Filters








75,203 Hits in 7.3 sec

Scalable-effort classifiers for energy-efficient machine learning

Swagath Venkataramani, Anand Raghunathan, Jie Liu, Mohammed Shoaib
2015 Proceedings of the 52nd Annual Design Automation Conference on - DAC '15  
In this paper, we propose scalable-effort classifiers, a new approach to optimizing the energy efficiency of supervised machine-learning classifiers.  ...  To address this inefficiency, we propose a systematic approach to design scalable-effort classifier that dynamically adjust their computational effort depending on the difficulty of the input data, while  ...  In the traditional approach shown in Fig. 1(a) , input training examples are used to build a decision boundary (model X) that separates data into two categories or classes.  ... 
doi:10.1145/2744769.2744904 dblp:conf/dac/VenkataramaniRL15 fatcat:2b7dau6d4vag3dka3gedf345e4

BDI Agent Architectures: A Survey

Lavindra de Silva, Felipe Meneguzzi, Brian Logan
2020 Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence  
In this paper, we survey the main approaches to each component of the BDI architecture, how these have been realised in agent programming languages, and discuss the trade-offs inherent in each approach  ...  In two-stage problems, instantiating the SAA approach requires to define a single set of decision variables for stage-1, and one set of decision variables per scenario for stage-2.  ...  Many such approaches focus on two-stage stochastic problems: these require to determine a single set of decisions for stage-1; after such decisions have been implemented, uncertain outcomes are observed  ... 
doi:10.24963/ijcai.2020/674 dblp:conf/ijcai/Filippo0M20 fatcat:cs2lgbvt7fdrrebqo5ldbnk3ee

Evolutionary Computation for Large-scale Multi-objective Optimization: A Decade of Progresses

Wen-Jing Hong, Peng Yang, Ke Tang
2021 International Journal of Automation and Computing  
This paper reviews the progress of evolutionary computation for large-scale multi-objective optimization from two angles.  ...  AbstractLarge-scale multi-objective optimization problems (MOPs) that involve a large number of decision variables, have emerged from many real-world applications.  ...  The first stage is to find a small set of Pareto-optimal solutions and the second is to reconstruct the entire front departing from a few points from such front.  ... 
doi:10.1007/s11633-020-1253-0 fatcat:q636cuvarbco7kmnvthnfysa54

Integrated facility location and capacity planning under uncertainty

Isabel Correia, Teresa Melo
2021 Computational and Applied Mathemathics  
While in the first model decisions related to capacity scalability are modeled as first-stage decisions, in the second model, capacity adjustments are deferred to the second stage.  ...  We propose two different frameworks for planning capacity decisions and present a two-stage stochastic model for each one of them.  ...  To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.  ... 
doi:10.1007/s40314-021-01560-0 fatcat:5gervis3urbnhlagsonvk5pbvu

Optimized combinatorial clustering for stochastic processes

Jumi Kim, Wookey Lee, Justin Jongsu Song, Soo-Bok Lee
2017 Cluster Computing  
As a new data processing era like Big Data, Cloud Computing, and Internet of Things approaches, the amount of data being collected in databases far exceeds the ability to reduce and analyze these data  ...  The algorithm outperforms conventional approaches through various numerical and qualitative thresholds like mean and standard deviation of accuracy and computation speed.  ...  Among two-stage procedures, the Bayesian decision-theoretic procedures have the best overall performance characteristics.  ... 
doi:10.1007/s10586-017-0763-1 fatcat:tgpierpyybdu5cuwes6qctn5oy

Automatically Learning Compact Quality-aware Surrogates for Optimization Problems [article]

Kai Wang, Bryan Wilder, Andrew Perrault, Milind Tambe
2020 arXiv   pre-print
Recent work has shown that including the optimization problem as a layer in the model training pipeline results in predictions of the unobserved parameters that lead to higher decision quality.  ...  fail to improve solution quality due to non-smoothness issues that arise when training through a complex optimization layer.  ...  The computations in this paper were run on the FASRC Cannon cluster supported by the FAS Division of Science Research Computing Group at Harvard University.  ... 
arXiv:2006.10815v2 fatcat:ixvcqrjgh5gfhgpg7ew6jdy43q

Multi-Objective Service Oriented Network Provisioning In Ultra-Scale Systems

Dragi Kimovski, Roland Matha, Sasko Ristov, Radu Prodan
2017 Zenodo  
To enable sustainable ultra-scale computing, there are multiple major challenges, which have to be tackled, such as, improved data distribution, increased systems scalability, enhanced fault tolerance,  ...  Regrettably, the current research initiatives in the area of ultra-scale computing are in a very early stage of research and are predominantly concentrated on the management of the computational and storage  ...  Acknowledgments This work is being accomplished as a part of project ENTICE: "dEcentralised repositories for traNsparent and efficienT vIrtual maChine opErations", funded  ... 
doi:10.5281/zenodo.1066698 fatcat:ybwody3vdbgfla3khubnd7epce

Power-efficient video encoding on resource-limited systems: A game-theoretic approach

Wen Ji, Jiangchuan Liu, Min Chen, Yiqiang Chen
2012 Future generations computer systems  
This paper is dedicated to developing a power-scalable video encoding (PSVE) strategy for energy-limited mobile terminals.  ...  With limited power supply, it is a challenging issue to support video applications with high resolution due to the complex functionality and high resource requirements.  ...  Power allocation control The proposed power-scalable control includes two stages. In the first stage, target power consumption are allocated at the frame level.  ... 
doi:10.1016/j.future.2011.04.002 fatcat:oxc7bdbg65amhgbywial63xnkm

