Learning A Deep l∞ Encoder for Hashing

Wang Zhangyang, Yang Yingzhen, Chang Shiyu, Ling Qing, Thomas S. Huang

Research output: Contribution to journalConference article

Abstract

We investigate the l∞-constrained representation which demonstrates robustness to quantization errors, utilizing the tool of deep learning. Based on the Alternating Direction Method of Multipliers (ADMM), we formulate the original convex minimization problem as a feed-forward neural network, named Deep l∞ Encoder, by introducing the novel Bounded Linear Unit (BLU) neuron and modeling the Lagrange multipliers as network biases. Such a structural prior acts as an effective network regularization, and facilitates the model initialization. We then investigate the effective use of the proposed model in the application of hashing, by coupling the proposed encoders under a supervised pairwise loss, to develop a Deep Siamese Network, which can be optimized from end to en1d. Extensive experiments demonstrate the impressive performances of the proposed model. We also provide an in-depth analysis of its behaviors against the competitors.

Original languageEnglish (US)
Pages (from-to)2174-2180
Number of pages7
JournalIJCAI International Joint Conference on Artificial Intelligence
Volume2016-January
StatePublished - Jan 1 2016
Event25th International Joint Conference on Artificial Intelligence, IJCAI 2016 - New York, United States
Duration: Jul 9 2016Jul 15 2016

Fingerprint

Lagrange multipliers
Feedforward neural networks
Neurons
Experiments
Deep learning

ASJC Scopus subject areas

  • Artificial Intelligence

Cite this

Zhangyang, W., Yingzhen, Y., Shiyu, C., Qing, L., & Huang, T. S. (2016). Learning A Deep l∞ Encoder for Hashing. IJCAI International Joint Conference on Artificial Intelligence, 2016-January, 2174-2180.

Learning A Deep l∞ Encoder for Hashing. / Zhangyang, Wang; Yingzhen, Yang; Shiyu, Chang; Qing, Ling; Huang, Thomas S.

In: IJCAI International Joint Conference on Artificial Intelligence, Vol. 2016-January, 01.01.2016, p. 2174-2180.

Research output: Contribution to journalConference article

Zhangyang, W, Yingzhen, Y, Shiyu, C, Qing, L & Huang, TS 2016, 'Learning A Deep l∞ Encoder for Hashing', IJCAI International Joint Conference on Artificial Intelligence, vol. 2016-January, pp. 2174-2180.
Zhangyang W, Yingzhen Y, Shiyu C, Qing L, Huang TS. Learning A Deep l∞ Encoder for Hashing. IJCAI International Joint Conference on Artificial Intelligence. 2016 Jan 1;2016-January:2174-2180.
Zhangyang, Wang ; Yingzhen, Yang ; Shiyu, Chang ; Qing, Ling ; Huang, Thomas S. / Learning A Deep l∞ Encoder for Hashing. In: IJCAI International Joint Conference on Artificial Intelligence. 2016 ; Vol. 2016-January. pp. 2174-2180.
@article{84e00430651e4cfea388559c26d70a89,
title = "Learning A Deep l∞ Encoder for Hashing",
abstract = "We investigate the l∞-constrained representation which demonstrates robustness to quantization errors, utilizing the tool of deep learning. Based on the Alternating Direction Method of Multipliers (ADMM), we formulate the original convex minimization problem as a feed-forward neural network, named Deep l∞ Encoder, by introducing the novel Bounded Linear Unit (BLU) neuron and modeling the Lagrange multipliers as network biases. Such a structural prior acts as an effective network regularization, and facilitates the model initialization. We then investigate the effective use of the proposed model in the application of hashing, by coupling the proposed encoders under a supervised pairwise loss, to develop a Deep Siamese Network, which can be optimized from end to en1d. Extensive experiments demonstrate the impressive performances of the proposed model. We also provide an in-depth analysis of its behaviors against the competitors.",
author = "Wang Zhangyang and Yang Yingzhen and Chang Shiyu and Ling Qing and Huang, {Thomas S.}",
year = "2016",
month = "1",
day = "1",
language = "English (US)",
volume = "2016-January",
pages = "2174--2180",
journal = "IJCAI International Joint Conference on Artificial Intelligence",
issn = "1045-0823",

}

TY - JOUR

T1 - Learning A Deep l∞ Encoder for Hashing

AU - Zhangyang, Wang

AU - Yingzhen, Yang

AU - Shiyu, Chang

AU - Qing, Ling

AU - Huang, Thomas S.

PY - 2016/1/1

Y1 - 2016/1/1

N2 - We investigate the l∞-constrained representation which demonstrates robustness to quantization errors, utilizing the tool of deep learning. Based on the Alternating Direction Method of Multipliers (ADMM), we formulate the original convex minimization problem as a feed-forward neural network, named Deep l∞ Encoder, by introducing the novel Bounded Linear Unit (BLU) neuron and modeling the Lagrange multipliers as network biases. Such a structural prior acts as an effective network regularization, and facilitates the model initialization. We then investigate the effective use of the proposed model in the application of hashing, by coupling the proposed encoders under a supervised pairwise loss, to develop a Deep Siamese Network, which can be optimized from end to en1d. Extensive experiments demonstrate the impressive performances of the proposed model. We also provide an in-depth analysis of its behaviors against the competitors.

AB - We investigate the l∞-constrained representation which demonstrates robustness to quantization errors, utilizing the tool of deep learning. Based on the Alternating Direction Method of Multipliers (ADMM), we formulate the original convex minimization problem as a feed-forward neural network, named Deep l∞ Encoder, by introducing the novel Bounded Linear Unit (BLU) neuron and modeling the Lagrange multipliers as network biases. Such a structural prior acts as an effective network regularization, and facilitates the model initialization. We then investigate the effective use of the proposed model in the application of hashing, by coupling the proposed encoders under a supervised pairwise loss, to develop a Deep Siamese Network, which can be optimized from end to en1d. Extensive experiments demonstrate the impressive performances of the proposed model. We also provide an in-depth analysis of its behaviors against the competitors.

UR - http://www.scopus.com/inward/record.url?scp=85006137678&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85006137678&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85006137678

VL - 2016-January

SP - 2174

EP - 2180

JO - IJCAI International Joint Conference on Artificial Intelligence

JF - IJCAI International Joint Conference on Artificial Intelligence

SN - 1045-0823

ER -