Post-training quantization method using multiple low-precision points achieves higher precision for critical weights.
problem Discretizing pre-trained deep neural networks without re-training.
method Multipoint quantization with efficient greedy selection and adaptive point number.
result Outperforms state-of-the-art methods on ImageNet classification and PASCAL VOC object detection.
HMQ improves quantization for edge devices with mixed precision.
problem Efficient quantization for edge devices with uniform, power-of-two thresholds.
method Introduces HMQ, a mixed precision quantization block that repurposes Gumbel-Softmax for searching over quantization schemes.
result Achieves competitive and state-of-the-art results on ImageNet despite restrictions.
Adaptive loss scaling speeds up and improves deep learning training.
problem Numerical underflow in mixed precision training.
method Adaptive loss scaling that automatically computes layer-wise loss scale values during training.
result Adaptive loss scaling leads to shorter convergence time and improved accuracy.
Efficient deep neural network (DNN) inference on mobile or embedded devices typically involves quantization of the network parameters and activations. In particular, mixed precision networks achieve better performance than networks with homogeneous bitwidth for the same size constraint. Since choosing the optimal bitwi…
Low-precision DNNs have been extensively explored in order to reduce the size of DNN models for edge devices. Recently, the posit numerical format has shown promise for DNN data representation and compute with ultra-low precision in [5..8]-bits. However, previous studies were limited to studying posit for DNN inference…
A critical part of multi-person multi-camera tracking is person re-identification (re-ID) algorithm, which recognizes and retains identities of all detected unknown people throughout the video stream. Many re-ID algorithms today exemplify state of the art results, but not much work has been done to explore the deployme…
SINGD improves KFAC for memory-efficiency and stability in low-precision training.
problem Memory inefficiency and numerical instability of KFAC in low-precision training.
method Formulated inverse-free KFAC update and imposed structures in Kronecker factors.
result SINGD is memory-efficient and numerically robust, often outperforming AdamW in half precision.
Efficient Bitwidth Search optimizes neural network quantization for better performance.
problem Finding optimal bitwidth for weights and activations of each layer efficiently.
method EBS algorithm reusing meta weights and binary decomposition for efficient mixed precision convolution.
result Mixed precision QNN outperforms uniform bitwidth and other techniques on CIFAR10 and ImageNet.
Mixed-precision CA-SGD for generalized linear models on GPUs
problem SGD communication bottleneck
method Mixed-precision CA-SGD
result Matches FP32 SGD loss within 0.5% on various problems
Deep neural networks have enabled progress in a wide variety of applications. Growing the size of the neural network typically results in improved accuracy. As model sizes grow, the memory and compute requirements for training these models also increases. We introduce a technique to train deep neural networks using hal…
Reduced precision computation for deep neural networks is one of the key areas addressing the widening compute gap driven by an exponential growth in model size. In recent years, deep learning training has largely migrated to 16-bit precision, with significant gains in performance and energy efficiency. However, attemp…
Unified formula for training dynamics of linear networks combining lazy and balanced regimes.
problem Training dynamics of linear networks in two distinct setups: lazy and balanced/active.
method Unified formula for the evolution of the learned matrix, combining lazy and balanced regimes.
result Unified formula allows for rapid convergence and low rank bias, proving a complete phase diagram.
MixMin finds optimal data mixtures for better model performance.
problem Finding the best data mixtures for improved model performance is challenging.
method Developed a gradient-based approach for optimizing a convex bi-level objective.
result MixMin mixtures uniformly improved model performance across various tasks.
Synchronized stochastic gradient descent (SGD) optimizers with data parallelism are widely used in training large-scale deep neural networks. Although using larger mini-batch sizes can improve the system scalability by reducing the communication-to-computation ratio, it may hurt the generalization ability of the models…
For a data holder, such as a hospital or a government entity, who has a privately held collection of personal data, in which the revealing and/or processing of the personal identifiable data is restricted and prohibited by law. Then, "how can we ensure the data holder does conceal the identity of each individual in the…
CogDL simplifies graph deep learning experiments and benchmarks.
problem Challenges in training and evaluating graph neural networks.
method Unified design for training and evaluation, mixed precision training, efficient sparse operators, ease of use.
result CogDL is the most competitive graph library for efficiency and ease of use.
This paper presents the first comprehensive empirical study demonstrating the efficacy of the Brain Floating Point (BFLOAT16) half-precision format for Deep Learning training across image classification, speech recognition, language modeling, generative networks and industrial recommendation systems. BFLOAT16 is attrac…
CoDeQ simplifies joint model compression by integrating pruning and quantization.
problem Joint pruning and quantization methods are complex and require additional procedures.
method CoDeQ uses a dead-zone quantizer to directly induce sparsity and learn quantization parameters.
result CoDeQ achieves high sparsity and low-precision accuracy with minimal bit operations.
