Real-time pruning during training reduces network size and training time.
problem Efficiently reducing neural network size without sacrificing accuracy.
method Activation density-based pruning during training.
result Up to 200x reduction in parameters and 60x reduction in inference compute operations.
WeldNet reduces complex dynamics to simpler, manageable segments.
problem Complex, high-dimensional time-dependent datasets from physical processes are costly to simulate.
method Windowed Encoders for Learning Dynamics, splitting time domain into windows for nonlinear dimension reduction and propagator training.
result WeldNet captures nonlinear latent structures and dynamics, outperforming existing methods.
Improved clustering speed for 20 clusters on CIFAR-100 dataset.
problem Training time complexity for VAEs with discrete latent variables is linear in the number of clusters.
method Applied a continuous relaxation to discrete variables in Gaussian Mixture VAE, reducing training time complexity to constant.
result Reduced training time from 47 hours to 6 hours for 20 clusters on CIFAR-100.
Testing the implementation of deep learning systems and their training routines is crucial to maintain a reliable code base. Modern software development employs processes, such as Continuous Integration, in which changes to the software are frequently integrated and tested. However, testing the training routines requir…
Paper proposes a method to estimate variance reduction in DNN training using importance sampling.
problem Challenges in assessing variance reduction during DNN training using importance sampling.
method Proposes a method for estimating variance reduction using minibatches sampled under importance sampling.
result Demonstrates consistent reduction in variance, improved training efficiency, and enhanced model accuracy.
This work simplifies Gaussian process regression for multiple outputs.
problem Exponential computational complexity in Gaussian process regression.
method Approximating the covariance kernel using eigenvalues and functions.
result Significant reduction in training and regression complexity.
Communication on heterogeneous edge networks is a fundamental bottleneck in Federated Learning (FL), restricting both model capacity and user participation. To address this issue, we introduce two novel strategies to reduce communication costs: (1) the use of lossy compression on the global model sent server-to-client;…
RSAC improves lightweight continuous learning efficiency.
problem Excessive training time and memory usage in continuous learning.
method Regularized subspace approximation classifier with feature reduction and regularization.
result RSAC achieves more efficient continuous learning than prior methods.
K-Nearest neighbor classifier (k-NNC) is simple to use and has little design time like finding k values in k-nearest neighbor classifier, hence these are suitable to work with dynamically varying data-sets. There exists some fundamental improvements over the basic k-NNC, like weighted k-nearest neighbors classifier (wh…
Improved time complexity for parallel stochastic optimization in heterogeneous systems.
problem Time complexity in parallel stochastic optimization for large-scale machine learning models.
method Proposes Rennala MVR, a variance-reduced extension of Rennala SGD based on momentum-based variance reduction.
result Variance reduction improves time complexity in relevant parameter regimes for parallel stochastic optimization in heterogeneous systems.
The diverse world of machine learning applications has given rise to a plethora of algorithms and optimization methods, finely tuned to the specific regression or classification task at hand. We reduce the complexity of algorithm design for machine learning by reductions: we develop reductions that take a method develo…
Automates PDE model reduction with time-scale separation.
problem Computational expense in solving high-dimensional PDEs.
method Combines autoencoder and time-continuous model for latent dynamics.
result Automatically learns independent temporal scales in complex systems.
Balanced Neural ODEs combine VAEs and Neural ODEs for efficient time series modeling.
problem Efficiently modeling systems with time-varying inputs and varying complexity.
method Combines VAEs for dimensionality reduction and Neural ODEs for dynamics, using variational parameters to adaptively learn.
result Balanced Neural ODEs (B-NODE) efficiently approximate Koopman operator without predefined dimensionality.
Graphical heuristic reduces and partitions large datasets for faster supervised training.
problem Training large datasets for classification tasks.
method Clustering and information graph construction for dataset reduction and partitioning.
result Significant speed-up in training run-time without compromising prediction accuracy.
Enhances supervised visualization for unseen data using autoencoders and random forest.
problem Lack of generalization to unseen test sets in supervised dimensionality reduction.
method Combines autoencoder and random forest proximities for out-of-sample extension.
result 40% reduction in training time with 10% of training data, achieving consistent quality.
