30,946 Hits in 6.0 sec

Memory-efficient inference in dynamic graphical models using multiple cores

Galen Andrew, Jeff A. Bilmes
2012 Journal of machine learning research  
We introduce the archipelagos algorithm for memory-efficient multi-core inference in dynamic graphical models.  ...  The archipelagos algorithm applies to any dynamic graphical model, including dynamic Bayesian networks, conditional random fields, and hidden conditional random fields.  ...  The opinions expressed in this work are those of the authors and do not necessarily reflect the views of the funding agency.  ... 
dblp:journals/jmlr/AndrewB12 fatcat:mnwq32upqzhm3mkwlgq4qfzr7i

Edge AI without Compromise: Efficient, Versatile and Accurate Neurocomputing in Resistive Random-Access Memory [article]

Weier Wan
2021 arXiv   pre-print
Resistive random-access memory (RRAM) based compute-in-memory (CIM) architectures promise to bring orders of magnitude energy-efficiency improvement by performing computation directly within memory.  ...  a high degree of versatility for diverse model architectures, record energy-efficiency 5× - 8× better than prior art across various computational bit-precisions, and inference accuracy comparable to software  ...  are used to process temporal data such as audio signals, data travel recurrently through the same layer for multiple time-steps; in probabilistic graphical models such as restricted Boltzmann machine (  ... 
arXiv:2108.07879v1 fatcat:uaz2gufthzasjkza7zikqnfv7y

Best Practices for the Deployment of Edge Inference: The Conclusions to Start Designing

Georgios Flamis, Stavros Kalapothas, Paris Kitsos
2021 Electronics  
The inference phase on the other hand, uses a trained network with new data. The sensitive optimization and back propagation phases are removed and forward propagation is only used.  ...  The complete development flow undergoes two distinct phases; training and inference. During training, all the weights are calculated through optimization and back propagation of the network.  ...  Each thread is executed on a GPU core (ALU) and multiple GPU cores are grouped together into blocks. Threads operating at the same block can inter-exchange data through shared memory.  ... 
doi:10.3390/electronics10161912 fatcat:3ywb6inqzvbfxb2vjve6ffvmiq

A Unified Optimization Approach for CNN Model Inference on Integrated GPUs

Leyuan Wang, Zhi Chen, Yizhi Liu, Yao Wang, Lianmin Zheng, Mu Li, Yida Wang
2019 Proceedings of the 48th International Conference on Parallel Processing - ICPP 2019  
For the device with Intel Graphics, we used Intel OpenVINO toolkit as the baseline, which does optimized model inference on Intel Graphics using Intel clDNN along with some graph-level optimizations.  ...  An efficient computation pattern will mostly use the data stored in the register files and hide the latency to retrieve data from farther memories.  ... 
doi:10.1145/3337821.3337839 dblp:conf/icpp/WangCLWZLW19 fatcat:ptvsneujwjdmhesvcrune7rqwy

Stealing Webpages Rendered on Your Browser by Exploiting GPU Vulnerabilities

Sangho Lee, Youngsok Kim, Jangwoo Kim, Jong Kim
2014 2014 IEEE Symposium on Security and Privacy  
Graphics processing units (GPUs) are important components of modern computing devices for not only graphics rendering, but also efficient parallel computations.  ...  We detect that both browsers leave rendered webpage textures in GPU memory, so that we can infer which webpages a victim user has visited by analyzing the remaining textures.  ...  GPUs utilize a large number of processing cores and a large amount of independent memory for efficiently processing graphics operations and computational workloads.  ... 
doi:10.1109/sp.2014.9 dblp:conf/sp/LeeKKK14 fatcat:7zval6ut5jfmvagb4znlemmtcy

Memory-Centric Neuromorphic Computing With Nanodevices

Damien Querlioz, Julie Grollier, Tifenn Hirtzlin, Jacques-Olivier Klein, Etienne Nowak, Elisa Vianello, Marc Bocquet, Jean-Michel Portal, Miguel Romera, Philippe Talatchian
2019 2019 IEEE Biomedical Circuits and Systems Conference (BioCAS)  
Second, we see that brains use the physics of their memory devices in a way much richer than only storage.  ...  Brains, by contrast, achieve superior energy efficiency by fusing logic and memory entirely. Currently, emerging memory nanodevices give us an opportunity to reproduce this concept.  ...  For these reasons, these models are extremely attractive for inference hardware [15] , [19] - [21] .  ... 
doi:10.1109/biocas.2019.8919010 dblp:conf/biocas/QuerliozGHKNVBP19 fatcat:yttn7g4clfevhpcnhjrkzbvzre

Parallel Exact Inference on a CPU-GPGPU Heterogenous System

Hyeran Jeon, Yinglong Xia, Viktor K. Prasanna
2010 2010 39th International Conference on Parallel Processing  
Exact inference is a key problem in exploring probabilistic graphical models, where the computational complexity varies dramatically as the parameters of the graphical models changes.  ...  The scheduler can merge multiple small cliques or split large cliques dynamically so as to maximize the utilization of the GPGPU resources.  ...  BACKGROUND Exact Inference A Bayesian network is a probabilistic graphical model that exploits conditional independence to represent compactly a joint distribution.  ... 
doi:10.1109/icpp.2010.15 dblp:conf/icpp/JeonXP10 fatcat:4gntmnvifjf2lgzmiffmk2oapm

