Parallelizes autoregressive generation using VSSM.
problem Autoregressive models' inability to parallelize generation.
method Variational SSM (VSSM) with parallelizable sampling and decoding.
result Parallel generation possible with VSSM.
On-device research index
arXiv research
A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
169,291 papers · 148 categories
Trend · papers per month
48 results for “Parallel States”
We prove a conjecture formulated by Pablo M. Chacon and Guillermo A. Lobos in [Pseudo-parallel Lagrangian submanifolds in complex space forms, Differential Geom. Appl.] stating that every Lagrangian pseudo-parallel submanifold of a complex space form of dimension at least 3 is semi-parallel.
dSMC improves parallel processing of state-space models.
problem Processing multiple observations efficiently in state-space models.
method A parallel-in-time particle smoother that reduces complexity to log(T).
result dSMC achieves O(log(T)) time complexity on parallel architectures.
Predictability enables efficient parallelization of nonlinear models.
problem Understanding which nonlinear state space models can be efficiently parallelized.
method Established a relationship between system dynamics and optimization problem conditioning, quantified by the largest Lyapunov exponent.
result Predictable systems can be evaluated in time, improving over conventional sequential approaches.
Paranom generates parallel anomaly datasets for improving LSTM-AD's accuracy.
problem Improving anomaly detection model accuracy.
method Parallel anomaly dataset generator design.
result Demonstrates usefulness in improving LSTM-AD's classification correctness.
Parallelizes MCTS for continuous domains using leaf and root parallelization.
problem Solving challenging tasks in continuous domains using MCTS.
method Extends existing parallelization strategies to continuous domains, focusing on leaf and root parallelization.
result Proposes two final selection strategies for continuous states in root parallelization.
NeLLoC improves image compression with parallel decoding.
problem Image compression with OOD generalization.
method Local autoregressive model with parallel decoding.
result Significant gains in compression runtime.
SySCD improves SCD scalability and speeds up training.
problem Scalability issues in parallel SCD algorithms.
method Developed a system-aware parallel SCD algorithm (SySCD) to avoid bottlenecks.
result Offers up to x42 speedup compared to state-of-the-art GLM solvers.
Optimizes parallel training of linear models, improving convergence.
problem Improving convergence of parallel training of linear models.
method Data partitioning scheme across threads to improve convergence.
result Achieved up to 42x speedup in convergence compared to state of the art implementations.
Stochastic gradient descent~(SGD) and its variants have become more and more popular in machine learning due to their efficiency and effectiveness. To handle large-scale problems, researchers have recently proposed several parallel SGD methods for multicore systems. However, existing parallel SGD methods cannot achieve…
A submanifold of a Riemannian symmetric space is called parallel if its second fundamental form is a parallel section of the appropriate tensor bundle. We classify parallel submanifolds of the Grassmannian $\rmG^+_2(\R^{n+2})$ which parameterizes the oriented 2-planes of the Euclidean space \,. Our main resul…
A new algorithm speeds up solving complex spin system problems.
problem Finding the lowest energy state in Ising models, which is hard and important.
method Mean-field Annealing from a Random State (MARS) algorithm, combining SA and MFA.
result MARS solves large Ising spin systems and maximum cut problems efficiently.
Cyclic Data Parallelism reduces memory usage and balances gradient communications.
problem Training large deep learning models requires efficient parallelism to scale.
method Cyclic Data Parallelism shifts micro-batches from simultaneous to sequential execution, balancing memory and gradient communications.
result Cyclic Data Parallelism reduces total memory usage and balances gradient communications.
Analysis of Wide and Deep Echo State Networks for Multiscale Spatiotemporal Time Series Forecastingeess.SP
Study on deep and wide echo state networks for forecasting complex time series.
problem Performance analysis of deep reservoir computing models.
method Investigates the impact of partitioning neurons and parallel pathways on forecasting accuracy.
result Wide and deep networks outperform shallow models in forecasting multiscale spatiotemporal data.
PL-NMF improves parallel NMF by optimizing data locality.
problem Data movement costs dominate in parallel NMF applications.
method Developed a parallel NMF algorithm based on HALS with data locality optimizations.
result Significant performance improvement over existing parallel NMF algorithms.
The Voice Conversion Challenge 2018: Promoting Development of Parallel and Nonparallel Methodseess.AS
Challenge evaluates voice conversion systems using parallel and non-parallel data.
problem Evaluate and compare different voice conversion systems.
method Hub and Spoke tasks with crowdsourced evaluation.
result Naturalness and similarity ratings provided for submitted systems.
