
This work improves policy-based training by proposing an evaluation balance objective for GFlowNets.
problem Reliable estimation of policy divergence under directed acyclic graphs remains challenging.
method Proposes an evaluation balance objective over partial episodes to measure policy divergence and improve policy-based training reliability.
result Evaluation balance strengthens policy-based training reliability and broadens its flexibility.
Modified training direction reduces generalization error in neural networks.
problem Reducing generalization error in neural networks.
method Theoretical analysis of modified natural gradient descent in function space.
result Modifying training direction in function space reduces total generalization error.
DoCoFL compresses model updates for cross-device federated learning.
problem Downlink compression for cross-device federated learning where clients may appear only once.
method Proposes DoCoFL framework for downlink compression in cross-device federated learning.
result Significant bi-directional bandwidth reduction with competitive accuracy.
Adaptive-SGD method optimizes machine learning training with dynamic batch and step sizes.
problem Optimizing machine learning training with adaptive batch and step sizes.
method Adaptive-SGD method that dynamically adjusts batch size and step size based on local curvature and probability of descent directions.
result Adaptive-SGD achieves global linear convergence on self-concordant functions and compares favorably to fine-tuned methods.
We explore a new research direction in Bayesian variational inference with discrete latent variable priors where we exploit Kronecker matrix algebra for efficient and exact computations of the evidence lower bound (ELBO). The proposed "DIRECT" approach has several advantages over its predecessors; (i) it can exactly co…
Despite their ability to memorize large datasets, deep neural networks often achieve good generalization performance. However, the differences between the learned solutions of networks which generalize and those which do not remain unclear. Additionally, the tuning properties of single directions (defined as the activa…
Federated learning trains models on remote devices without sharing data.
problem Training models on remote devices with limited data.
method Develops new methods for distributed optimization and privacy.
result New challenges and approaches for large-scale machine learning.
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.
GOLS finds activation functions affect training robustness, especially ReLU.
problem Investigate how different activation functions impact GOLS in neural network training.
method Identify SNN-GPPs for GOLS, analyze activation function effects on gradient continuity.
result GOLS robust for most activation functions but sensitive to ReLU.
A new method optimizes slicing directions for SW distances to improve high-dimensional probability measure comparison.
problem Challenging identification of informative slicing directions for SW distances.
method Constrained learning approach to optimize slicing directions, using continuous relaxations and gradient-based primal-dual approach.
result Demonstrated efficacy in learning more informative slicing directions on various high-dimensional data.
Proposes a neural network model to improve predictions in biased datasets.
problem Reduces bias in predictions from datasets with selection bias.
method Integrates meta information about bias direction and quantity into a neural network model.
result Improves prediction accuracy in electoral polls with biased training data.
Direct Feedback Alignment performs well on diverse deep learning tasks and architectures.
problem The limitations of backpropagation in parallelizing and scaling to modern deep learning tasks.
method Direct Feedback Alignment approach applied to neural view synthesis, recommender systems, geometric learning, and natural language processing.
result Direct Feedback Alignment successfully trains a wide range of state-of-the-art deep learning architectures with performance close to fine-tuned backpropagation.
Equity-Directed Bootstrapping improves model performance across groups in imbalanced datasets.
problem Improving model performance across different groups in imbalanced datasets.
method Equity-Directed Bootstrapping to balance training data with respect to both labels and group identity.
result The equity-directed bootstrap brings test set sensitivities and specificities closer to satisfying the equal odds criterion.
Bayesian optimization framework for hyperparameter tuning with directional derivatives.
problem Efficient hyperparameter tuning for machine learning models.
method Bayesian optimization with directional derivatives to seek more complex models.
result Demonstrated improved performance on various machine learning tasks.
Machine learning predicts critical points for directed percolation models.
problem Determining critical points for directed percolation models.
method Supervised and unsupervised machine learning algorithms (CNN and DBSCAN) were used.
result Machine learning accurately predicts critical points for both models.
This work trains a model to predict human driving directions from road scenes.
problem Defining implicit rules of human behavior for autonomous vehicles.
method Self-supervised learning of probabilistic network model.
result Model successfully generalizes to new road scenes.
CProp scales learning rate based on past gradient direction.
problem Adaptive learning rate scheduling during training.
method Gradient scaling method based on past gradient conformity.
result Significant gain in training speed on SGD and adaptive methods.
DCGANs generate drainage networks quickly from samples.
problem High computational costs in generating large numbers of drainage networks.
method DCGANs trained with connectivity-informed directional information.
result Connectivity-informed DCGANs outperform other methods in reproducing accurate drainage networks.
Study shows Direct Feedback Alignment fails to offer more efficient scaling than backpropagation.
problem Understanding and optimizing training methods for neural networks.
method Use of scaling laws to compare Direct Feedback Alignment (DFA) and backpropagation.
result DFA fails to offer more efficient scaling than backpropagation.
In this paper, we explore and detail our experiments in a high-dimensionality, multi-class image classification problem often found in the automatic recognition of Sign Languages. Here, our efforts are directed towards comparing the characteristics, advantages and drawbacks of creating and training Support Vector Machi…