We improve Gaussian copula models for imputing mixed data types with precise approximations.
problem Imputing missing values with mixed data types in surveys and medical applications.
method We use randomized quasi-Monte Carlo procedures for direct and arbitrarily precise approximations of model estimation and imputation.
result Our method yields lower errors for model parameters and imputed values compared to existing methods.
Three RFF-based methods for nonlinear causal discovery in mixed data.
problem Nonlinear causal discovery in mixed data with computational constraints.
method FFML, TRFF, and FFCI methods for score-based, constraint-based, and hybrid causal discovery.
result FFML and TRFF methods provide complementary performance in causal discovery.
New techniques improve 16-bit training accuracy without 32-bit units.
problem Training deep learning models with only 16-bit floating-point units.
method Studied BFloat16 units and applied stochastic rounding and Kahan summation techniques.
result Up to 7% absolute validation accuracy gain in 16-bit-FPU training.
PAIN network improves imputation for mixed datasets.
problem Missing data in diverse scientific domains.
method Dynamic adaptive imputation using statistical methods, random forests, and autoencoders.
result PAIN outperforms traditional imputation methods in preserving data distributions.
The study analyzes numerical stability in large language models using mixed-precision arithmetic.
problem Numerical stability of large language models using low-precision arithmetic.
method Developed a mixed-precision analysis of transformer inference, deriving bounds for condition numbers and forward error.
result Established that numerical stability is determined by the interplay between weight magnitude and the growth of the residual stream.
Spectral methods improve signal recovery in mixed GLMs with precise asymptotics.
problem Estimating multiple signals from unlabeled observations in mixed GLMs.
method Developed exact asymptotics for spectral methods in a proportional regime.
result Optimized spectral method combined with a linear estimator minimizes estimation error.
Optimal SD improves ridge regression performance strictly and precisely.
problem Improving ridge regression performance through self-distillation.
method Analyzes unconstrained SD for ridge regression, deriving optimal mixing weight and asymptotic risk.
result Optimal SD strictly improves ridge regression performance, with exact risk equivalents derived.
Bayesian Bits unifies quantization and pruning through gradient optimization.
problem Joint mixed precision quantization and pruning for efficient neural networks.
method Gradient-based optimization with a novel bit width decomposition and learnable stochastic gates.
result Bayesian Bits achieves better accuracy vs. efficiency trade-off compared to static bit width networks.
Improved deep learning model deployment on tiny MCUs with mixed-precision quantization.
problem Memory limitations prevent accurate deployment of DNN models on tiny MCUs.
method Automated mixed-precision quantization using Reinforcement Learning for MCU constraints.
result Mixed-precision models achieve high accuracy with uniform quantization policies.
Generalized Precision Matrix for scalable estimation of nonparametric Markov networks.
problem Estimating conditional independence structure in general distributions for all data types.
method Generalized Precision Matrix (GPM) for mixed-type variables, regularized score matching framework for scalability.
result Validated theoretical results and demonstrated scalability in various settings.
The paper explores knotoids, pseudo knotoids, braidoids, and pseudo braidoids on the torus.
problem The study of knotoids, pseudo knotoids, braidoids, and pseudo braidoids on the torus.
method Introducing new knotoid and braidoid concepts, isotopy theorems, state sum formulas, and Alexander and Markov theorems.
result Formulation and proof of Alexander and Markov theorems for mixed knotoids and mixed pseudo knotoids.
We study the regular conditional law of mixed Gaussian Volterra processes under the influence of model disturbances. More precisely, we study prediction of Gaussian Volterra processes driven by a Brownian motion in a case where the Brownian motion is not observable, but only a noisy version is observed. As an applicati…
Deep learning has led to tremendous advancements in the field of Artificial Intelligence. One caveat however is the substantial amount of compute needed to train these deep learning models. Training a benchmark dataset like ImageNet on a single machine with a modern GPU can take upto a week, distributing training on mu…
We prove the existence of C^{\infty} local solutions to a class of mixed type Monge-Ampere equations in the plane. More precisely, the equation changes type to finite order across two smooth curves intersecting transversely at a point. Existence of C^{\infty} global solutions to a corresponding class of linear mixed ty…
This paper presents a novel end-to-end methodology for enabling the deployment of low-error deep networks on microcontrollers. To fit the memory and computational limitations of resource-constrained edge-devices, we exploit mixed low-bitwidth compression, featuring 8, 4 or 2-bit uniform quantization, and we model the i…
EMPIR combines low and full precision DNNs to enhance robustness against adversarial attacks.
problem Vulnerability of DNNs to adversarial attacks that misclassify inputs with small perturbations.
method Ensemble of quantized DNN models with different numerical precisions.
result EMPIR ensembles increase adversarial robustness by 42.6% on average across different tasks.
