The Experts below are selected from a list of 88698 Experts worldwide ranked by ideXlab platform

Hongkai Xiong - One of the best experts on this subject based on the ideXlab platform.

  • pc darts partial channel connections for memory efficient architecture search
    International Conference on Learning Representations, 2020
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-net and searching for an optimal architecture. In this paper, we present a novel approach, namely Partially-Connected DARTS, by sampling a small part of super-net to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We solve it by introducing edge normalization, which adds a new set of edge-level hyper-parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoy both faster speed and higher training stability. Experiment results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 within merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) within 3.8 GPU-days for search. Our code has been made available at https://www.dropbox.com/sh/on9lg3rpx1r6dkf/AABG5mt0sMHjnEJyoRnLEYW4a?dl=0.

  • pc darts partial channel connections for memory efficient architecture search
    arXiv: Computer Vision and Pattern Recognition, 2019
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-network and searching for an optimal architecture. In this paper, we present a novel approach, namely, Partially-Connected DARTS, by sampling a small part of super-network to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We alleviate it using edge normalization, which adds a new set of edge-level parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoys both faster speed and higher training stability. Experimental results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 with merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) using 3.8 GPU-days for search. Our code has been made available at: this https URL.

  • pc darts partial channel connections for memory efficient differentiable architecture search
    2019
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-network and searching for an optimal architecture. In this paper, we present a novel approach, namely, Partially-Connected DARTS, by sampling a small part of super-network to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We alleviate it using edge normalization, which adds a new set of edge-level parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoys both faster speed and higher training stability. Experimental results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 with merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) using 3.8 GPU-days for search. Our code has been made available at: this https URL.

Lingxi Xie - One of the best experts on this subject based on the ideXlab platform.

  • pc darts partial channel connections for memory efficient architecture search
    International Conference on Learning Representations, 2020
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-net and searching for an optimal architecture. In this paper, we present a novel approach, namely Partially-Connected DARTS, by sampling a small part of super-net to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We solve it by introducing edge normalization, which adds a new set of edge-level hyper-parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoy both faster speed and higher training stability. Experiment results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 within merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) within 3.8 GPU-days for search. Our code has been made available at https://www.dropbox.com/sh/on9lg3rpx1r6dkf/AABG5mt0sMHjnEJyoRnLEYW4a?dl=0.

  • pc darts partial channel connections for memory efficient architecture search
    arXiv: Computer Vision and Pattern Recognition, 2019
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-network and searching for an optimal architecture. In this paper, we present a novel approach, namely, Partially-Connected DARTS, by sampling a small part of super-network to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We alleviate it using edge normalization, which adds a new set of edge-level parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoys both faster speed and higher training stability. Experimental results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 with merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) using 3.8 GPU-days for search. Our code has been made available at: this https URL.

  • pc darts partial channel connections for memory efficient differentiable architecture search
    2019
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-network and searching for an optimal architecture. In this paper, we present a novel approach, namely, Partially-Connected DARTS, by sampling a small part of super-network to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We alleviate it using edge normalization, which adds a new set of edge-level parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoys both faster speed and higher training stability. Experimental results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 with merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) using 3.8 GPU-days for search. Our code has been made available at: this https URL.

Tongseng Quah - One of the best experts on this subject based on the ideXlab platform.

  • a graphical operating environment for neural network expert systems
    International Joint Conference on Neural Network, 1991
    Co-Authors: Tongseng Quah
    Abstract:

    A window-based platform, known as the Graphical Environment for Neuronet Expert Systems (GENES), is proposed. The platform provides the user with an easy-to-learn, easy-to-use operating environment for creating, training, editing, and enhancing neural-network-based expert systems. The underlying neural logic network (NELONET) has been shown to be capable of doing logical inferencing and is used in two large-scale-Operation expert systems. Building on top of the X-window system and the OPENLOOK user interface, GENES inherits the select-and-Perform Operation strategy for neural network objects. The system's knowledge base contains simple network elements that correspond to rules in a conventional system. During the inference process, these network elements are linked up dynamically to form a large neural network which will operate according to the NELONET activation rules. >

Qi Tian - One of the best experts on this subject based on the ideXlab platform.

  • pc darts partial channel connections for memory efficient architecture search
    International Conference on Learning Representations, 2020
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-net and searching for an optimal architecture. In this paper, we present a novel approach, namely Partially-Connected DARTS, by sampling a small part of super-net to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We solve it by introducing edge normalization, which adds a new set of edge-level hyper-parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoy both faster speed and higher training stability. Experiment results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 within merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) within 3.8 GPU-days for search. Our code has been made available at https://www.dropbox.com/sh/on9lg3rpx1r6dkf/AABG5mt0sMHjnEJyoRnLEYW4a?dl=0.

