Improved analysis of accelerated noisy power method for PCA.
problem Inexact matrix-vector products in PCA settings.
method Improved analysis of Accelerated Noisy Power Method under milder perturbation conditions.
result Worst-case optimal convergence rate with relaxed noise conditions.
We accelerate the power method for strong low-rank approximation using fast sketching.
problem Efficiency bottleneck in power method for large target ranks.
method Developed an algorithmic and theoretical framework for accelerating the power method using fast sketching.
result Simple and provably efficient methods for singular value decomposition, low-rank factorization, and Nyström approximation.
Principal component analysis (PCA) is one of the most powerful tools in machine learning. The simplest method for PCA, the power iteration, requires O(1/Δ) full-data passes to recover the principal component of a matrix with eigen-gap Δ. Lanczos, a significantly more complex method, achieves an accelerated…
We analyze Riemannian accelerated methods using a new framework.
problem Understanding Riemannian accelerated gradient methods.
method Riemannian A-HPE framework, focusing on Euclidean A-HPE insights and metric distortion control.
result Characterization of acceleration for various Riemannian methods.
Recent technological advances have proliferated the available computing power, memory, and speed of modern Central Processing Units (CPUs), Graphics Processing Units (GPUs), and Field Programmable Gate Arrays (FPGAs). Consequently, the performance and complexity of Artificial Neural Networks (ANNs) is burgeoning. While…
Accelerators with power-law memory are proposed in the framework of the discrete time approach. To describe discrete accelerators we use the capital stock adjustment principle, which has been suggested by Matthews.The suggested discrete accelerators with memory describe the economic processes with the power-law memory …
Superposition accelerates training to a universal power-law exponent.
problem Training dynamics in neural networks.
method Teacher-student framework and analytic theory.
result Superposition leads to a universal power-law exponent of ~1, independent of data and channel statistics.
Computer vision performances have been significantly improved in recent years by Convolutional Neural Networks(CNN). Currently, applications using CNN algorithms are deployed mainly on general purpose hardwares, such as CPUs, GPUs or FPGAs. However, power consumption, speed, accuracy, memory footprint, and die size sho…
FMCIT accelerates CI tests for causal discovery, maintaining power and efficiency.
problem High computational complexity in CI tests limits practical applicability of causal discovery methods.
method Flow Matching-based Conditional Independence Test (FMCIT) that leverages flow matching for fast CI tests.
result FMCIT effectively controls type-I error and maintains high testing power under the alternative hypothesis.
New factorial power constants improve optimization convergence rates.
problem Optimization convergence rates depend on various constants.
method Proposes using factorial powers for defining these constants.
result Factorial powers simplify or improve convergence rates of optimization methods.
New method accelerates CNNs for mobile devices by approximating tensors and quantizing weights.
problem Efficiently compress and accelerate CNNs for mobile devices.
method Low-rank tensor approximation in Tucker format combined with quantization of weights and activations.
result Our method significantly improves CNN performance on various classification tasks.
Hardware-accelerated RBM solves large combinatorial problems and integer factorization.
problem Solving large combinatorial optimization and integer factorization problems.
method Logically synthesized RBM architecture, hardware acceleration, and efficient training methods.
result Hardware-accelerated RBM factorizes 16-bit numbers with 10000x speed and 32x power improvements.
Stochastic gradient descent (\textsc{Sgd}) methods are the most powerful optimization tools in training machine learning and deep learning models. Moreover, acceleration (a.k.a. momentum) methods and diagonal scaling (a.k.a. adaptive gradient) methods are the two main techniques to improve the slow convergence of \text…
Memory bandwidth bottleneck is a major challenges in processing machine learning (ML) algorithms. In-memory acceleration has potential to address this problem; however, it needs to address two challenges. First, in-memory accelerator should be general enough to support a large set of different ML algorithms. Second, it…
Replica exchange Langevin diffusion accelerates nonconvex optimization.
problem Nonconvex optimization challenges in machine learning.
method Replica exchange Langevin diffusion, discretization analysis.
result Replica exchange accelerates convergence to global minima.
Accelerates TPP sampling with speculative decoding for faster sequence generation.
problem Efficiently sampling from complex temporal point processes.
method Adapting speculative decoding techniques from language models to TPPs.
result Achieves significant speedup (2-6x) while maintaining distributional accuracy.
