The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Este artigo aborda códigos de superposição esparsa (SSCs) de curto comprimento sobre o canal de ruído gaussiano branco aditivo. A passagem aproximada de mensagens (AMP) amortecida é usada para decodificar SSCs curtos com dicionários gaussianos independentes de média zero e distribuídos de forma idêntica. Para projetar fatores de amortecimento em AMP por meio de aprendizado profundo, este artigo constrói redes de decodificação de AMP amortecidas e desdobradas. Um método de recozimento para aprendizado profundo é proposto para projetar fatores de amortecimento quase ótimos com alta probabilidade. No recozimento, os fatores de amortecimento são primeiro otimizados por meio de aprendizado profundo no regime de baixa relação sinal-ruído (SNR). Em seguida, os fatores de amortecimento obtidos são ajustados para os valores iniciais na descida gradiente estocástica, o que otimiza os fatores de amortecimento para SNR ligeiramente maior. A repetição deste processo de recozimento projeta fatores de amortecimento no regime de alta SNR. Simulações numéricas mostram que o recozimento atenua a flutuação nos fatores de amortecimento aprendidos e supera a pesquisa exaustiva com base em um fator de amortecimento independente da iteração.
Toshihiro YOSHIDA
Toyohashi University of Technology
Keigo TAKEUCHI
Toyohashi University of Technology
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Toshihiro YOSHIDA, Keigo TAKEUCHI, "Deep Learning of Damped AMP Decoding Networks for Sparse Superposition Codes via Annealing" in IEICE TRANSACTIONS on Fundamentals,
vol. E106-A, no. 3, pp. 414-421, March 2023, doi: 10.1587/transfun.2022TAP0009.
Abstract: This paper addresses short-length sparse superposition codes (SSCs) over the additive white Gaussian noise channel. Damped approximate message-passing (AMP) is used to decode short SSCs with zero-mean independent and identically distributed Gaussian dictionaries. To design damping factors in AMP via deep learning, this paper constructs deep-unfolded damped AMP decoding networks. An annealing method for deep learning is proposed for designing nearly optimal damping factors with high probability. In annealing, damping factors are first optimized via deep learning in the low signal-to-noise ratio (SNR) regime. Then, the obtained damping factors are set to the initial values in stochastic gradient descent, which optimizes damping factors for slightly larger SNR. Repeating this annealing process designs damping factors in the high SNR regime. Numerical simulations show that annealing mitigates fluctuation in learned damping factors and outperforms exhaustive search based on an iteration-independent damping factor.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2022TAP0009/_p
Copiar
@ARTICLE{e106-a_3_414,
author={Toshihiro YOSHIDA, Keigo TAKEUCHI, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Deep Learning of Damped AMP Decoding Networks for Sparse Superposition Codes via Annealing},
year={2023},
volume={E106-A},
number={3},
pages={414-421},
abstract={This paper addresses short-length sparse superposition codes (SSCs) over the additive white Gaussian noise channel. Damped approximate message-passing (AMP) is used to decode short SSCs with zero-mean independent and identically distributed Gaussian dictionaries. To design damping factors in AMP via deep learning, this paper constructs deep-unfolded damped AMP decoding networks. An annealing method for deep learning is proposed for designing nearly optimal damping factors with high probability. In annealing, damping factors are first optimized via deep learning in the low signal-to-noise ratio (SNR) regime. Then, the obtained damping factors are set to the initial values in stochastic gradient descent, which optimizes damping factors for slightly larger SNR. Repeating this annealing process designs damping factors in the high SNR regime. Numerical simulations show that annealing mitigates fluctuation in learned damping factors and outperforms exhaustive search based on an iteration-independent damping factor.},
keywords={},
doi={10.1587/transfun.2022TAP0009},
ISSN={1745-1337},
month={March},}
Copiar
TY - JOUR
TI - Deep Learning of Damped AMP Decoding Networks for Sparse Superposition Codes via Annealing
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 414
EP - 421
AU - Toshihiro YOSHIDA
AU - Keigo TAKEUCHI
PY - 2023
DO - 10.1587/transfun.2022TAP0009
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E106-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2023
AB - This paper addresses short-length sparse superposition codes (SSCs) over the additive white Gaussian noise channel. Damped approximate message-passing (AMP) is used to decode short SSCs with zero-mean independent and identically distributed Gaussian dictionaries. To design damping factors in AMP via deep learning, this paper constructs deep-unfolded damped AMP decoding networks. An annealing method for deep learning is proposed for designing nearly optimal damping factors with high probability. In annealing, damping factors are first optimized via deep learning in the low signal-to-noise ratio (SNR) regime. Then, the obtained damping factors are set to the initial values in stochastic gradient descent, which optimizes damping factors for slightly larger SNR. Repeating this annealing process designs damping factors in the high SNR regime. Numerical simulations show that annealing mitigates fluctuation in learned damping factors and outperforms exhaustive search based on an iteration-independent damping factor.
ER -