SWAP uses large mini-batches to train DNNs faster with good generalization.
problem Training deep neural networks with small mini-batches is time-consuming.
method SWAP computes an approximate solution with large mini-batches and refines it by averaging weights of multiple parallel models.
result SWAP trains models as well as small-batch training but in significantly less time.
The well-known Mori-Zwanzig theory tells us that model reduction leads to memory effect. For a long time, modeling the memory effect accurately and efficiently has been an important but nearly impossible task in developing a good reduced model. In this work, we explore a natural analogy between recurrent neural network…
A new neural network reduces high-dimensional time-series data for faster classification.
problem Classifying high-dimensional time-series patterns efficiently.
method Developed a time-series discriminant component network (TSDCN) using TSDCA for dimensionality reduction and classification.
result The TSDCN achieves high-accuracy classification and reduces training time.
MARS optimizes large model training by reducing variance, outperforming AdamW.
problem Training large models efficiently and scalably.
method Unified optimization framework MARS combining preconditioned gradient updates and variance reduction.
result MARS outperforms AdamW in training GPT-2 models.
FasTrCaps reduces CapsNet training time by 58.6% while maintaining accuracy.
problem The high training time of Capsule Networks (CapsNets).
method Integrating multiple lightweight optimizations and a novel learning rate policy (WarmAdaBatch) into FasTrCaps framework.
result FasTrCaps framework can reduce CapsNet training time by 58.6% while preserving accuracy.
Paper explores reducing precision in SVM for faster text classification.
problem Efficiency in multi-class text classification training.
method Comparison of SVM trained with reduced precision (16-bit, half) vs original.
result Reduced precision training maintains text classification accuracy.
The paper extends Marsden-Weinstein reduction to mechanical presymplectic structures for time-dependent Hamiltonian systems.
problem Limitations of Marsden-Weinstein reduction for cosymplectic structures in time-dependent Hamiltonian systems.
method Developed Marsden-Weinstein reduction for mechanical presymplectic structures.
result Mechanical presymplectic structures provide a more suitable framework for time-dependent Hamiltonian systems than cosymplectic structures.
Paper speeds up large foundation models for time series data.
problem Resource-intensive foundation models limit accessibility.
method Dimensionality reduction techniques, including PCA and neural network adapters.
result Up to 10x speedup and 4.5x more datasets fit on a single GPU.
We present a theoretical analysis and empirical evaluations of a novel set of techniques for computational cost reduction of classifiers that are based on learned transform and soft-threshold. By modifying optimization procedures for dictionary and classifier training, as well as the resulting dictionary entries, our t…
The study analyzes sharpness dynamics in neural networks, revealing mechanisms and conditions.
problem Understanding sharpness in neural network training.
method Fixed point analysis and edge of stability analysis in a simplified 2-layer linear network.
result Reveals mechanisms behind sharpness trends, conditions for edge of stability, and a period-doubling route to chaos.
RMDA trains structured neural networks with regularization and variance reduction.
problem Training structured neural networks with desired properties.
method RMDA algorithm for structured NNs with regularization and variance reduction.
result RMDA achieves desired structures identical to regularizer's at stationary points.
Paper develops an efficient approach to reduce HPO time.
problem Challenges in determining optimal hyperparameters due to large number and training time.
method Nested Latin hypercube design for initialization, truncated additive Gaussian process model for calibration, sequential model-based algorithm for optimization.
result Demonstrates competitive performance on various machine learning models.
New method reduces PDE surrogate model training costs by selectively acquiring time steps.
problem High computational cost of generating training data for PDE surrogate models.
method STAP (Selective Time-Step Acquisition for PDEs) framework that acquires only important time steps.
result Demonstrated effectiveness on several benchmark PDEs, reducing training costs.
In the covariate shift learning scenario, the training and test covariate distributions differ, so that a predictor's average loss over the training and test distributions also differ. In this work, we explore the potential of extreme dimension reduction, i.e. to very low dimensions, in improving the performance of imp…
We present local discriminative Gaussian (LDG) dimensionality reduction, a supervised dimensionality reduction technique for classification. The LDG objective function is an approximation to the leave-one-out training error of a local quadratic discriminant analysis classifier, and thus acts locally to each training po…
Predict training time without training using a linearized model.
problem Predicting the training time of a deep network without actual training.
method Approximate training dynamics with a linearized model and solve a low-dimensional SDE in function space.
result Predict training time within 20% error margin and 30-45x reduction in cost.
