Dynamic Sparse Training finds efficient sparse networks from scratch.
problem Finding efficient sparse neural networks.
method Jointly optimizes network parameters and sparsity with trainable thresholds.
result Achieves state-of-the-art performance with minimal performance loss.
New pruning methods improve dynamic sparse training performance.
problem Improving dynamic sparse training performance.
method Design and empirical analysis of pruning criteria.
result Most pruning methods yield similar results, but magnitude-based pruning performs best in low-density regimes.
ADAPT method controls balance between sparse generator and discriminator in GANs.
problem High computational complexity in GANs training.
method Balanced Dynamic Sparse Training (ADAPT) method.
result ADAPT method achieves a good trade-off between performance and computational cost.
FedDST trains sparse sub-networks to improve efficiency in federated learning.
problem Efficiently train large models on resource-limited edge devices with limited network bandwidth.
method Dynamic sparse training to reduce model size and communication.
result FedDST outperforms existing methods in federated learning, especially in non-i.i.d. settings.
Modern deep neural networks are typically highly overparameterized. Pruning techniques are able to remove a significant fraction of network parameters with little loss in accuracy. Recently, techniques based on dynamic reallocation of non-zero parameters have emerged, allowing direct training of sparse networks without…
Predictive Sparse Manifold Transform learns dynamic video sequences.
problem Learning and predicting natural dynamics in video sequences.
method Two-layer framework: sparse coding and manifold learning.
result PSMT with dynamic embedding space outperforms static baselines in future frame prediction.
Large SGD step sizes lead to sparse feature learning in neural networks.
problem Sparse feature learning in neural networks with large step sizes.
method Empirical observations and theoretical analysis of SGD dynamics.
result Large step sizes induce implicit regularization leading to sparse predictors.
Dynamic sparseness reduces neural network computation by selectively omitting parts of computations.
problem Reducing the computational and memory footprint of neural networks.
method Combining dynamic sparseness with block-wise matrix-vector multiplications to selectively omit parts of computations.
result The proposed method outperforms static sparseness and achieves similar perplexities at half the computational cost.
Dynamic model pruning improves performance on deep neural networks without retraining.
problem High memory and latency requirements for deep neural networks on low-end devices.
method Dynamic allocation of sparsity pattern and feedback signal to reactivate pruned weights.
result Sparse models achieve state-of-the-art performance with no additional retraining.
We propose to execute deep neural networks (DNNs) with dynamic and sparse graph (DSG) structure for compressive memory and accelerative execution during both training and inference. The great success of DNNs motivates the pursuing of lightweight models for the deployment onto embedded devices. However, most of the prev…
We investigate the difficulties of training sparse neural networks and make new observations about optimization dynamics and the energy landscape within the sparse regime. Recent work of \citep{Gale2019, Liu2018} has shown that sparse ResNet-50 architectures trained on ImageNet-2012 dataset converge to solutions that a…
Study shows how specialized attention circuits emerge during transformer training.
problem Understanding the mechanisms of transformer training dynamics at large scales.
method Controlled sparse modular addition task; monitoring token evolution via visual sandbox.
result Specialized attention circuits (clustering heads) naturally emerge during training.
Graph neural networks have become increasingly popular in recent years due to their ability to naturally encode relational input data and their ability to scale to large graphs by operating on a sparse representation of graph adjacency matrices. As we look to scale up these models using custom hardware, a natural assum…
A new method reduces feature size in CRFs for faster training.
problem Challenges in solving sparse CRFs for large-scale applications.
method Safe dynamic screening method exploiting dual optimum estimation.
result Significant speedup in training CRFs without loss of accuracy.
Efficiently constructs sparse ROMs for high-dimensional data using causation entropy.
problem Creating effective reduced-order models for high-dimensional dynamical data.
method Uses causation entropy to identify important terms and construct ROMs with varying sparsity.
result Demonstrates the effectiveness of causation entropy in constructing sparse ROMs for chaotic systems with skewed statistics.
A new distributed method speeds up sparse model training.
problem Efficiently training models with massive samples and high-dimensional features.
method Distributed Dynamic Safe Screening (DDSS) method for sparsity regularized models.
result Achieves linear convergence rate and eliminates almost all inactive features.
dLDS models neural dynamics as sparse combinations of simpler components.
problem Understanding complex neural dynamics at a population level.
method Proposes a decomposed dynamical system model trained through dictionary learning.
result Model efficiently captures and demix diverse neural dynamics.
