检索结果-内蒙古大学图书馆

arXiv 2017年

作者： Mishra, Asit Marr, Debbie Accelerator Architecture Lab Intel Labs

Deep learning networks have achieved state-of-the-art accuracies on computer vision workloads like image classification and object detection. The performant systems, however, typically involve big models with numerous parameters. Once trained, a challenging aspect for such top performing models is deployment on resource constrained inference systems - the models (often deep networks or wide networks or both) are compute and memory intensive. Low-precision numerics and model compression using knowledge distillation are popular techniques to lower both the compute requirements and memory footprint of these deployed models. In this paper, we study the combination of these two techniques and show that the performance of low-precision networks can be significantly improved by using knowledge distillation techniques. Our approach, Apprentice, achieves state-of-the-art accuracies using ternary precision and 4-bit precision for variants of ResNet architecture on ImageNet dataset. We present three schemes using which one can apply knowledge distillation techniques to various stages of the train-and-deploy pipeline. Copyright © 2017, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Low Precision RNNs: Quantizing RNNs Without Losing Accuracy

arXiv

引用

arXiv 2017年

作者： Kapur, Supriya Mishra, Asit Marr, Debbie Accelerator Architecture Lab Intel Corporation

Similar to convolution neural networks, recurrent neural networks (RNNs) typically suffer from over-parameterization. Quantizing bit-widths of weights and activations results in runtime efficiency on hardware, yet it often comes at the cost of reduced accuracy. This paper proposes a quantization approach that increases model size with bit-width reduction. This approach will allow networks to perform at their baseline accuracy while still maintaining the benefits of reduced precision and overall model size reduction. Copyright © 2017, The Authors. All rights reserved.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

WRPN: Training and inference using wide reduced-precision networks

arXiv

引用

arXiv 2017年

作者： Mishra, Asit Cook, Jeffrey J. Nurvitadhi, Eriko Marr, Debbie Accelerator Architecture Lab Intel Labs

For computer vision applications, prior works have shown the efficacy of reducing numeric precision of model parameters (network weights) in deep neural networks but also that reducing the precision of activations hurts model accuracy much more than reducing the precision of model parameters. We study schemes to train networks from scratch using reduced-precision activations without hurting the model accuracy. We reduce the precision of activation maps (along with model parameters) using a novel quanti-zation scheme and increase the number of filter maps in a layer, and find that this scheme compensates or surpasses the accuracy of the baseline full-precision network. As a result, one can significantly reduce the dynamic memory footprint, memory bandwidth, computational energy and speed up the training and inference process with appropri- A te hardware support. We call our scheme WRPN-wide reduced-precision networks. We report results using our proposed schemes and show that our results are better than previously reported accuracies on ILSVRC-12 dataset while being computationally less expensive compared to previ-ously reported reduced-precision networks. Copyright © 2017, The Authors. All rights reserved.

关键词： Chemical activation

来源：评论

学校读者我要写书评

暂无评论

Accelerating Deep Convolutional Networks using low-precision and sparsity

Accelerating Deep Convolutional Networks using low-precision...

引用

IEEE International Conference on Acoustics, Speech and Signal Processing

作者： Ganesh Venkatesh Eriko Nurvitadhi Debbie Marr Accelerator Architecture Lab Intel Corporation United States of America

ISBN: (纸本)9781509041183

We explore techniques to significantly improve the compute efficiency and performance of Deep Convolution Networks without impacting their accuracy. To improve the compute efficiency, we focus on achieving high accuracy with extremely low-precision (2-bit) weight networks, and to accelerate the execution time, we aggressively skip operations on zero-values. We achieve the highest reported accuracy of 76.6% Top-1/93% Top-5 on the Imagenet object classification challenge with low-precision network while reducing the compute requirement by ～3× compared to a full-precision network that achieves similar accuracy. Furthermore, to fully exploit the benefits of our low-precision networks, we build a deep learning accelerator core, DLAC, that can achieve up to 1 TFLOP/mm~2 equivalent for single-precision floating-point operations (～2 TFLOP/mm~2 for half-precision), which is ～5× better than Linear Algebra Core [16] and ～4× better than previous deep learning accelerator proposal [8].