This research analyzes and accelerates score-based diffusion models using discretization and Hessian information.
problem Theoretical foundations and convergence analysis of score-based diffusion models.
method Investigation of various discretization schemes, including Euler, exponential integrators, and midpoint randomization. Proposal of an accelerated sampler based on local linearization method.
result Hessian-based approach achieves faster convergence rates of order $\widetilde{\mathcal{O}}\left(\frac{1}{\varepsilon}
ight)$, significantly improving upon vanilla diffusion models.
The development of machine learning is promoting the search for fast and stable minimization algorithms. To this end, we suggest a change in the current gradient descent methods that should speed up the motion in flat regions and slow it down in steep directions of the function to minimize. It is based on a "power grad…
Unified framework for accelerating DNNs on resource-limited platforms.
problem Accelerating DNN execution on resource-limited platforms.
method Block-based pruning framework with reweighted regularization.
result First universal framework for both CNNs and RNNs with real-time acceleration and no accuracy compromise.
PCNN prunes CNN weights efficiently for hardware acceleration.
problem Efficiently compressing CNN models for hardware acceleration.
method PCNN uses a novel Sparsity Pattern Mask (SPM) to encode and prune weights.
result PCNN achieves up to 8.4X compression with minimal accuracy loss.
Survey of RL methods for optimizing power grid topologies.
problem Optimizing power grid operation with adaptive control strategies.
method Reinforcement Learning (RL) for dynamic and uncertain environments.
result Comprehensive evaluation of RL-based methods for power grid topology optimization.
Machine Learning (ML) is making a strong resurgence in tune with the massive generation of unstructured data which in turn requires massive computational resources. Due to the inherently compute- and power-intensive structure of Neural Networks (NNs), hardware accelerators emerge as a promising solution. However, with …
ALONE uses neural networks to accelerate cine MRI without needing ground truth.
problem Accelerated MRI reconstruction without using ground truth.
method Unsupervised learning of shallow CNNs to approximate MRI patches.
result ALONE outperforms TV and DIC methods in cine MRI reconstruction.
A new method solves diagonally constrained SDPs quickly and accurately.
problem Solving large-scale diagonally constrained SDPs efficiently.
method Combines momentum from convex optimization with coordinate descent and matrix factorization.
result Local linear convergence and first-order critical point convergence proved.
GPU-accelerates multiuser detection for 5G URLLC systems.
problem Efficiently detecting payloads in OFDM frames with ultra-low latency.
method Implemented partially linear multiuser detection in RKHSs on a GPU-accelerated platform.
result Sub-millisecond latency detection in 5G URLLC systems.
RVI accelerates encoderless VI for faster convergence.
problem Slow convergence in encoderless VI methods.
method Introduces Relay Variational Inference (RVI) for faster learning.
result RVI outperforms existing methods in convergence speed and performance.
We propose a generic algorithmic building block to accelerate training of machine learning models on heterogeneous compute systems. Our scheme allows to efficiently employ compute accelerators such as GPUs and FPGAs for the training of large-scale machine learning models, when the training data exceeds their memory cap…
Paper proposes NASAIC framework for co-designing neural architectures and heterogeneous ASICs.
problem Designing efficient neural architectures and ASICs for multiple tasks.
method Build ASIC templates and propose NASAIC framework for simultaneous design of architectures and ASICs.
result NASAIC ensures design specifications and maximizes accuracy with minimal performance loss.
A Python package for GPU-accelerated signature kernel computation.
problem Efficient computation of signature kernels for sequential data.
method GPU-accelerated algorithms and tensor sketches.
result New algorithm outperforms existing methods.
Parallelizes feedforward computation using nonlinear equation solving.
problem Sequential nature of feedforward computation limits parallelization.
method Frame feedforward computation as solving nonlinear equations; use Jacobi or Gauss-Seidel methods for parallel updates.
result Accelerates feedforward computation with reduced parallelizable iterations.
In recent years, Convolutional Neural Network (CNN) based methods have achieved great success in a large number of applications and have been among the most powerful and widely used techniques in computer vision. However, CNN-based methods are computational-intensive and resource-consuming, and thus are hard to be inte…
Convolutional Neural Networks (CNN) has become more popular choice for various tasks such as computer vision, speech recognition and natural language processing. Thanks to their large computational capability and throughput, GPUs ,which are not power efficient and therefore does not suit low power systems such as mobil…
PoWER-BERT speeds up BERT inference by eliminating redundant word-vectors.
problem Improving BERT inference speed without sacrificing accuracy.
method Eliminating redundant word-vectors using a self-attention-based significance measure and learning the number of vectors to eliminate.
result Up to 4.5x reduction in inference time with <1% loss in accuracy on GLUE benchmark.
