Paper proposes MABN to stabilize BN in small batch sizes.
problem Weakness of BN in small batch sizes limits its use in tasks with small batch sizes.
method Identifies and addresses two extra batch statistics in BN's backward propagation.
result MABN completely restores BN's performance in small batch sizes without additional nonlinear operations.
A new method for learning Bayesian neural networks using layerwise inference.
problem Learning Bayesian neural networks efficiently and accurately.
method Bayesian layerwise inference, treating neural networks as stacked Bayesian linear models, with pseudo-targets defined by backpropagated gradients.
result The method converges quickly and performs well on various benchmarks.
Variational inference has had great success in scaling approximate Bayesian inference to big data by exploiting mini-batch training. To date, however, this strategy has been most applicable to models of independent data. We propose an extension to state space models of time series data based on a novel generative model…
SGDM accelerates faster than SGD with large batch sizes and permits broader learning rates.
problem Understanding the role of momentum in SGDM and its convergence rates.
method Analysis of SGDM convergence rates under strongly convex settings, including finite-sample rates and asymptotic normality of the averaged estimator.
result SGDM converges faster than SGD with large batch sizes and permits broader learning rates.
STORM-PG uses momentum for faster policy gradient updates.
problem Improving policy gradient methods for reinforcement learning.
method Introduces STORM-PG, a SARAH-based algorithm with exponential moving average.
result Achieves O(1/ε3) sample complexity, matching best-known rate. Batch normalization makes deep residual networks train faster.
problem Training deep residual networks with large depths.
method Downscaling the residual branch by a normalizing factor early in training.
result Normalized residual blocks compute functions close to the identity function early in training.
This work provides a scaling rule for model EMA optimization across batch sizes.
problem Training dynamics and performance differences across batch sizes when using model EMA.
method Developed a scaling rule for model EMA optimization, demonstrating its validity across various architectures and data modalities.
result Enabled SSL methods like BYOL to train at larger batch sizes without performance degradation.
Use of an autoencoder (AE) as a normal model is a state-of-the-art technique for unsupervised-anomaly detection in sounds (ADS). The AE is trained to minimize the sample mean of the anomaly score of normal sounds in a mini-batch. One problem with this approach is that the anomaly score of rare-normal sounds becomes hig…
New method predicts spatio-temporal data with short and long-range dependence.
problem Uncertainty in predicting the distribution of mixed moving average fields.
method Theory-guided machine learning approach using generalized Bayesian algorithm.
result Fixed-time and any-time PAC Bayesian bounds for ensemble forecasts.
This work analyzes centered binary Restricted Boltzmann Machines (RBMs) and binary Deep Boltzmann Machines (DBMs), where centering is done by subtracting offset values from visible and hidden variables. We show analytically that (i) centering results in a different but equivalent parameterization for artificial neural …
This paper introduces sample-averaged Q-learning for better RL performance.
problem Improving reinforcement learning algorithms by managing uncertainty.
method Integrates statistical inference into Q-learning through sample averaging and functional central limit theorem.
result Establishes a unified theoretical foundation for sample-averaged Q-learning.
BEMA reduces bias in EMA, leading to faster convergence and better performance.
problem Stochasticity in language model fine-tuning destabilizes training.
method Bias-Corrected Exponential Moving Average (BEMA) augmentation of EMA.
result BEMA leads to significantly improved convergence rates and final performance.
Takagi-Sugeno-Kang (TSK) fuzzy systems are flexible and interpretable machine learning models; however, they may not be easily optimized when the data size is large, and/or the data dimensionality is high. This paper proposes a mini-batch gradient descent (MBGD) based algorithm to efficiently and effectively train TSK …
GPA improves LLM training speed by 8.71% for Llama-160M models.
problem Training Large Language Models (LLMs) with high memory overhead and slow convergence.
method Generalized Primal Averaging (GPA) extends Nesterov's method to eliminate memory-intensive two-loop structure.
result GPA achieves up to 10.13% speedup over AdamW in training Llama-1B model.
A new method normalizes activations to match batch normalization without batch dependence.
problem Performance degradation with batch-independent normalization techniques.
method Proxy-Normalizing Activations
result Proxy-Normalization technique emulates batch normalization's behavior and performance.
Explains the difference between EMA and moving EMA, focusing on market trend indicators.
problem Understanding the difference between exponential moving average and moving exponential average.
method Explains the mathematical tools and definitions of trend indicators.
result Discusses the properties of the MACD indicator and its use in market trend analysis.