关键词： Deep Neural Networks Ternary-weight Convolutions accelerator

来源：评论

学校读者我要写书评

暂无评论

Search for high-mass exclusive γγ → WW and γγ → ZZ production in proton-proton collisions at √s=13 TeV

引用

JOURNAL OF HIGH ENERGY PHYSICS 2023年第7期2023卷 229页

作者： Tumasyan, A. Adam, W. Andrejkovic, J. W. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Del Valle, A. Escalante Hussain, P. S. Jeitler, M. Krammer, N. Lechner, L. Liko, D. Mikulec, I. Paulitsch, P. Pitters, F. M. Schieck, J. Schoefbeck, R. Schwarz, D. Templ, S. Waltenberger, W. Wulz, C. -E. Darwish, M. R. Janssen, T. Kello, T. Rejeb Sfar, H. Van Mechelen, P. Bols, E. S. D'Hondt, J. De Moor, A. Delcourt, M. El Faham, H. Lowette, S. Moortgat, S. Morton, A. Mueller, D. Sahasransu, A. R. Tavernier, S. Van Doninck, W. Vannerom, D. Clerbaux, B. De Lentdecker, G. Favart, L. Hohov, D. Jaramillo, J. Lee, K. Mahdavikhorrami, M. Makarenko, I. Malara, A. Paredes, S. Petre, L. Postiau, N. Starling, E. Thomas, L. Vanden Bemden, M. Vander Velde, C. Vanlaer, P. Dobur, D. Knolle, J. Lambrecht, L. Mestdach, G. Niedziela, M. Rendon, C. Roskas, C. Samalan, A. Skovpen, K. Tytgat, M. Van Den Bossche, N. Vermassen, B. Wezenbeek, L. Benecke, A. Bruno, G. Bury, F. Caputo, C. David, P. Delaere, C. Donertas, I. S. Giammanco, A. Jaffel, K. Jain, Sa. Lemaitre, V. Mondal, K. Prisciandaro, J. Taliercio, A. Tran, T. T. Vischia, P. Wertz, S. Alves, G. A. Coelho, E. Hensel, C. Moraes, A. Teles, P. Rebello Alda Junior, W. L. Pereira, M. Alves Gallo Filho, M. Barroso Ferreira Malbouisson, H. Brandao Carvalho, W. Chinellato, J. Da Costa, E. M. Da Silveira, G. G. De Jesus Damiao, D. Dos Santos Sousa, V. De Souza, S. Fonseca Martins, J. Mora Herrera, C. Mota Amarilo, K. Mundim, L. Nogima, H. Santoro, A. Do Amaral, S. M. Silva Sznajder, A. Thiel, M. De Araujo, F. Torres Da Silva Pereira, A. Vilela Bernardes, C. A. Calligaris, L. Tomei, T. R. Fernandez Perez Gregores, E. M. Mercadante, P. G. Novaes, S. F. Padula, Sandra S. Aleksandrov, A. Hadjiiska, R. Iaydjiev, P. Misheva, M. Rodozov, M. Shopova, M. Sultanov, G. Dimitrov, A. Ivanov, T. Litov, L. Pavlov, B. Petkov, P. Petrov, A. Shumka, E. Cheng, T. Javaid, T. Mittal, M. Yuan, L. Ahmad, M. Bauer, G. Hu, Z. Lezki, S. Yi, K. Chen, G. M. Chen, H. S. Chen, M. Iemmi, F. Jiang, C. H. Kapoor, A. Liao Yerevan Phys Inst Yerevan Armenia Inst Hochenergiephys Vienna Austria Univ Antwerp Antwerp Belgium Vrije Univ Brussel Brussels Belgium Univ Libre Bruxelles Brussels Belgium Univ Ghent Ghent Belgium Catholic Univ Louvain Louvain La Neuve Belgium Ctr Brasileiro Pesquisas Fis Rio De Janeiro Brazil Univ Estado Rio De Janeiro Rio De Janeiro Brazil Fed Univ ABC Univ Estadual Paulista Sao Paulo Brazil Bulgarian Acad Sci Inst Nucl Res & Nucl Energy Sofia Bulgaria Univ Sofia Sofia Bulgaria Beihang Univ Beijing Peoples R China Tsinghua Univ Dept Phys Beijing Peoples R China Inst High Energy Phys Beijing Peoples R China Peking Univ State Key Lab Nucl Phys & Technol Beijing Peoples R China Sun Yat Sen Univ Guangzhou Peoples R China Fudan Univ Inst Modern Phys Shanghai Peoples R China Fudan Univ Key Lab Nucl Phys & Ionbeam Applicat MOE Shanghai Peoples R China Zhejiang Univ Hangzhou Zhejiang Peoples R China Univ Los Andes Bogota Colombia Univ Antioquia Medellin Colombia Univ Split Fac Elect Engn Mech Engn & Naval Architecture Split Croatia Univ Split Fac Sci Split Croatia Inst Rudjer Boskov Zagreb Croatia Univ Cyprus Nicosia Cyprus Charles Univ Prague Prague Czech Republic Escuela Politec Nacl Quito Ecuador Univ San Francisco Quito Quito Ecuador Acad Sci Res & Technol Arab Republ Egypt Egyptian Network High Energy Phys Cairo Egypt Fayoum Univ Ctr High Energy Phys CHEP FU Al Fayyum Egypt NICPB Tallinn Estonia Univ Helsinki Dept Phys Helsinki Finland Helsinki Inst Phys Helsinki Finland Lappeenranta Lahti Univ Technol Lappeenranta Finland Univ Paris Saclay IRFU CEA Gif Sur Yvette France Inst Polytech Paris Lab Leprince Ringuet CNRS Ecole PolytechIN2P3 Palaiseau France Univ Strasbourg CNRS IPHC UMR 7178 Strasbourg France Inst Phys 2 Infinis Lyon IP2I Villeurbanne France Georgian Tech Univ Tbilisi Georgia Rhein Westfal TH Aachen Phys Inst 1 Aachen Germany Rhein Westfal TH Aachen Phys Inst 3 Aachen Germany Rhein Westfal TH A