In recent years, deep neural networks (DNN) have demonstrated significant business impact in large scale analysis and classification tasks such as speech recognition, visual object detection, pattern extraction, etc. Training of large DNNs, however, is universally considered as time consuming and computationally intens…
Improved KSD test for faster GoF testing.
problem Slow and computationally intractable KSD tests.
method Nyström acceleration for KSD estimation.
result Asymptotic properties preserved by Nyström acceleration.
Paper explores second-order optimization in first-order methods, proving and disproving the necessity of square root.
problem Understanding and optimizing first-order optimization methods using second-order information.
method Rigorously proves Nesterov Accelerated Gradient uses past and current gradients to approximate Hessian. Relates adaptive methods to Natural Gradient Descent. Introduces AdaSqrt algorithm to remove square root in denominator.
result New algorithm AdaSqrt comparable to first-order methods on MNIST and beats Adam on CIFAR-10, casting doubt on the necessity of square root.
Gradient Boosting Machine (GBM) is an extremely powerful supervised learning algorithm that is widely used in practice. GBM routinely features as a leading algorithm in machine learning competitions such as Kaggle and the KDDCup. In this work, we propose Accelerated Gradient Boosting Machine (AGBM) by incorporating Nes…
Accelerated magnetic resonance (MR) scan acquisition with compressed sensing (CS) and parallel imaging is a powerful method to reduce MR imaging scan time. However, many reconstruction algorithms have high computational costs. To address this, we investigate deep residual learning networks to remove aliasing artifacts …
A new KDE model prevents singular solutions and accelerates optimization for probabilistic modeling.
problem Adapting to varying densities in data regions for probabilistic modeling.
method Adaptive KDE model with individual bandwidths, LOO-MLL criterion, and modified EM algorithm.
result The proposed models prevent singular solutions and have promising performance.
Interpreting gradient methods as fixed-point iterations, we provide a detailed analysis of those methods for minimizing convex objective functions. Due to their conceptual and algorithmic simplicity, gradient methods are widely used in machine learning for massive data sets (big data). In particular, stochastic gradien…
NumPyro accelerates probabilistic programming with JAX transformations.
problem Efficiently handling probabilistic models with hardware acceleration.
method Composable effects and program transformations in JAX.
result Iterative NUTS formulation JIT compiled for speed.
New mechanism found for power laws including Zipf's law.
problem Understanding the ubiquity of power law distributions.
method Introduced nonlinear self-excited Hawkes processes with fast-accelerating intensities.
result Wide class of nonlinear Hawkes processes have power law intensity PDFs.
Emerging resistive random-access memory (ReRAM) has recently been intensively investigated to accelerate the processing of deep neural networks (DNNs). Due to the in-situ computation capability, analog ReRAM crossbars yield significant throughput improvement and energy reduction compared to traditional digital methods.…
LightOn OPUs accelerate randomized numerical linear algebra, reducing computational costs.
problem Computational bottleneck in randomization step for large-scale linear algebra.
method Near constant-time linear random projections from LightOn OPUs.
result Significant acceleration of RandNLA algorithms with negligible precision loss.
Adaptive tuning of latent space for non-stationary data.
problem Learning from large, non-stationary systems with quick characteristic changes.
method Adaptive tuning of low-dimensional latent space based on real-time feedback.
result Improved prediction of time-varying charged particle beam properties.
Optimized neural networks for Edge TPU achieve high accuracy in real-time image classification.
problem Designing neural networks for hardware accelerators to achieve optimal performance.
method Hardware-aware neural architecture search and model customization for Edge TPU.
result Improved accuracy-latency tradeoff on Pixel 4's Edge TPU compared to existing models.
VegasFlow accelerates complex simulations across various hardware platforms.
problem Complex calculations and simulations requiring high-dimensional integrals.
method Monte Carlo integration techniques using Vegas algorithm and TensorFlow.
result Significantly faster performance on various hardware platforms.
A new family of momentum coefficients improves the convergence rate of accelerated algorithms.
problem Improving the convergence rate of accelerated gradient methods for strongly convex functions.
method Introducing a family of controllable momentum coefficients for forward-backward accelerated methods.
result Established a controllable $O\left(1/k^{2α}
ight)$ convergence rate for the NAG-α method.