Study shows gradient variance increases during deep learning training, contrary to common belief.
problem Understanding and minimizing gradient variance in deep learning models.
method Gradient Clustering method using stratified sampling to minimize gradient variance.
result Gradient variance increases during training, and smaller learning rates coincide with higher variance.
We show Vector Autoregressive Moving Average models with scalar Moving Average components could be estimated by generalized least square (GLS) for each fixed moving average polynomial. The conditional variance of the GLS model is the concentrated covariant matrix of the moving average process. Under GLS the likelihood …
New moving average adapts weight dynamically based on polynomial and wavefunction.
problem Lagging traditional moving averages in adjusting to changes in data.
method Develops a moving average with weight as a polynomial of a wavefunction from an eigenproblem.
result Immediate 'switch' without lag, adapting to changes in data.
CBN improves batch normalization for small mini-batch sizes.
problem Reduced effectiveness of Batch Normalization in small mini-batch sizes.
method CBN uses statistics from multiple recent iterations, compensating for network weight changes via Taylor polynomials.
result CBN outperforms original batch normalization and direct iteration statistics in object detection and image classification.
Batch normalization biases linear models towards uniform margins, improving performance in binary classification.
problem Understanding the implicit bias of batch normalization in linear models and neural networks.
method Analyzing gradient descent convergence on linear models and two-layer CNNs with batch normalization.
result Gradient descent with batch normalization in linear models converges to a uniform margin classifier with an exponential convergence rate.
We extend average edge order results to normal 3-pseudomanifolds.
problem Determining the average edge order of normal 3-pseudomanifolds.
method Extending previous results on 3-manifolds to 3-pseudomanifolds with singularities.
result For a normal 3-pseudomanifold K, μ0(K)≥730, with equality if and only if K is a specific triangulation of RP2. A key component of most neural network architectures is the use of normalization layers, such as Batch Normalization. Despite its common use and large utility in optimizing deep architectures, it has been challenging both to generically improve upon Batch Normalization and to understand the circumstances that lend them…
Online Normalization is a new technique for normalizing the hidden activations of a neural network. Like Batch Normalization, it normalizes the sample dimension. While Online Normalization does not use batches, it is as accurate as Batch Normalization. We resolve a theoretical limitation of Batch Normalization by intro…
Optimal algorithms for Riemannian optimization with reduced complexity.
problem Stochastic optimization on Riemannian manifolds with limited data.
method Zeroth-order Riemannian Averaging Stochastic Approximation algorithms using Riemannian moving-average estimators and novel geometric conditions.
result Achieves optimal sample complexities for generating approximate first-order stationary solutions.
New algorithms solve complex multi-level optimization problems with improved efficiency.
problem Smooth stochastic multi-level composition optimization problems.
method Two algorithms using moving-average and linearized stochastic estimates.
result Achieved sample complexities of O(1/ε^4) and O(1/ε^6).
TaskNorm improves meta-learning performance by rethinking batch normalization.
problem Challenges in batch normalization for meta-learning with deep networks.
method Developed TaskNorm, a novel approach to batch normalization for meta-learning.
result TaskNorm consistently improves meta-learning performance across various datasets and meta-learning approaches.
Batch Normalization (BN) uses mini-batch statistics to normalize the activations during training, introducing dependence between mini-batch elements. This dependency can hurt the performance if the mini-batch size is too small, or if the elements are correlated. Several alternatives, such as Batch Renormalization and G…
This study uses moving average cluster entropy to analyze financial market dynamics.
problem Understanding long-range dependence in financial markets.
method Moving average cluster entropy approach applied to ARFIMA and FBM processes.
result Long-range positive correlation in financial markets is linked to the cluster entropy behavior.
A new method normalizes flow mixtures for better inference across different data types.
problem Inference failure across diverse posterior geometries in normalizing flows.
method Introduces a two-stage framework with a stable global weighting mechanism based on sEMA.
result Achieves consistent NLL improvements and stable weight trajectories over baselines.