A search is performed for exclusive high-mass gamma gamma -> WW and gamma gamma -> ZZ production in proton-proton collisions using intact forward protons reconstructed in near-beam detectors, with both weak bosons decaying into boosted and merged jets. The analysis is based on a sample of proton-proton collisions collected by the CMS and TOTEM experiments at root s = 13 TeV, corresponding to an integrated luminosity of 100 fb(-1). No excess above the standard model background prediction is observed, and upper limits are set on the pp -> pWWp and pp -> pZZp cross sections in a fiducial region defined by the diboson invariant mass m(VV) > 1 TeV (with V = W, Z) and proton fractional momentum loss 0.04 < xi < 0.20. The results are interpreted as new limits on dimension-6 and dimension-8 anomalous quartic gauge couplings.

关键词： Forward Physics Hadron-Hadron Scattering Vector Boson Production

来源：评论

学校读者我要写书评

暂无评论

Performance of CMS muon reconstruction from proton-proton to heavy ion collisions

引用

JOURNAL OF INSTRUMENTATION 2024年第9期19卷 P09012页

作者： Hayrapetyan, A. Tumasyan, A. Adam, W. Andrejkovic, J. W. Arnold, B. Bergauer, H. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Del Valle, A. Escalante Hussain, P. S. Jeitler, M. Krammer, N. Liko, D. Mikulec, I. Schieck, J. Schoefbeck, R. Schwarz, D. Sonawane, M. Templ, S. Waltenberger, W. Wulz, C-E Darwish, M. R. Janssen, T. Van Mechelen, P. Bols, E. S. D'Hondt, J. Dansana, S. De Moor, A. Delcourt, M. El Faham, H. Lowette, S. Makarenko, I. Morton, A. Muller, D. Sahasransu, A. R. Tavernier, S. Tytgat, M. Van Putte, S. Vannerom, D. Clerbaux, B. De Lentdecker, G. Favart, L. Hohov, D. Jaramillo, J. Khalilzadeh, A. Lee, K. Mahdavikhorrami, M. Malara, A. Paredes, S. Petre, L. Postiau, N. Thomas, L. Vanden Bemden, M. Vander Velde, C. Vanlaer, P. De Coen, M. Dobur, D. Knolle, J. Lambrecht, L. Mestdach, G. Rendon, C. Samalan, A. Skovpen, K. Van den Bossche, N. Vermassen, B. Wezenbeek, L. Benecke, A. Bruno, G. Caputo, C. Delaere, C. Donertas, I. S. Giammanco, A. Jaffel, K. Jain, Sa. Lemaitre, V. Lidrych, J. Mastrapasqua, P. Mondal, K. Tran, T. T. Wertz, S. Alves, G. A. Coelho, E. Hensel, C. Menezes De Oliveira, T. Moraes, A. Rebello Teles, P. Soeiro, M. Alda Junior, W. L. Alves Gallo Pereira, M. Barroso Ferreira Filho, M. Brandao Malbouisson, H. Carvalho, W. Chinellato, J. Da Costa, E. M. Da Silveira, G. G. De Jesus Damiao, D. Fonseca De Souza, S. Martins, J. Mora Herrera, C. Mota Amarilo, K. Mundim, L. Nogima, H. Santoro, A. Silva Do Amaral, S. M. Sznajder, A. Thiel, M. Vilela Pereira, A. Bernardes, C. A. Calligaris, L. Fernandez Perez Tomei, T. R. Gregores, E. M. Mercadante, P. G. Novaes, S. F. Orzari, B. Padula, Sandra S. Aleksandrov, A. Antchev, G. Hadjiiska, R. Iaydjiev, P. Misheva, M. Shopova, M. Sultanov, G. Dimitrov, A. Ivanov, T. Litov, L. Pavlov, B. Petkov, P. Petrov, A. Shumka, E. Keshri, S. Thakur, S. Cheng, T. Guo, Q. Javaid, T. Mittal, M. Yuan, L. Bauer, G. Hu, Z. Yi, K. Chen, G. M. Chen, H. S. Chen, M. Iemmi, F. Jiang, C. H. Kapoor, A. Liao, H. Liu, Z-A