New method detects essential tori in mixed singularity links.
problem Detecting essential tori in mixed singularity link complements.
method Analyzing properties of defining mixed polynomials.
result Explicit criteria for essential tori existence.
Contrastive learning estimates transition kernels for continuous-time stochastic processes.
problem Estimating transition kernels for continuous-time stochastic processes without labeled data.
method Contrastive learning applied to strong-mixing continuous-time stochastic processes.
result Contrastive learning can estimate transition kernels for small-to-mid-range intervals in the diffusion case.
The paper develops new inequalities for Markov chain sums, linking them to mixing time.
problem Establishing concentration inequalities for Markov chain sums.
method Developed novel concentration inequalities for geometrically ergodic Markov chains, linking bounds to mixing time constants.
result Explicit bounds for additive functionals of Markov chains, linked to Rosenthal inequality constants and mixing properties.
New method trains Boltzmann machines without supervision.
problem Training unsupervised learning models.
method Mixed binary quadratic feasibility problem formulation.
result Theory validated on XOR patterns.
BayesBoost combines boosting and Bayesian methods for linear mixed models, improving uncertainty estimation and variable selection.
problem Lack of straightforward uncertainty estimation for parameters in high-dimensional linear mixed models.
method BayesBoost: Combines boosting and Bayesian inference for linear mixed models.
result Improves uncertainty estimation and variable selection in linear mixed models.
A new data augmentation method selects mixed classes based on class distances for better performance.
problem Improving recognition accuracy in object recognition using deep learning.
method Calculates class distances and selects mixed data from suitable classes dynamically.
result Improves recognition performance on general and long-tailed image recognition datasets.
New study on No-U-Turn Sampler for accelerated mixing in Hamiltonian Monte Carlo.
problem Achieving accelerated convergence in Hamiltonian Monte Carlo.
method Combining concentration of measure and coupling analysis for mixing.
result Rigorous mixing guarantees for the No-U-Turn Sampler in certain Gaussian distributions.
Investigates market dynamics with informed traders and high-frequency traders.
problem Trading large orders in a market with multiple high-frequency traders.
method Analyzes a three-period Kyle's model with a normal-speed informed trader and multiple anticipatory high-frequency traders under different inventory pressures.
result Surprising results: improving HFTs' speed or prediction can harm them but benefit the informed trader.
Paper reduces hyperparameters in mixed-categorical Gaussian processes for green aircraft optimization.
problem High-dimensional mixed-categorical Gaussian processes with many hyperparameters.
method Innovative dimension reduction algorithm using partial least squares regression.
result Significant reduction in fuel consumption (439 kg) for a green aircraft.
The study prevents model collapse in overparameterized linear regression by mixing real and synthetic labels.
problem Preventing model collapse in overparameterized linear regression.
method Iterative mixing of real and synthetic labels, deriving generalization error formulae.
result Optimal mixing ratio converges to the reciprocal of the golden ratio for isotropic features.
Batchboost stabilizes training by mixing and pairing samples, improving accuracy.
problem Stabilizing training in machine learning, especially avoiding overfitting and underfitting.
method Batchboost pipeline with three stages: pairing, mixing, and feeding. Mixing uses mixup technique.
result Batchboost achieves 0.5-3% better accuracy than mixup on CIFAR-10 & Fashion-MNIST.
Markov chain Monte Carlo (MCMC) algorithms are simple and extremely powerful techniques to sample from almost arbitrary distributions. The flaw in practice is that it can take a large and/or unknown amount of time to converge to the stationary distribution. This paper gives sufficient conditions to guarantee that univa…
Paper extends nonparametric regression bounds for dependent β-mixing samples.
problem Analyzing error in nonparametric regression with dependent data.
method Extends uniform deviation inequalities from independent to dependent β-mixing samples. result Derives generalization bounds for nonparametric regression with dependent data.
The study examines mixing times of data-augmentation Gibbs samplers for high-dimensional probit regression.
problem Investigating convergence properties of data-augmentation samplers for Bayesian probit regression.
method Using recent results on Gibbs samplers for log-concave targets, the study provides non-asymptotic bounds on mixing times.
result Explicit non-asymptotic bounds on mixing times depend on design matrix and prior precision, holding uniformly over responses.