Monte Carlo (MC) methods are widely used for Bayesian inference and optimization in statistics, signal processing and machine learning. A well-known class of MC methods are Markov Chain Monte Carlo (MCMC) algorithms. In order to foster better exploration of the state space, specially in high-dimensional applications, s…
Enhances parallelism in decentralized learning for larger networks.
problem Scalability limitations in decentralized learning with increasing number of machines.
method Proposes Decentralized Anytime SGD, a novel algorithm that extends parallelism threshold.
result Establishes a theoretical upper bound on parallelism surpassing current state-of-the-art.
New heuristics for parallel Bayesian optimization.
problem Challenges in parallel and scalable Bayesian optimization.
method Review and propose practical heuristic algorithms.
result Simple, heuristic algorithms for parallel Bayesian optimization.
New LSTM structure improves text representation efficiency.
problem Sequential limitations of Bi-directional LSTMs.
method Sentence-State LSTM with parallel word states and simultaneous information exchange.
result Sentence-State LSTM achieves competitive performance on text benchmarks.
A parallel algorithm learns efficient Kronecker product dictionaries.
problem Sparse representation of 2D signals like images and hyperspectral data.
method Highly parallelizable algorithm for learning separable dictionaries.
result Competitive sparse representations at lower computational cost.
New deep ESN architectures improve memory capacity and prediction accuracy.
problem Improving memory capacity and prediction accuracy of ESNs.
method Two new deep ESN architectures: parallel and series. Analysis of memory capacity and prediction accuracy.
result Parallel deep ESNs have equivalent memory capacity to shallow ESNs, while series deep ESNs have smaller memory capacity.
New method solves blind inverse problems by optimizing both operator and image parameters.
problem Solving blind inverse problems with known forward operator.
method Parallel reverse diffusion guided by gradients from intermediate stages.
result State-of-the-art performance on blind deblurring and imaging through turbulence.
StarGAN-VC: Non-parallel many-to-many voice conversion with star generative adversarial networkscs.SD
Paper proposes StarGAN-VC for non-parallel voice conversion.
problem Non-parallel many-to-many voice conversion.
method Variant of GAN called StarGAN for simultaneous many-to-many mappings.
result Obtained higher sound quality and speaker similarity than state-of-the-art methods.
Study surfaces with constant mean curvature in 4D spaces.
problem Classify surfaces with parallel mean curvature in homogeneous 4-manifolds.
method Survey and provide a common framework for classification results.
result Existence of holomorphic quadratic differentials on these surfaces.
Parallel-in-time solver reduces ODE simulation time from linear to logarithmic.
problem Efficiently solving ordinary differential equations (ODEs) with reduced computational cost.
method Formulated a parallel-in-time probabilistic numerical ODE solver using time-parallel formulation of iterated extended Kalman smoothers.
result Reduces span cost from linear to logarithmic in the number of time steps.
New vector quantization method reduces relevance of parallel components in database points.
problem Scaling maximum inner product search to massive databases.
method Developed anisotropic vector quantization loss functions.
result Achieves state-of-the-art results on public benchmarks.
AgEBO-Tabular combines NAS and hyperparameter tuning for fast, high-performing tabular models.
problem Developing high-performing predictive models for large tabular data sets is challenging.
method Combines aging evolution NAS and asynchronous Bayesian optimization for hyperparameter tuning in data-parallel training.
result Automatically discovered neural network models outperform state-of-the-art AutoML ensembles in inference speed by two orders of magnitude.
ASHA optimizes hyperparameters for large models in parallel.
problem Large hyperparameter spaces and long training times in modern learning models.
method ASHA algorithm exploiting parallelism and early-stopping.
result ASHA outperforms existing methods and scales linearly with workers.
Study of bound states in quantum layers with confining potentials.
problem Investigating bound states in quantum layers with confining potentials.
method Developed a general approach using parallel coordinates based on the surface but outside its cut locus.
result Discrete eigenvalues exist for certain quantum layers with positive total Gauss curvature.
New parallel GCN techniques boost graph embedding scalability and efficiency.
problem Scalability issues in graph embedding models.
method Parallelization techniques for graph sampling and feature propagation.
result Superior scalable performance on large graphs without accuracy loss.
Parle speeds up SGD training by 2-4x with better results.
problem Training deep networks efficiently and accurately.
method Parallelizes SGD by performing more computation locally.
result Achieves state-of-the-art error rates on benchmarks.
New method accelerates Parallel Tempering using neural samplers.
problem Challenges in sampling from high-dimensional, multimodal distributions.
method Leverages neural samplers to reduce overlap between distributions.
result Improves sample quality and reduces computational cost.