Monti, F. Sharma, R. Song, J. N. Tao, J. Wan Yerevan Phys Inst Yerevan Armenia Inst Hochenergiephys Vienna Austria Univ Antwerp Antwerp Belgium Vrije Univ Brussel Brussels Belgium Univ Libre Bruxelles Brussels Belgium Univ Ghent Ghent Belgium Univ Ghent Ghent Belgium Catholic Univ Louvain Louvain La Neuve Belgium Ctr Brasileiro Pesquisas Fis Rio De Janeiro Brazil Univ Estado Rio de Janeiro Rio De Janeiro Brazil Univ Fed ABC Univ Estadual Paulista Sao Paulo Brazil Bulgarian Acad Sci Inst Nucl Res & Nucl Energy Sofia Bulgaria Univ Sofia Sofia Bulgaria Univ Tarapaca Inst Alta Invest Casilla 7 D Arica Chile Beihang Univ Beijing Peoples R China Tsinghua Univ Dept Phys Beijing Peoples R China Inst High Energy Phys Beijing Peoples R China Peking Univ State Key Lab Nucl Phys & Technol Beijing Peoples R China Sun Yat Sen Univ Guangzhou Peoples R China Univ Sci & Technol China Hefei Peoples R China Fudan Univ Inst Modern Phys Shanghai Peoples R China Fudan Univ Key Lab Nucl Phys & Ion Beam Applicat MOE Shanghai Peoples R China Zhejiang Univ Hangzhou Zhejiang Peoples R China Univ Los Andes Bogota Colombia Univ Antioquia Medellin Colombia Univ Split Fac Elect Engn Mech Engn & Naval Architecture Split Croatia Univ Split Fac Sci Split Croatia Inst Rudjer Boskovic Zagreb Croatia Univ Cyprus Nicosia Cyprus Charles Univ Prague Prague Czech Republic Escuela Politec Nacl Quito Ecuador Univ San Francisco Quito Quito Ecuador Acad Sci Res & Technol Arab Republ Egypt Egyptian Network High Energy Phys Cairo Egypt Fayoum Univ Ctr High Energy Phys Al Fayyum Egypt NICPB Tallinn Estonia Univ Helsinki Dept Phys Helsinki Finland Helsinki Inst Phys Helsinki Finland Lappeenranta Lahti Univ Technol Lappeenranta Finland Univ Paris Saclay IRFU CEA Gif Sur Yvette France Ecole Polytech IN2P3 CNRS Lab Leprince RinguetInst Polytech Paris Palaiseau France Univ Strasbourg CNRS IPHC UMR 7178 Strasbourg France Inst Phys 2 Infinis Lyon IP2I Villeurbanne France Georgian Tech Univ Tbilisi Georgia

The performance of muon tracking, identification, triggering, momentum resolution, and momentum scale has been studied with the CMS detector at the LHC using data collected at root s(NN) = 5.02 TeV in proton-proton (pp) and lead-lead (PbPb) collisions in 2017 and 2018, respectively, and at root s(NN) = 8.16 TeV in proton-lead (pPb) collisions in 2016. Muon efficiencies, momentum resolutions, and momentum scales are compared by focusing on how the muon reconstruction performance varies from relatively small occupancy pp collisions to the larger occupancies of pPb collisions and, finally, to the highest track multiplicity PbPb collisions. We find the efficiencies of muon tracking, identification, and triggering to be above 90% throughout most of the track multiplicity range. The momentum resolution and scale are unaffected by the detector occupancy. The excellent muon reconstruction of the CMS detector enables precision studies across all available collision systems.