New model leads to optimal test loss in sparse linear regression.
problem Sparse linear regression with low test loss despite interpolating training data.
method Developed a new parametrization of the model that combines benefits of ℓ1 and ℓ2 norms.
result Training via gradient descent leads to an interpolator with near-optimal test loss.
Proposes a flexible MGP model for dynamic, sparse correlations.
problem Handling dynamic and sparse correlations in multivariate data.
method Non-stationary MGP with dynamic spike-and-slab prior and EM algorithm.
result Captures dynamic and sparse correlations effectively.
DLNs dynamics change with variance, leading to saddle-to-saddle training phases.
problem Understanding the dynamics of DLNs with varying initialization variance.
method Analyzing the phase transition of DLNs' dynamics as variance changes.
result Gradient descent visits a sequence of saddles, reaching a sparse global minimum.
DistGP models multi-robot mapping with distributed Gaussian process learning.
problem Collaborative mapping by multiple robots with limited local data.
method Sparse Gaussian process with factorisation and distributed training via GBP.
result DistGP achieves superior accuracy and robustness compared to DiNNO.
Deep weight factorization improves neural network training through smooth optimization of sparse penalties.
problem Challenges in applying sparse regularization in neural networks due to non-differentiability of penalties.
method Introduces deep weight factorization, decomposing weights into multiple factors for smooth optimization of L 1 L_1 L 1 -penalized networks. result Deep weight factorization outperforms shallow factorization and pruning methods consistently across various architectures and datasets.
Dynamic models learn from sparse, interacting sub-systems.
problem Learning robust models for systems with local views and spatial locations.
method Abstracting the system as a collection of sparsely interacting sub-systems, each with a learned topology informed by spatial structure.
result Models are more robust to the number of available views and generalize better to novel tasks.
Deep neural network reconstructs traffic speeds from sparse vehicle data.
problem Reconstructing traffic speeds from limited probe vehicle data.
method Convolutional neural network architecture for spatio-temporal learning.
result The method can reconstruct traffic speeds with low probe vehicle penetration.
Improved trajectory prediction for team sports using sparse outputs.
problem Challenging to train deep learning models for player trajectory prediction.
method Sparse trajectory prediction and constant acceleration interpolation.
result Interpolation improves performance for all tested models.
New method finds sparse networks without labels, improving performance.
problem Sparse connectivity in neural networks to reduce memory and energy demands.
method Neural Tangent Transfer method to find sparse networks without labels.
result Sparse networks achieve higher classification performance and faster convergence.
Prototype for adaptive electron microscopy scans reduces dose and time.
problem Reduce electron microscopy scan time and dose with minimal loss.
method Adaptive partial scanning with reinforcement learning.
result Reinforcement learning trained neural network optimizes scan paths.
We present Distributed Equivalent Substitution (DES) training, a novel distributed training framework for large-scale recommender systems with dynamic sparse features. DES introduces fully synchronous training to large-scale recommendation system for the first time by reducing communication, thus making the training of…
UCB-TQL learns from multiple tasks with shared dynamics and adapts to task-specific variations.
problem Transfer reinforcement learning with composite MDPs where tasks share core dynamics but have sparse differences.
method UCB-TQL, a novel transfer RL algorithm for composite MDPs.
result Achieved a regret bound of i l d e O ( e H 5 N ) ilde{O}(\sqrt{eH^5N}) i l d e O ( e H 5 N ) that scales independently of the ambient dimension. Powerpropagation makes neural networks inherently sparse.
problem Training sparse neural networks to reduce computational footprint and model size.
method Introduces a new weight-parameterisation technique exploiting gradient descent dynamics.
result Models trained with Powerpropagation have a higher density of zero weights, allowing for more efficient pruning.
A hybrid method combines data assimilation and machine learning to predict chaotic dynamics from sparse noisy data.
problem Predicting chaotic dynamics from sparse and noisy observations.
method Iterative application of ensemble Kalman filter for data assimilation and neural network for model emulation.
result The hybrid method successfully predicts chaotic dynamics up to two Lyapunov times, retrieves positive Lyapunov exponents, and more energetic frequencies.
This study explains RL training dynamics in LLMs, focusing on token-level optimization and reasoning pattern reshaping.
problem Understanding the training dynamics of RL in LLMs to improve their reasoning capabilities.
method Empirical analysis and theoretical modeling of RL training process, focusing on reasoning patterns and token optimization.
result RL primarily optimizes a sparse subset of critical tokens, reshaping reasoning pattern distributions and affecting model performance.