Optimal automatic multi-pass shader partitioning by dynamic programming

Alan Heirich
2005 Proceedings of the ACM SIGGRAPH/EUROGRAPHICS conference on Graphics hardware - HWWS '05  
Experimental results on a set of test cases with a commercial prerelease compiler for a popular high level shading language showed a DP algorithm had an average runtime cost of O(n 1.14966 ) which is less  ...  This demonstrates that efficient and optimal automatic shader partitioning can be an emergent byproduct of a DP-based code generator for a very high performance GPU.  ...  When the first stage has been computed the set P satisfies the Principle of Optimality for paths of length n. The best p ∈ P represents the globally optimal solution.  ... 
doi:10.1145/1071866.1071881 fatcat:wqads3nupffptocv3kz6ubrscq

Learning Fast Optimizers for Contextual Stochastic Integer Programs

Vinod Nair, Dj Dvijotham, Iain Dunning, Oriol Vinyals
2018 Conference on Uncertainty in Artificial Intelligence  
We present a novel reinforcement learning (RL) approach to learning a fast and highly scalable solver for a two-stage stochastic integer program in the large-scale data setting.  ...  We also propose learning a policy to compute a bound on the objective using dual decomposition.  ...  This provides another perspective of our work as a hybrid approach that combines a learned contextual solver with a MIP solver in the learning loop to efficiently optimize stochastic two-stage MIPs.  ... 
dblp:conf/uai/NairDDV18 fatcat:qpc6fa5s4jdcrjs2xtmcuyky5m

Scalable Feature Extraction for Coarse-to-Fine JPEG 2000 Image Classification

A. Descampe, C. De Vleeschouwer, P. Vandergheynst, B. Macq
2011 IEEE Transactions on Image Processing  
An efficient way to learn and optimize such cascade is proposed.  ...  An original representation, called integral volume, is first proposed to compute local image features progressively from the compressed code-stream, on any spatial image area, regardless of the codeblock  ...  The computational cost measures the fraction of bits that need to be entropically decoded to make a decision on all images.  ... 
doi:10.1109/tip.2011.2126584 pmid:21411407 fatcat:w2aefecchvehvg7ibkolwre32y

Improving Symbolic System-Level Synthesis by Solver Coordination and Domain-Specific Heuristics

Christian Haubelt, Alexander Rausch
2022 Electronics  
Based on the answer set solver's decisions for binding and routing, a background theory solver then computes time-triggered schedules to resolve resource access conflicts.  ...  computational utilization per tile, the coordinated synthesis approach scales up to 5< [...]  ...  As manual approaches are likely to fail in this setting, symbolic approaches based on satisfiability modulo theories (SMT) emerged as scalable automated decision making approaches in the area of system  ... 
doi:10.3390/electronics11121888 doaj:31d27f8d6a774a1fb08d76845d5f06f6 fatcat:mk5mttzyzvgszjbwr4ulewe6ve

Complexity-distortion tradeoffs in vector matching based on probabilistic partial distance techniques

K. Lengwehasatit, A. Ortega
1999 Proceedings DCC'99 Data Compression Conference (Cat. No. PR00096)  
We propose an approach that computes a partial distance metric and uses prior probabilistic knowledge of the reliability of the estimate to decide on whether to stop the distance computation.  ...  In this paper we consider the problem of searching for the best match for an input among a set of vectors, according to some predetermined metric.  ...  Hypothesis Testing Fast Matching One drawback of the DTFM approach is that it does not provide any computation scalability, i.e., DTFM achieves the same solution as optimal matching but does not allow  ... 
doi:10.1109/dcc.1999.755689 dblp:conf/dcc/LengwehasatitO99 fatcat:2uf32msj35hi7akgqui4eqlm3e

The case for a versatile storage system

Samer Al-Kiswany, Abdullah Gharaibeh, Matei Ripeanu
2010 ACM SIGOPS Operating Systems Review  
To deal with these limitations we propose a new operational approach: versatile storage, an application-optimized and highly configurable storage system that harnesses node-local resources, is configured  ...  Storage systems in emerging large-scale (a.k.a. peta-scale) computing systems often introduce a performance or scalability bottleneck.  ...  Moreover, this approach often introduces a scalability bottleneck.  ... 
doi:10.1145/1740390.1740394 fatcat:d3imzypy3jbj7birhf2vcnkz6a

Petascaling Machine Learning Applications With Mr-Mpi

Cevdet Aykanat
2014 Zenodo  
This whitepaper addresses applicability of the Map/Reduce paradigm for scalable and easy parallelization of fundamental data mining approaches with the aim of exploring/enabling processing of terabytes  ...  The obtained results show that utilization of the Map/Reduce paradigm can be a strong enhancer on the road to petascale.  ...  We believe that the MR-MPI library has the possibility to make a great impact in scientific computing since it eases parallel programming while providing high scalability for HPC platforms.  ... 
doi:10.5281/zenodo.823040 fatcat:f4olnrq2zfcsjksd7e4k5ej5vy
« Previous Showing results 1 — 15 out of 75,203 results