关键词： Instrumentation and methods for heavy-ion reactions and fission studies Large detector systems for particle and astroparticle physics Muon spectrometers

来源：评论

学校读者我要写书评

暂无评论

Measurement of the production cross section for a W boson in association with a charm quark in proton-proton collisions at √s=13 TeV

引用

EUROPEAN PHYSICAL JOURNAL C 2024年第1期84卷 27页

作者： Tumasyan, A. Adam, W. Andrejkovic, J. W. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Del Valle, A. Escalante Hussain, P. S. Jeitler, M. Krammer, N. Lechner, L. Liko, D. Mikulec, I. Paulitsch, P. Pitters, F. M. Schieck, J. Schoefbeck, R. Schwarz, D. Sonawane, M. Templ, S. Waltenberger, W. Wulz, C-E Darwish, M. R. Janssen, T. Kello, T. Sfar, H. Rejeb Van Mechelen, P. Bols, E. S. D'Hondt, J. De Moor, A. Delcourt, M. El Faham, H. Lowette, S. Moortgat, S. Morton, A. Mueller, D. Sahasransu, A. R. Tavernier, S. Van Doninck, W. Vannerom, D. Clerbaux, B. De Lentdecker, G. Favart, L. Hohov, D. Jaramillo, J. Lee, K. Mahdavikhorrami, M. Makarenko, I. Malara, A. Paredes, S. Petre, L. Postiau, N. Thomas, L. Vanden Bemden, M. Vander Velde, C. Vanlaer, P. Dobur, D. Knolle, J. Lambrecht, L. Mestdach, G. Rendon, C. Samalan, A. Skovpen, K. Tytgat, M. Van den Bossche, N. Vermassen, B. Wezenbeek, L. Benecke, A. Bruno, G. Bury, F. Caputo, C. David, P. Delaere, C. Donertas, I. S. Giammanco, A. Jaffel, K. Jain, Sa. Lemaitre, V. Mondal, K. Taliercio, A. Tran, T. T. Vischia, P. Wertz, S. Alves, G. A. Coelho, E. Hensel, C. Moraes, A. Rebello Teles, P. Alda Junior, W. L. Alves Gallo Pereira, M. Barroso Ferreira Filho, M. Brandao Malbouisson, H. Carvalho, W. Chinellato, J. Da Costa, E. M. Da Silveira, G. G. De Jesus Damiao, D. Dos Santos Sousa, V. Fonseca De Souza, S. Martins, J. Mora Herrera, C. Mota Amarilo, K. Mundim, L. Nogima, H. Santoro, A. Silva Do Amaral, S. M. Sznajder, A. Thiel, M. Vilela Pereira, A. Bernardes, C. A. Calligaris, L. Fernandez Perez Tomei, T. R. Gregores, E. M. Mercadante, P. G. Novaes, S. F. Padula, Sandra S. Aleksandrov, A. Antchev, G. Hadjiiska, R. Iaydjiev, P. Misheva, M. Rodozov, M. Shopova, M. Sultanov, G. Dimitrov, A. Ivanov, T. Litov, L. Pavlov, B. Petkov, P. Petrov, A. Shumka, E. Thakur, S. Cheng, T. Javaid, T. Mittal, M. Yuan, L. Ahmad, M. Bauer, G. Hu, Z. Lezki, S. Yi, K. Chen, G. M. Chen, H. S. Chen, M. Iemmi, F. Jiang, C. H. Kapoor, A. Liao, H. Liu, Z-A Milosevic, V. Monti, F. Sharma, R. Tao Yerevan Phys Inst Yerevan Armenia Inst Hochenergiephys Vienna Austria Univ Antwerp Antwerp Belgium Vrije Univ Brussel Brussels Belgium Univ Libre Bruxelles Brussels Belgium Univ Ghent Ghent Belgium Catholic Univ Louvain Louvain La Neuve Belgium Ctr Brasileiro Pesquisas Fis Rio De Janeiro Brazil Univ Estado Rio de Janeiro Rio De Janeiro Brazil Univ Fed ABC Univ Estadual Paulista Sao Paulo Brazil Bulgarian Acad Sci Inst Nucl Res & Nucl Energy Sofia Bulgaria Univ Sofia Sofia Bulgaria Univ Tarapaca Inst Alta Invest Casilla 7 D Arica Chile Beihang Univ Beijing Peoples R China Tsinghua Univ Dept Phys Beijing Peoples R China Inst High Energy Phys Beijing Peoples R China Peking Univ State Key Lab Nucl Phys & Technol Beijing Peoples R China Sun Yat Sen Univ Guangzhou Peoples R China Univ Sci & Technol China Hefei Peoples R China Fudan Univ Inst Modern Phys Shanghai Peoples R China Fudan Univ Key Lab Nucl Phys & Ion Beam Applicat MOE Shanghai Peoples R China Zhejiang Univ Hangzhou Zhejiang Peoples R China Univ Los Andes Bogota Colombia Univ Antioquia Medellin Colombia Univ Split Fac Elect Engn Mech Engn & Naval Architecture Split Croatia Univ Split Fac Sci Split Croatia Inst Rudjer Boskovic Zagreb Croatia Univ Cyprus Nicosia Cyprus Charles Univ Prague Prague Czech Republic Escuela Politec Nacl Quito Ecuador Univ San Francisco Quito Quito Ecuador Acad Sci Res & Technol Arab Republ Egypt Egyptian Network High Energy Phys Cairo Egypt Fayoum Univ Ctr High Energy Phys CHEP FU Al Fayyum Egypt NICPB Tallinn Estonia Univ Helsinki Dept Phys Helsinki Finland Helsinki Inst Phys Helsinki Finland Lappeenranta Lahti Univ Technol Lappeenranta Finland Univ Paris Saclay IRFU CEA Gif Sur Yvette France Inst Polytech Paris CNRS IN2P3 Lab Leprince RinguetEcole Polytech Palaiseau France Univ Strasbourg CNRS IPHC UMR 7178 Strasbourg France Inst Phys 2 Infinis Lyon IP2I Villeurbanne France Georgian Tech Univ Tbilisi Georgia Rhein Westfal TH A