Paper compares Lagrangian reduction methods for rigid body systems.
problem Modeling and reduction of rigid body systems with rotors.
method Euler-Poincaré reduction by the whole group and reduction by stages.
result Equivalence of equations and conservation laws are tracked.
To train good supervised and semi-supervised object classifiers, it is critical that we not waste the time of the human experts who are providing the training labels. Existing active learning strategies can have uneven performance, being efficient on some datasets but wasteful on others, or inconsistent just between ru…
The Long-Short-Term-Memory Recurrent Neural Networks (LSTM RNNs) are a popular class of machine learning models for analyzing sequential data. Their training on modern GPUs, however, is limited by the GPU memory capacity. Our profiling results of the LSTM RNN-based Neural Machine Translation (NMT) model reveal that fea…
Enhances robustness of AT frameworks to multiple perturbations without increasing training complexity.
problem Defending against the union of multiple perturbations in adversarial training.
method SNAP technique that augments a network with shaped noise to enhance robustness.
result 14%-to-20% improvement in adversarial accuracy for ResNet-18 on CIFAR-10.
A new framework learns clustering and dimensionality reduction together.
problem Challenges in clustering high-dimensional data.
method Gradient-based manifold optimization for joint learning.
result Better performance compared to existing clustering algorithms.
Despite their successes in the field of self-learning AI, Convolutional Neural Networks (CNNs) suffer from having too many trainable parameters, impacting computational performance. Several approaches have been proposed to reduce the number of parameters in the visual domain, the Inception architecture [Szegedy et al.,…
LIIT uses large deviations to train neural networks faster.
problem Training neural networks requires large amounts of data, leading to inefficiencies.
method LIIT uses large deviations principle to generate and update training samples.
result LIIT achieves similar classification performance with smaller, more representative samples.
Spatiotemporal sequence prediction is an important problem in deep learning. We study next-frame(s) video prediction using a deep-learning-based predictive coding framework that uses convolutional, long short-term memory (convLSTM) modules. We introduce a novel reduced-gate convolutional LSTM(rgcLSTM) architecture that…
Proposes a federated learning approach for industrial asset failure prediction.
problem Lack of data and privacy concerns in industrial prognostics.
method Two-stage federated learning: dimension reduction and parameter estimation.
result Validated the approach using simulated and real data.
We propose a novel technique for faster deep neural network training which systematically applies sample-based approximation to the constituent tensor operations, i.e., matrix multiplications and convolutions. We introduce new sampling techniques, study their theoretical properties, and prove that they provide the same…
TTPUDR uses tensor-train decomposition for high-dimensional data analysis.
problem High-dimensional data analysis challenges.
method Tensor-train decomposition and manifold optimization.
result TTPUDR significantly outperforms past methods and state-of-the-art methods.
New method finds efficient low-rank neural networks during training.
problem High memory and computational demands of neural networks.
method Restricts weight matrices to a low-rank manifold and updates low-rank factors.
result Significantly reduced time and memory resources required for training and evaluation.
Study reduces redundant information in multi-modal datasets.
problem Redundant information increases data size and costs.
method Multimodal framework using entropy, mutual info, Lasso, SHAP, PCA, topic modeling.
result Significant feature space can be pruned with minimal loss in performance.
New optimizer MARS-M combines variance reduction with Muon for faster LLM training.
problem Training large-scale neural networks efficiently.
method Integrates MARS variance reduction with Muon optimizer.
result MARS-M converges to a first-order stationary point at a rate of ildeO(T−1/3). Pruning improves model generalization in over-parameterized models, contradicting traditional theories.
problem Pruning's effect on generalization in over-parameterized models.
method Empirical study on standard pruning algorithms and additional regularization effects.
result Pruning leads to better training and regularization, improving generalization.
Proposes faster neural network learning by using subsets of training data.
problem Manual design and validation of network topologies is time-consuming.
method Exploits subsets of training data at each incremental training step and performs online hyperparameter selection.
result Significantly reduces overall training time while maintaining performance.
How can we train a statistical mixture model on a massive data set? In this work we show how to construct coresets for mixtures of Gaussians. A coreset is a weighted subset of the data, which guarantees that models fitting the coreset also provide a good fit for the original data set. We show that, perhaps surprisingly…