Combines local MCMC chains to speed up sampling.
problem Designing efficient MCMC chains that mix well over the whole state space.
method Combining parallel chains prioritized by kernel Stein discrepancy, combining samples using novel probability estimation.
result Significant speedups in sampling from multimodal distributions.
Timelike Surfaces with Parallel Normalized Mean Curvature Vector Field in the Minkowski 4-Spacemath.DG
Study timelike surfaces with parallel mean curvature in Minkowski 4-space.
problem Existence and uniqueness of timelike surfaces with parallel mean curvature.
method Introduce canonical parameters and prove existence and uniqueness theorem.
result Each timelike surface with parallel mean curvature is determined by three geometric functions.
New BO methods exploit parallel experiments, reducing search time and improving solution quality.
problem Limitation of Bayesian optimization in exploiting parallel experiments.
method Propose new parallel BO paradigms that exploit the structure of the system to partition the design space.
result Significantly reduce search time and increase probability of finding global solutions.
Mesh-TensorFlow enables efficient deep learning on large clusters.
problem Memory constraints and inefficiency in batch-splitting for large models.
method Introduces Mesh-TensorFlow for specifying general tensor computations across a multi-dimensional mesh of processors.
result Trains Transformer models with up to 5 billion parameters on TPU meshes of up to 512 cores.
The study characterizes compact homogeneous manifolds with Bismut parallel torsion.
problem Characterizing compact homogeneous manifolds with specific geometric properties.
method Investigating Hermitian manifolds with Bismut parallel torsion, focusing on locally homogeneous manifolds.
result Characterization of compact Chern flat BTP manifolds and properties of BTP compact Hermitian locally homogeneous manifolds.
Efficient event generation for collider phenomenology using parallel Langevin sampling and learned Stein diagnostics.
problem Event generation for precision collider phenomenology.
method Parallel Langevin sampling with learned Stein diagnostics.
result Relaxation time is estimated using a data-driven approach.
The object of this article is to compute the holonomy group of the normal connection of complex parallel submanifolds of the complex projective space. We also give a new proof of the classification of complex parallel submanifolds by using a normal holonomy approach. Indeed, we explain how these submanifolds can be reg…
Paper tackles robust knowledge transfer in parallel RL tasks.
problem Transfer knowledge from low-tier to high-tier tasks in parallel RL without shared dynamics or reward functions.
method Identifies Optimal Value Dominance condition and proposes online learning algorithms for both tasks.
result Achieves constant regret on partial states and near-optimal regret when tasks are dissimilar.
Paper proposes DCT for efficient hybrid parallel training of large recommendation models.
problem Training large recommendation models at scale with efficient communication.
method Dynamic Communication Thresholding (DCT) for both Data Parallelism and Model Parallelism.
result Reduces communication by 100x and 20x during DP and MP, respectively, improving training time by 37%.
KalMamba improves RL efficiency with probabilistic SSMs.
problem Efficiency in learning and inference for probabilistic SSMs in RL.
method Combines Mamba's scalability with Kalman filtering for efficient probabilistic SSMs.
result KalMamba outperforms state-of-the-art SSMs in RL, especially on longer sequences.
Proposes a method to improve SLMC for multimodal distributions.
problem Difficulty of applying SLMC to multimodal distributions.
method Parallel adaptive annealing with VAE-SLMC.
result Can proficiently obtain accurate samples from multimodal distributions.
Data consistency networks for (calibration-less) accelerated parallel MR image reconstructioneess.IV
New networks improve MRI image reconstruction without calibration.
problem Calibration issues in parallel MR image reconstruction.
method Data consistency layers in deep CNN networks.
result Proposed methods outperform existing techniques.
A new GNN model SPIN achieves state-of-the-art performance on diverse real-world datasets.
problem Graph classification efficiency and accuracy.
method Parallel neighborhood aggregations (PA-GNNs) and SPIN model.
result SPIN model achieves state-of-the-art performance on diverse real-world datasets.
QEM uses parallel importance weighting for fast approximate Bayesian inference.
problem Bayesian inference challenges in large models with many observations and latent variables.
method Expectation Maximization (EM) with massively parallel importance weighting.
result QEM is faster and more scalable than RWS and VI.
A fundamental task in machine learning and related fields is to perform inference on Bayesian networks. Since exact inference takes exponential time in general, a variety of approximate methods are used. Gibbs sampling is one of the most accurate approaches and provides unbiased samples from the posterior but it has hi…