The strange quark content of the proton is probed through the measurement of the production cross section for a W boson and a charm (c) quark in proton-proton collisions at a center-of-mass energy of 13 TeV. The analysis uses a data sample corresponding to a total integrated luminosity of 138 fb(-1) collected with the CMS detector at the LHC. The W bosons are identified through their leptonic decays to an electron or a muon, and a neutrino. Charm jets are tagged using the presence of a muon or a secondary vertex inside the jet. The W + c production cross section and the cross section ratio R-c(+/-) = sigma(W+ + (c) over bar)/sigma(W- + c) are measured inclusively and differentially as functions of the transverse momentum and the pseudorapidity of the lepton originating from the W boson decay. The precision of the measurements is improved with respect to previous studies, reaching 1% in R-c(+/-) = 0.950 +/- 0.005 (stat)+/- 0.010 (syst). The measurements are compared with theoretical predictions up to next-to-next-to-leading order in perturbative quantum chromodynamics.

关键词： Bosons

来源：评论

学校读者我要写书评

暂无评论

High performance binary neural networks on the Xeon+FPGA™ platform

High performance binary neural networks on the Xeon+FPGA™ p...

引用

International Conference on Field Programmable Logic and Applications

作者： Duncan J. M. Moss Eriko Nurvitadhi Jaewoong Sim Asit Mishra Debbie Marr Suchit Subhaschandra Philip H. W. Leong FPGA Product Team Intel Corporation The University of Sydney Accelerator Architecture Lab

Convolutional neural networks (CNNs) are deployed in a wide range of image recognition, scene segmentation and object detection applications. Achieving state of the art accuracy in CNNs often results in large models and complex topologies that require significant compute resources to complete in a timely manner. Binarised neural networks (BNNs) have been proposed as an optimised variant of CNNs, which constrain the weights and activations to +1 or -1 and thus offer compact models and lower computational complexity per operation. This paper presents a high performance BNN accelerator on the Intel®Xeon+FPGA™ platform. The proposed accelerator is designed to take advantage of the Xeon+FPGA system in a way that a specialised FPGA architecture can be targeted for the most compute intensive parts of the BNN whilst other parts of the topology can be handled by the Xeon™ CPU. The implementation is evaluated by comparing the raw compute performance and energy efficiency for key layers in standard CNN topologies against an Nvidia Titan X Pascal GPU and other published FPGA BNN accelerators. The results show that our single-package integrated Arria™ 10 FPGA accelerator coupled with a high-end Xeon CPU can offer comparable performance and better energy efficiency than a high-end discrete Titan X GPU card. In addition, our solution delivers the best performance compared to previous BNN FPGA implementations.

