46 Hits in 4.1 sec

Robust Gradient Descent via Moment Encoding with LDPC Codes [article]

Raj Kumar Maity, Ankit Singh Rawat, Arya Mazumdar
2019 arXiv   pre-print
We, instead, propose to encode the second-moment of the data with a low density parity-check (LDPC) code.  ...  We show that for a random model for stragglers, the proposed moment encoding based gradient descent method can be viewed as the stochastic gradient descent method.  ...  We compare our LDPC codes based (rate= 1/2) moment encoding scheme with the recently proposed data encoding (with MDS/Gaussian matrices) scheme of Karakus et al.  ... 
arXiv:1805.08327v2 fatcat:bghtp26hhjbutjnxx6jffb3e5q

Boost decoding performance of finite geometry LDPC codes with deep learning tactics [article]

Guangwen Li, Xiao Yu
2022 arXiv   pre-print
We adopt the similar decoding framework to seek a low-complexity and high-performance decoder for a class of finite geometry LDPC codes in short and moderate block lengths.  ...  Consequently, for the referred LDPC codes, their rigorous algebraic structure promotes the feasibility of cutting down the number of trainable parameters even to only one, whereas incurring marginal performance  ...  For a class of finite-geometry LDPC (FG-LDPC) codes [24] , the unique characteristic of encoding in linear time makes it advantageous over LDPC codes with other constructions.  ... 
arXiv:2205.00481v1 fatcat:7fa6datnhfeixgoodeohax2ltq

Ricean K-factor Estimation based on Channel Quality Indicator in OFDM Systems using Neural Network [article]

Kun Wang
2018 arXiv   pre-print
Ricean channel model is widely used in wireless communications to characterize the channels with a line-of-sight path.  ...  Most existing works estimate K factor based on either maximum-likelihood criterion or higher-order moments, and the existing works are targeted at K-factor estimation at receiver side.  ...  With the objective function in hand, we can compute the gradient by using backpropagation algorithm and thus apply gradient descent.  ... 
arXiv:1808.06537v1 fatcat:i4uhcpa3yvdlvpgfxktvnp7dvm

Table of Contents [EDICS]

2020 IEEE Transactions on Signal Processing  
Luo 3644 Federated Variance-Reduced Stochastic Gradient Descent With Robustness to Byzantine Attacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ...  Krishna 3103 Efficient QP-ADMM Decoder for Binary LDPC Codes and Its Performance Analysis . . . . . . . J. Bai, Y. Wang, and Q.  ...  Li 1500 Resolving Range Ambiguity via Multiple-Input Multiple-Output Radar With Element-Pulse Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ... 
doi:10.1109/tsp.2020.3045363 fatcat:wcnvdcy3rvhblh7rtxfe6gz4re

Fundamental Limits of Approximate Gradient Coding [article]

Sinong Wang, Jiashang Liu, Ness Shroff
2019 arXiv   pre-print
We show that, to exactly recover the gradient with high probability, the computation load is lower bounded by O((n)/(n/s)). We also propose a code that exactly matches such lower bound.  ...  We identify a fundamental three-fold tradeoff for any approximate gradient coding scheme d≥ O((1/ϵ)/(n/s)), where d is the computation load, ϵ is the error of gradient.  ...  Mazumdar, "Robust gradient descent via moment encoding with ldpc codes," SysML, 2018. [20] M. Ye and E.  ... 
arXiv:1901.08166v1 fatcat:bppo7wujfvfgbkzavlcj3nddvi

2020 Index IEEE Transactions on Signal Processing Vol. 68

2020 IEEE Transactions on Signal Processing  
., One-Step Prediction for Discrete Time-Varying Nonlinear Systems With Unknown Inputs and Correlated Noises; TSP  ...  Ye, Y., +, TSP 2020 5842-5854 Quickly Finding the Best Linear Model in High Dimensions via Projected Gradient Descent.  ...  Ibrahim, M.S., +, TSP 2020 1897-1909 Federated Variance-Reduced Stochastic Gradient Descent With Robustness to Byzantine Attacks. Wu, Z., +, TSP 2020 4583-4596 Functional Nonlinear Sparse Models.  ... 
doi:10.1109/tsp.2021.3055469 fatcat:6uswtuxm5ba6zahdwh5atxhcsy

Coded Distributed Computing with Partial Recovery [article]

Emre Ozfatura and Sennur Ulukus and Deniz Gunduz
2021 arXiv   pre-print
Coded computation techniques provide robustness against straggling workers in distributed computing.  ...  In this paper, we first introduce a novel coded matrix-vector multiplication scheme, called coded computation with partial recovery (CCPR), which benefits from the advantages of both coded and uncoded  ...  Mazumdar, “Robust gradient de- 5273–5282, 2020. scent via moment encoding and ldpc codes,” in 2019 IEEE International [38] B. Hasırcıoğlu, J. Gómez-Vilardebó, and D.  ... 
arXiv:2007.02191v2 fatcat:a7rx5w4yrbhnbnh4uf2s2h3eke