Training state-of-the-art, deep neural networks is computationally expensive. One way to reduce the training time is to normalize the activities of the neurons. A recently introduced technique called batch normalization uses the distribution of the summed input to a neuron over a mini-batch of training cases to compute…
We propose a simple but effective multi-source domain generalization technique based on deep neural networks by incorporating optimized normalization layers that are specific to individual domains. Our approach employs multiple normalization methods while learning separate affine parameters per domain. For each domain,…
In this work, we investigate Batch Normalization technique and propose its probabilistic interpretation. We propose a probabilistic model and show that Batch Normalization maximazes the lower bound of its marginalized log-likelihood. Then, according to the new probabilistic model, we design an algorithm which acts cons…
We define a C^1 distance between submanifolds of a riemannian manifold M and show that, if a compact submanifold N is not moved too much under the isometric action of a compact group G, there is a G-invariant submanifold C^1-close to N. The proof involves a procedure of averaging nearby submanifolds of riemannian manif…
The paper tackles batch policy learning in Markov Decision Processes, focusing on average reward maximization.
problem Maximizing long-term average reward in Markov Decision Processes with batch learning.
method Doubly robust estimator for average reward, optimization algorithm for optimal policy, finite-sample regret guarantee.
result The proposed method achieves semiparametric efficiency and provides a finite-sample regret guarantee.
Zero-shot anomaly detection method using batch normalization.
problem Adapting anomaly detectors to new normal data distributions without training data.
method Adaptive Centered Representations (ACR) with batch normalization.
result First zero-shot AD results for tabular data and image data.
Federated learning for Bayesian clustering of large datasets.
problem Bayesian model-based clustering of large-scale binary and categorical data.
method Federated variational inference with local merge and delete moves in parallel batches, followed by global merge moves.
result Empirical validation shows superior performance compared to existing algorithms.
Recurrent Neural Networks (RNNs) are powerful models for sequential data that have the potential to learn long-term dependencies. However, they are computationally expensive to train and difficult to parallelize. Recent work has shown that normalizing intermediate representations of neural networks can significantly im…
Normalization techniques such as Batch Normalization have been applied successfully for training deep neural networks. Yet, despite its apparent empirical benefits, the reasons behind the success of Batch Normalization are mostly hypothetical. We here aim to provide a more thorough theoretical understanding from a clas…
New streaming methods improve convergence rates for optimization problems.
problem Optimizing large-scale, sequential data problems.
method Time-varying mini-batches and Polyak-Ruppert averaging for gradient-based algorithms.
result Time-varying mini-batches and averaging achieve optimal convergence and variance reduction.
This research investigates if deep neural networks can be trained without batch normalization.
problem Training deep neural networks efficiently without batch normalization.
method Detailed study of batch normalization, comparison with other methods, and adaptation of training process.
result It is possible to train deep neural networks effectively without batch normalization.
We propose a novel unsupervised domain adaptation framework based on domain-specific batch normalization in deep neural networks. We aim to adapt to both domains by specializing batch normalization layers in convolutional neural networks while allowing them to share all other model parameters, which is realized by a tw…
As an indispensable component, Batch Normalization (BN) has successfully improved the training of deep neural networks (DNNs) with mini-batches, by normalizing the distribution of the internal representation for each hidden layer. However, the effectiveness of BN would diminish with scenario of micro-batch (e.g., less …
The paper defines normal forms for rational 3-tangles and shows a sequence of moves to transform one form to another.
problem Understanding and manipulating rational 3-tangles.
method Definition of normal forms and sequence of normal jump moves.
result There is a sequence of normal jump moves leading to equivalent normal forms of rational 3-tangles.
Develops methods for reliable inference on batched bandit data.
problem Need for reliable inference methods based on adaptively-collected data from bandit algorithms.
method Introduces Batched OLS (BOLS) estimator for reliable inference on bandit data.
result BOLS is asymptotically normal and robust to non-stationarity in the baseline reward.
Overfitting frequently occurs in deep learning. In this paper, we propose a novel regularization method called Drop-Activation to reduce overfitting and improve generalization. The key idea is to drop nonlinear activation functions by setting them to be identity functions randomly during training time. During testing, …
Develops a flexible batched experimentation framework for limited adaptivity.
problem Challenges of continual reallocation in bandit algorithms with delayed feedback.
method Computational framework leveraging Gaussian sequential experiment and dynamic programming.
result Improves statistical power over standard methods, even compared to Bayesian bandit algorithms.
New method trains deep ResNets without normalization, achieving state-of-the-art performance.
problem Training deep ResNets without normalization layers leads to instability and lower accuracy.
method Adaptive gradient clipping and Normalizer-Free ResNets design.
result Normalizer-Free ResNets achieve 86.5% top-1 accuracy on ImageNet, matching EfficientNet-B7.