关键词： Field programmable gate arrays Graphics processing units Topology Computer architecture Performance evaluation Network topology IP networks

来源：评论

学校读者我要写书评

暂无评论

Low Temperature Nanoscale Oxygen-Ion Intercalation into Epitaxial MoO₂ Thin Films

引用

JOURNAL OF PHYSICAL CHEMISTRY C 2017年第6期121卷 3410-3415页

作者： Ahn, Eun Young Lee, Joonhyuk Koh, Yoon Young Lee, Jaekwang Park, Byeong-Gyu Kim, Jae-Young Lee, Inwon Lee, Chan-Woo Jeen, Hyoungjeen Pusan Natl Univ Dept Phys Busan 46241 South Korea Pohang Univ Sci & Technol Pohang Accelerator Lab Pohang 37673 South Korea Pusan Natl Univ Dept Naval Architecture & Ocean Engn Busan 46241 South Korea Korea Inst Energy Res Energy Mat & Proc Res Div Convers Mat Lab Daejeon 34129 South Korea

In transition metal oxides (TMOs), lots of physical phenomena such as metal-insulator transitions (MIT), magnetism, and ferroelectricity are closely related to the amounts of oxygen contents. Thus, understanding surface oxidation process in TMOs and its effect are important for enhancing performances of modern electronic and electrochemical devices due to miniaturization of those devices. In this regard, MoO2+x (0 <= x <= 1) is an interesting TMO, which shows MIT driven by the change of its oxygen content, i.e. metallic MoO2 and insulating MoO3. Hence, understanding thermally driven oxygen intercalation into MoO2 is very important. In this work, we conducted in situ postannealing of as-grown epitaxial MoO2 thin films at different temperatures in oxidative condition to investigate the thermal effect on oxygen ion intercalation and resultant MIT in MoO2+x. Through the spectroscopic techniques such as spectroscopic ellipsometry and X-ray absorption spectroscopy, we observed that oxygen ions can intercalate into MoO2 and trigger a phase transition in nanoscale at surprisingly low-temperature as low as 250 C. In addition, after oxygen annealing at 350 C, we find that both hybridization and interband transition energy between O 2p and Mo 4d t(2g) are significantly shifted to low energy nearly 0.2 eV, which clearly supports that the electronic transition of MoO2+x is predominantly driven by change of oxygen contents.

关键词： Thin films Oxygen Oxidation Electrical conductivity Intercalation

来源：评论

学校读者我要写书评

暂无评论

Search for exotic decays of the Higgs boson to a pair of pseudoscalars in the μμbb and ττbb final states