Paper improves reinforcement learning in multi-scene tasks.
problem Reducing sample variance in multi-scene reinforcement learning.
method Sparse dynamic value estimation using Gaussian mixture models.
result Significant improvements in reward scores and navigation efficiency.
Deep learning predicts dynamics from sparse data.
problem Predicting spatiotemporal dynamics from sparse data.
method Spatially dimension-independent deep learning framework.
result Predicts dynamics from sparse data sites.
We address the issue of estimating the topology and dynamics of sparse linear dynamic networks in a hyperparameter-free setting. We propose a method to estimate the network dynamics in a computationally efficient and parameter tuning-free iterative framework known as SPICE (Sparse Iterative Covariance Estimation). The …
Early alignment in neural networks leads to sparse representations but hinders convergence.
problem The implicit bias of gradient descent during early training phases.
method Quantitative description of early alignment phase in small initialisation, one hidden layer networks.
result Early alignment induces a sparse representation but also hinders convergence to global minima.
New method learns dynamics from sparse data using geometric constraints.
problem Learning dynamics from sparse, undersampled data.
method Reformulates inference as a stochastic control problem, using geometry-driven path augmentation.
result Accurately recovers stochastic dynamics from extremely undersampled data.
FLASH-MAX predicts electromagnetic fields from sparse data in seconds.
problem Predicting homogeneous electromagnetic fields from sparse pointwise observations.
method Exact-by-construction neural network architecture that satisfies Maxwell's equations symbolically.
result FLASH-MAX achieves sub-1% relative validation error from 1K sparse observations in seconds.
New techniques improve the accuracy of identifying nonlinear systems from noisy data.
problem Identifying nonlinear dynamical systems from noisy state measurements.
method Comparative study of local and global smoothing techniques to denoise state measurements and improve sparse regression methods.
result Global smoothing methods outperform local methods in improving the accuracy of governing equation recovery.
We consider the problem of modeling multivariate time series with parsimonious dynamical models which can be represented as sparse dynamic Bayesian networks with few latent nodes. This structure translates into a sparse plus low rank model. In this paper, we propose a Gaussian regression approach to identify such a mod…
HierGP improves emulator efficiency for sparse, structured data.
problem Sparse, structured data in expensive simulations.
method Hierarchical shrinkage GP framework with cumulative shrinkage priors.
result HierGP identifies structured sparse features from limited data.
Parsimonious Dynamic Mode Decomposition selects sparse modes robustly.
problem Manual tuning of sparsity parameters in traditional DMD.
method Time-delay embedding and Orthogonal Matching Pursuit.
result Autonomously determines optimally sparse subset of modes.
LD-EnSF speeds up data assimilation with sparse observations.
problem Efficiently assimilate sparse and noisy data into complex dynamical systems.
method LD-EnSF uses latent dynamics networks and history-aware LSTM encoders to process sparse observations without full-space simulations.
result Achieves significant speedups over existing methods while maintaining high accuracy.
Exploration in environments with continuous control and sparse rewards remains a key challenge in reinforcement learning (RL). Recently, surprise has been used as an intrinsic reward that encourages systematic and efficient exploration. We introduce a new definition of surprise and its RL implementation named Variation…
Model shows loss curve with two distinct exponents due to sparse activations.
problem Sparse activations impact neural network scaling laws.
method Introduced a model for neural scaling laws under sparse activations, derived asymptotic population loss, and analyzed gradient-descent dynamics.
result Loss curve exhibits double-descent peak near interpolation threshold with two distinct scaling exponents.
Paper improves DOA estimation in sparse arrays using Siamese neural networks.
problem Challenges in DOA estimation with limited snapshots in sparse linear arrays.
method Introduces a Siamese neural network with a sparse augmentation layer for enhanced signal feature embedding.
result Demonstrates improved DOA estimation accuracy in sparse arrays.
Framework learns continuous dynamics from sparse trajectories.
problem Learning dynamics from sparsely sampled and high-dimensional trajectories.
method Interpolative Multi-Marginal Flow Matching (IMMFM) framework.
result IMMFM outperforms existing methods in forecasting and downstream tasks.
Improved SINDy autoencoder for identifying noisy dynamical systems.
problem Robust identification of noisy dynamical systems from data.
method Incorporates noise-separating neural network structures into SINDy autoencoder architecture.
result Accurately recovers latent dynamics and estimates measurement noise from noisy observations.