Technical Program

2021 2020 IEEE Information Theory Workshop (ITW)  
The Schalkwijk-Kailath (SK) feedback scheme is a capacity-achieving coding scheme for the point-to-point white Gaussian channel with feedback.  ...  It is shown here to be effective also in assisting the encoder: it achieves the helper capacity on the single-user Gaussian channel, on the multiple-access Gaussian channel, on the Exponential channel,  ...  Among them are certain moment equivalence assumptions often used in the robust statistics literature.  ... 
doi:10.1109/itw46852.2021.9457668 fatcat:j425ygeajrbd5esztbe5zgygty

Communication-Efficient Edge AI: Algorithms and Systems [article]

Yuanming Shi, Kai Yang, Tao Jiang, Jun Zhang, Khaled B. Letaief
2020 arXiv   pre-print
When the loss function is the squared loss, it was proposed in [100] to encode the second moment of the data matrix with a low density parity-check (LDPC) code to mitigate the effect of the stragglers  ...  They also indicated that the moment encoding based gradient descent algorithm can be viewed as a stochastic gradient descent method, which provides opportunities to obtain convergence guarantees for the  ... 
arXiv:2002.09668v1 fatcat:nhasdzb7t5dt5brs2r7ocdzrnm

A Survey of Coded Distributed Computing [article]

Jer Shyuan Ng, Wei Yang Bryan Lim, Nguyen Cong Luong, Zehui Xiong, Alia Asheralieva, Dusit Niyato, Cyril Leung, Chunyan Miao
2020 arXiv   pre-print
Coding theoretic techniques have proved effective in WiFi and cellular systems to deal with channel noise.  ...  To address these issues, coded distributed computing (CDC), i.e., a combination of coding theoretic techniques and distributed computing, has been recently proposed as a promising solution.  ...  data based on pair-wise balanced scheme and provide a rigorous convergence analysis of the proposed coding scheme [15] LDPC Codes Encode the second moment of the data points [151] , [152] Encoded  ... 
arXiv:2008.09048v1 fatcat:riy4dxvuc5ae3krz7lf25zkg6m

Table of Contents

2021 2021 IEEE International Symposium on Information Theory (ISIT)  
CODING D5-T1-S5.1: ON GRADIENT CODING WITH PARTIAL RECOVERY ............................................................. 2274 Sahasrajit Sarmasarkar, Indian Institute of Technology, Bombay, India; V.  ...  Lalitha, International Institute of Information Technology, Hyderabad, India; Nikhil Karamchandani, Indian Institute of Technology, Bombay, India xlvii D5-T1-S5.2: APPROXIMATE GRADIENT CODING WITH OPTIMAL  ... 
doi:10.1109/isit45174.2021.9518212 fatcat:vkpbkhd3kzaxtcwe53osdnt6ku

Machine Learning for Multimedia Communications

Nikolaos Thomos, Thomas Maugey, Laura Toni
2022 Sensors  
Then, training is done by means of the stochastic gradient descent algorithm. The decoder achieves improved decoding performance and reduced complexity.  ...  It is proposed to exploit the lifting structure of the protograph LDPC codes, which are the base codes of 5G-NR LDPC codes.  ... 
doi:10.3390/s22030819 pmid:35161566 pmcid:PMC8840624 fatcat:nmz7s6ei3bdddarbt2bqhf6beu

Joint Alignment of Multiple Protein–Protein Interaction Networks via Convex Optimization

Somaye Hashemifar, Qixing Huang, Jinbo Xu
2016 Journal of Computational Biology  
HSA is robust, accurate, and outperforms or rivals existing computational tools when evaluated on simulated and real datasets in diverse cell types.  ...  Similar to conventional gradient descent, by sensing the food gradient, each cell has its own belief about the location of the food source.  ...  We conclude by discussing how the efficient and robust bacterial gradient descent algorithms we developed can be used by distributed sensors or wireless networks that operate under strict communication  ... 
doi:10.1089/cmb.2016.0025 pmid:27428933 fatcat:tr4e3u3hqjaoziavhhsklxkkmi

Parameter Estimation and Classification via Supervised Learning in the Wireless Physical Layer: A Survey and Tutorial

Kyle W. McClintick, Galahad M. Wernsing, Paulo Victor R. Ferreira, Alexander M. Wyglinski
2021 IEEE Access  
layer applications: Dynamic Spectrum Access (DSA), channel corrections, Automatic Gain Control (AGC), Multiple Input Multiple Output (MIMO) control, Analog to Digital (ADC) conversion, and Automatic Coding  ...  , as well as to provide the reader with the necessary tools and techniques needed for addressing open challenges to be experienced by future wireless networks.  ...  Recent improvements to control codes include turbo codes [108] and Low-Density Parity-Check (LDPC) codes [109] .  ... 
doi:10.1109/access.2021.3128813 fatcat:kj6xxdefzbbttanqdlwgq7ekqq

Message Passing Algorithms for Phase Noise Tracking Using Tikhonov Mixtures [article]

Shachar Shayovitz, Dan Rapaheli
2013 arXiv   pre-print
The algorithm is based on the Sum & Product Algorithm (SPA) with phase noise messages modeled as Tikhonov mixtures.  ...  The global algorithms attempt to solve (35) by gradient descent type solutions which are very computationally demanding.  ...  In Fig. 3 and 4 , we show the BER and PER results for an 8PSK constellation with an LDPC code of length 4608 with code rate 0.89.  ... 
arXiv:1306.3693v2 fatcat:rk36dsicvbep5g3qs477x73shi
« Previous Showing results 1 — 15 out of 46 results