引用

EUROPEAN PHYSICAL JOURNAL C 2024年第5期84卷 493页

作者： Hayrapetyan, A. Tumasyan, A. Adam, W. Andrejkovic, J. W. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Del Valle, A. Escalante Hussain, P. S. Jeitler, M. Krammer, N. Liko, D. Mikulec, I. Schieck, J. Schoefbeck, R. Schwarz, D. Sonawane, M. Templ, S. Waltenberger, W. Wulz, C. -E. Darwish, M. R. Janssen, T. Van Mechelen, P. Bols, E. S. D'Hondt, J. Dansana, S. De Moor, A. Delcourt, M. El Faham, H. Lowette, S. Makarenko, I. Mueller, D. Sahasransu, A. R. Tavernier, S. Tytgat, M. Van Putte, S. Vannerom, D. Clerbaux, B. De Lentdecker, G. Favart, L. Hohov, D. Jaramillo, J. Khalilzadeh, A. Lee, K. Mahdavikhorrami, M. Malara, A. Paredes, S. Petre, L. Postiau, N. Thomas, L. Vanden Bemden, M. Vander Velde, C. Vanlaer, P. De Coen, M. Dobur, D. Hong, Y. Knolle, J. Lambrecht, L. Mestdach, G. Rendon, C. Samalan, A. Skovpen, K. Van Den Bossche, N. Wezenbeek, L. Benecke, A. Bruno, G. Caputo, C. Delaere, C. Donertas, I. S. Giammanco, A. Jaffel, K. Jain, Sa. Lemaitre, V. Lidrych, J. Mastrapasqua, P. Mondal, K. Tran, T. T. Wertz, S. Alves, G. A. Coelho, E. Hensel, C. Menezes De Oliveira, T. Moraes, A. Rebello Teles, P. Soeiro, M. Alda Junior, W. L. Alves Gallo Pereira, M. Barroso Ferreira Filho, M. Brandao Malbouisson, H. Carvalho, W. Chinellato, J. Da Costa, E. M. Da Silveira, G. G. De Jesus Damiao, D. Fonseca De Souza, S. Martins, J. Mora Herrera, C. Mota Amarilo, K. Mundim, L. Nogima, H. Santoro, A. Silva Do Amaral, S. M. Sznajder, A. Thiel, M. Vilela Pereira, A. Bernardes, C. A. Calligaris, L. Fernandez Perez Tomei, T. R. Gregores, E. M. Mercadante, P. G. Novaes, S. F. Orzari, B. Padula, Sandra S. Aleksandrov, A. Antchev, G. Hadjiiska, R. Iaydjiev, P. Misheva, M. Shopova, M. Sultanov, G. Dimitrov, A. Ivanov, T. Litov, L. Pavlov, B. Petkov, P. Petrov, A. Shumka, E. Keshri, S. Thakur, S. Cheng, T. Guo, Q. Javaid, T. Mittal, M. Yuan, L. Bauer, G. Hu, Z. Yi, K. Chen, G. M. Chen, H. S. Chen, M. Iemmi, F. Jiang, C. H. Kapoor, A. Liao, H. Liu, Z. -A. Monti, F. Shahzad, M. A. Sharma, R. Song, J. N. Tao, J. Wang, C. Wang, J. Wan Yerevan Phys Inst Yerevan Armenia Inst Hochenergiephys Vienna Austria Univ Antwerp Antwerp Belgium Vrije Univ Brussel Brussels Belgium Univ Libre Bruxelles Brussels Belgium Univ Ghent Ghent Belgium Catholic Univ Louvain Louvain La Neuve Belgium Ctr Brasileiro Pesquisas Fis Rio De Janeiro Brazil Univ Estado Rio de Janeiro Rio De Janeiro Brazil Univ Fed ABC Univ Estadual Paulista Sao Paulo Brazil Bulgarian Acad Sci Inst Nucl Res & Nucl Energy Sofia Bulgaria Univ Sofia Sofia Bulgaria Univ Tarapaca Inst Alta Invest Casilla 7 D Arica Chile Beihang Univ Beijing Peoples R China Tsinghua Univ Dept Phys Beijing Peoples R China Inst High Energy Phys Beijing Peoples R China Peking Univ State Key Lab Nucl Phys & Technol Beijing Peoples R China Sun Yat Sen Univ Guangzhou Peoples R China Univ Sci & Technol China Hefei Peoples R China Fudan Univ Inst Modern Phys Shanghai Peoples R China Fudan Univ Key Lab Nucl Phys & Ion Beam Applicat MOE Shanghai Peoples R China Zhejiang Univ Hangzhou Zhejiang Peoples R China Univ Los Andes Bogota Colombia Univ Antioquia Medellin Colombia Univ Split Fac Elect Engn Mech Engn & Naval Architecture Split Croatia Univ Split Fac Sci Split Croatia Inst Rudjer Boskovic Zagreb Croatia Univ Cyprus Nicosia Cyprus Charles Univ Prague Prague Czech Republic Escuela Politec Nacl Quito Ecuador Univ San Francisco Quito Quito Ecuador Acad Sci Res & Technol Arab Republ Egypt Egyptian Network High Energy Phys Cairo Egypt Fayoum Univ Ctr High Energy Phys CHEP FU Al Fayyum Egypt NICPB Tallinn Estonia Univ Helsinki Dept Phys Helsinki Finland Helsinki Inst Phys Helsinki Finland Lappeenranta Lahti Univ Technol Lappeenranta Finland Univ Paris Saclay CEA IRFU Gif Sur Yvette France Ecole Polytech Inst Polytech Paris CNRS Lab Leprince RinguetIN2P3 Palaiseau France Univ Strasbourg CNRS IPHC UMR 7178 Strasbourg France Inst Phys 2 Infinis Lyon IP2I Villeurbanne France Georgian Tech Univ Tbilisi Georgia Rhein Westfal TH A

A search for exotic decays of the Higgs boson (H) with a mass of 125 GeV to a pair of light pseudoscalars a(1) is performed in final states where one pseudoscalar decays to two b quarks and the other to a pair of muons or tau leptons. A data sample of proton-proton collisions at root s = 13 TeV corresponding to an integrated luminosity of 138 fb(-1) recorded with the CMS detector is analyzed. No statistically significant excess is observed over the standard model backgrounds. Upper limits are set at 95% confidence level (CL) on the Higgs boson branching fraction to mu mu bb and to tau tau bb, via a pair of a(1)s. The limits depend on the pseudoscalar mass m(a1) and are observed to be in the range (0.17-3.3) x10(-4) and (1.7-7.7) x10(-2) in the mu mu bb and tau tau bb final states, respectively. In the framework of models with two Higgs doublets and a complex scalar singlet (2HDM+S), the results of the two final states are combined to determine upper limits on the branching fraction B(H -> a(1)a(1) -> llbb) at 95% CL, with l being a muon or a tau lepton. For different types of 2HDM+S, upper bounds on the branching fraction B(H -> a(1)a(1)) are extracted from the combination of the two channels. In most of the Type II 2HDM+S parameter space, B(H -> a(1)a(1)) values above 0.23 are excluded at 95% CL for m(a1) values between 15 and 60 GeV.

关键词： Scattering parameters

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：