Heterogeneous computing for epidemiological model fitting and simulation

Thomas Kovac, Tom Haber, Frank Van Reeth, Niel Hens
2018 BMC Bioinformatics  
Conclusions: Utilizing GPUs for parameter inference can bring considerable increases in performance using average host systems with high-end consumer GPUs.  ...  The ultimate goal is to infer model parameters that enable the model to correctly describe observed data.  ...  In contrast, simulating multiple small models simultaneously can be done efficiently on GPUs.  ... 
doi:10.1186/s12859-018-2108-3 pmid:29548279 pmcid:PMC5857139 fatcat:b7jcmjvy2ndtnlstir5tk7cf2a

Bayesian Analogical Cybernetics [article]

Adam Safron
2019 arXiv   pre-print
It has been argued that all of cognition can be understood in terms of Bayesian inference. It has also been argued that analogy is the core of cognition.  ...  From the Bayesian perspective of the Free Energy Principle and Active Inference framework, thought is constituted by dynamics of cascading belief propagation through the nodes of probabilistic generative  ...  In addition to MCMC, analytic approaches based on variational (or approximate) inference are increasingly being used for their computational efficiency.  ... 
arXiv:1911.02362v2 fatcat:f7hmtrggvbggvh4izlfgomh4ae

Deep Probabilistic Programming [article]

Dustin Tran, Matthew D. Hoffman, Rif A. Saurous, Eugene Brevdo, Kevin Murphy, David M. Blei
2017 arXiv   pre-print
In addition, Edward can reuse the modeling representation as part of inference, facilitating the design of rich variational models and generative adversarial networks.  ...  For flexibility, Edward makes it easy to fit the same model using a variety of composable inference methods, ranging from point estimation to variational inference to MCMC.  ...  Stan only used 1 CPU as it leverages multiple cores by running HMC chains in parallel. Stan also used double-precision floating point as it does not allow single-precision.  ... 
arXiv:1701.03757v2 fatcat:f3zxlird3bbpblw2fcrq3zuypm

Efficient Implementation of MrBayes on Multi-GPU

J. Bao, H. Xia, J. Zhou, X. Liu, G. Wang
2013 Molecular biology and evolution  
By dynamically adjusting the task granularity to adapt to input data size and hardware configuration, it makes full use of GPU cores with different data sets.  ...  Recently, graphics processor unit (GPU) has shown its power as a coprocessor (or rather, an accelerator) in many fields.  ...  Acknowledgments This work was partially supported by National Natural Science Foundation of China (grant numbers 60903028 and 61070014) and by Key Projects in the Tianjin Science & Technology Pillar Program  ... 
doi:10.1093/molbev/mst043 pmid:23493260 pmcid:PMC3649675 fatcat:q3qs3dnsardcznob5rect276fa

The Tensor-Core Correlator

J. W. Romein
2021 Astronomy and Astrophysics  
Often, these instruments use graphics processing units (GPUs) to correlate the incoming data streams as GPUs are fast, energy efficient, flexible, and programmable with a reasonable programming effort.  ...  The library hides the complexity of the use of tensor cores and can be easily integrated into the GPU pipelines of existing and future instruments, leading to a significant reduction in costs and energy  ...  We kindly thank NVIDIA for providing us with an A100 GPU. A52, page 3 of 4 A&A 656, A52 (2021)  ... 
doi:10.1051/0004-6361/202141896 fatcat:b5e4hq2lyvc5bi34aqhvyoyfpm

Learning in a Distributed Software Architecture for Large-Scale Neural Modeling [chapter]

Jasmin Léveillé, Heather Ames, Benjamin Chandler, Anatoli Gorchetchnikov, Ennio Mingolla, Sean Patrick, Massimiliano Versace
2012 Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering  
On the other hand, recent work highlights that certain constraints on neural models must be imposed on neural and synaptic dynamics in order to take advantage of such systems.  ...  Progress on large-scale simulation of neural models depends in part on the availability of suitable hardware and software architectures.  ...  Beyond graphics processors, single-purpose hardware offers multiple additional orders of magnitude in power efficiency.  ... 
doi:10.1007/978-3-642-32615-8_65 fatcat:pltmv7jx5veoba4yigun26pz4a


Xiaofeng Wu, Jia Rao, Wei Chen, Hang Huang, Chris Ding, Heng Huang
2021 Proceedings of the 22nd International Middleware Conference  
This results in less interference and the elimination of out-of-memory errors.  ...  Spatial and temporal multitasking on GPU have been studied in the literature, but popular deep learning frameworks, such as Tensor-Flow and PyTorch, lack the support of GPU sharing among multiple DL models  ...  This work was supported in part by U.S. NSF grants CCF-1845706 and IIS-1852606.  ... 
doi:10.1145/3464298.3493391 fatcat:2fsjulz7gjahfe6n2533lcpgzi

Efficient Bayesian inference in stochastic chemical kinetic models using graphical processing units [article]

Jarad Niemi, Matthew Wheeler
2011 arXiv   pre-print
We show how graphical processing units can be efficiently utilized for parameter estimation in systems that hitherto were inestimable.  ...  A goal of systems biology is to understand the dynamics of intracellular systems.  ...  The content of the information herein does not necessarily reflect the position or policy of the Government and no official endoresement should be inferred.  ... 
arXiv:1101.4242v1 fatcat:ufaolqdq7jf6hifu7xfre7sfzy
« Previous Showing results 1 — 15 out of 30,946 results