Neural Inference Search for Multiloss Segmentation Models

Sam Slade; Li Zhang; Haoqian Huang; Houshyar Asadi; Chee Peng Lim; Yonghong Yu; Dezong Zhao; Hanhe Lin; Rong Gao

doi:10.1109/tnnls.2023.3282799

Neural Inference Search for Multiloss Segmentation Models

Sam Slade, Li Zhang^*, Haoqian Huang, Houshyar Asadi, Chee Peng Lim, Yonghong Yu, Dezong Zhao, Hanhe Lin, Rong Gao

^*Corresponding author for this work

Computer and Information Sciences

Research output: Contribution to journal › Article › peer-review

2 Downloads (Pure)

Abstract

-dimensional matrix rotation for local exploitation. A scheduling mechanism is also introduced in NIS to manage the contributions of these three novel search behaviors in stages. NIS optimizes learning and multiloss parameters simultaneously. Compared with state-of-the-art segmentation methods and those optimized with other well-known search algorithms, NIS-optimized models show significant improvements across multiple performance metrics on five segmentation datasets. NIS also reliably yields better solutions as compared with a variety of search methods for solving numerical benchmark functions.

Original language	English
Pages (from-to)	1-15
Number of pages	15
Journal	IEEE Transactions on Neural Networks and Learning Systems
Early online date	16 Jun 2023
DOIs	https://doi.org/10.1109/tnnls.2023.3282799
Publication status	E-pub ahead of print - 16 Jun 2023

Access to Document

10.1109/tnnls.2023.3282799

108245427_AAMAccepted author manuscript, 1.93 MB

Cite this

@article{962b196362874923917e2e2e7fd87757,

title = "Neural Inference Search for Multiloss Segmentation Models",

abstract = " -dimensional matrix rotation for local exploitation. A scheduling mechanism is also introduced in NIS to manage the contributions of these three novel search behaviors in stages. NIS optimizes learning and multiloss parameters simultaneously. Compared with state-of-the-art segmentation methods and those optimized with other well-known search algorithms, NIS-optimized models show significant improvements across multiple performance metrics on five segmentation datasets. NIS also reliably yields better solutions as compared with a variety of search methods for solving numerical benchmark functions.",

author = "Sam Slade and Li Zhang and Haoqian Huang and Houshyar Asadi and Lim, {Chee Peng} and Yonghong Yu and Dezong Zhao and Hanhe Lin and Rong Gao",

note = "Funding information: This work was supported in part by the European Regional Development Fund (ERDF); in part by RPPTV Ltd., through the Joint Funding of a Ph.D. Studentship via the Intensive Industrial Innovation Program North East (IIIPNE) under Grant 25R17P01847; and in part by Innovate U.K. Smart Grants.",

year = "2023",

month = jun,

day = "16",

doi = "10.1109/tnnls.2023.3282799",

language = "English",

pages = "1--15",

journal = "IEEE Transactions on Neural Networks and Learning Systems",

issn = "2162-237X",

publisher = "IEEE",

}

TY - JOUR

T1 - Neural Inference Search for Multiloss Segmentation Models

AU - Slade, Sam

AU - Zhang, Li

AU - Huang, Haoqian

AU - Asadi, Houshyar

AU - Lim, Chee Peng

AU - Yu, Yonghong

AU - Zhao, Dezong

AU - Lin, Hanhe

AU - Gao, Rong

N1 - Funding information: This work was supported in part by the European Regional Development Fund (ERDF); in part by RPPTV Ltd., through the Joint Funding of a Ph.D. Studentship via the Intensive Industrial Innovation Program North East (IIIPNE) under Grant 25R17P01847; and in part by Innovate U.K. Smart Grants.

PY - 2023/6/16

Y1 - 2023/6/16

N2 - -dimensional matrix rotation for local exploitation. A scheduling mechanism is also introduced in NIS to manage the contributions of these three novel search behaviors in stages. NIS optimizes learning and multiloss parameters simultaneously. Compared with state-of-the-art segmentation methods and those optimized with other well-known search algorithms, NIS-optimized models show significant improvements across multiple performance metrics on five segmentation datasets. NIS also reliably yields better solutions as compared with a variety of search methods for solving numerical benchmark functions.

AB - -dimensional matrix rotation for local exploitation. A scheduling mechanism is also introduced in NIS to manage the contributions of these three novel search behaviors in stages. NIS optimizes learning and multiloss parameters simultaneously. Compared with state-of-the-art segmentation methods and those optimized with other well-known search algorithms, NIS-optimized models show significant improvements across multiple performance metrics on five segmentation datasets. NIS also reliably yields better solutions as compared with a variety of search methods for solving numerical benchmark functions.

U2 - 10.1109/tnnls.2023.3282799

DO - 10.1109/tnnls.2023.3282799

M3 - Article

C2 - 37327096

SP - 1

EP - 15

JO - IEEE Transactions on Neural Networks and Learning Systems

JF - IEEE Transactions on Neural Networks and Learning Systems

SN - 2162-237X

ER -