  • PC-DARTS: Partial Channel Connections for Memory-Efficient Architecture Search
    2020
    Co-Authors: Xu Yuhui, Xie Lingxi, Zhang Xiaopeng, Chen Xin, Qi Guo-jun, Qi Tian, Xiong Hongkai
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-network and searching for an optimal architecture. In this paper, we present a novel approach, namely, Partially-Connected DARTS, by sampling a small part of super-network to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We alleviate it using edge normalization, which adds a new set of edge-level parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoys both faster speed and higher training stability. Experimental results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 with merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) using 3.8 GPU-days for search. Our code has been made available at: https://github.com/yuhuixu1993/PC-DARTS.Comment: Accepted by ICLR202

  • pc darts partial channel connections for memory efficient architecture search
    arXiv: Computer Vision and Pattern Recognition, 2019
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-network and searching for an optimal architecture. In this paper, we present a novel approach, namely, Partially-Connected DARTS, by sampling a small part of super-network to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We alleviate it using edge normalization, which adds a new set of edge-level parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoys both faster speed and higher training stability. Experimental results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 with merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) using 3.8 GPU-days for search. Our code has been made available at: this https URL.

  • pc darts partial channel connections for memory efficient differentiable architecture search
    2019
    Co-Authors: Lingxi Xie, Qi Tian, Xiaopeng Zhang, Xin Chen, Hongkai Xiong
    Abstract:

    Differentiable architecture search (DARTS) provided a fast solution in finding effective network architectures, but suffered from large memory and computing overheads in jointly training a super-network and searching for an optimal architecture. In this paper, we present a novel approach, namely, Partially-Connected DARTS, by sampling a small part of super-network to reduce the redundancy in exploring the network space, thereby Performing a more efficient search without comprising the Performance. In particular, we Perform Operation search in a subset of channels while bypassing the held out part in a shortcut. This strategy may suffer from an undesired inconsistency on selecting the edges of super-net caused by sampling different channels. We alleviate it using edge normalization, which adds a new set of edge-level parameters to reduce uncertainty in search. Thanks to the reduced memory cost, PC-DARTS can be trained with a larger batch size and, consequently, enjoys both faster speed and higher training stability. Experimental results demonstrate the effectiveness of the proposed method. Specifically, we achieve an error rate of 2.57% on CIFAR10 with merely 0.1 GPU-days for architecture search, and a state-of-the-art top-1 error rate of 24.2% on ImageNet (under the mobile setting) using 3.8 GPU-days for search. Our code has been made available at: this https URL.

Dalehaug, Kjell Harald - One of the best experts on this subject based on the ideXlab platform.

  • An analysis of drilling Operation efficiency
    University of Stavanger Norway, 2011
    Co-Authors: Dalehaug, Kjell Harald
    Abstract:

    Master's thesis in Petroleum engineeringSeadrill’s constant strive to achieve their goal in setting the standard in drilling pursues the focus on Operations excellence. The company’s fleet is expanding with frequent new Mobile Offshore Drilling Units (MODU) and this might involve challenges with keeping up Operations excellence from the first day. In this thesis the main object is to Perform Operation efficiency analyzes across three of Seadrill’s semisubmersible rigs; West Venture, West Phoenix and West Eminence. The chosen Operations for analyses are: · Unrestricted tripping of drillpipe in cased hole · Unrestricted running of casing into the well · Running and pulling BOP The analyses were based on Daily Operations Database Application (DODA) reports, which is Seadrill’s daily Operation reporting system. The reports contain specified amount of hours spent to Perform the Operation and the tripping distance, and tripping rates were determined and compared with the rigs. In the BOP Operations the analyses emphasizes in addition on Operation efficiency when placing BOP on wellhead and unlatch BOP from wellhead including the related Operations. The well data are based on exploration wells, except from the first analysis which contain Operations from other wells

  • An analysis of drilling Operation efficiency
    University of Stavanger Norway, 2011
    Co-Authors: Dalehaug, Kjell Harald
    Abstract:

    Seadrill’s constant strive to achieve their goal in setting the standard in drilling pursues the focus on Operations excellence. The company’s fleet is expanding with frequent new Mobile Offshore Drilling Units (MODU) and this might involve challenges with keeping up Operations excellence from the first day. In this thesis the main object is to Perform Operation efficiency analyzes across three of Seadrill’s semisubmersible rigs; West Venture, West Phoenix and West Eminence. The chosen Operations for analyses are: · Unrestricted tripping of drillpipe in cased hole · Unrestricted running of casing into the well · Running and pulling BOP The analyses were based on Daily Operations Database Application (DODA) reports, which is Seadrill’s daily Operation reporting system. The reports contain specified amount of hours spent to Perform the Operation and the tripping distance, and tripping rates were determined and compared with the rigs. In the BOP Operations the analyses emphasizes in addition on Operation efficiency when placing BOP on wellhead and unlatch BOP from wellhead including the related Operations. The well data are based on exploration wells, except from the first analysis which contain Operations from other wells