Quantile based decision making rule of the neural networks committee for ill-posed approximation problems

Kruglov, I. A.; Mishulina, O. A.; Bakirov, M. B.

doi:10.1016/j.neucom.2012.01.038

Cited by 7 publications

(5 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Concomitant with the substantial progress made in semiconductor technologies and novel computing architectures 1 – 9 , artificial neural network (ANN)-related machine learning applications are being extensively utilized in many fields, including computer vision 10 , natural language processing 11 , emotion detection 12 , speech recognition 13 , medical image analysis 14 , 15 , and decision-making 16 , 17 . However, to solve complex tasks in a timely manner, ANNs require massive amounts of resources, both regarding computing speed and energy consumption.…”

Section: Introductionmentioning

confidence: 99%

Photonic machine learning with on-chip diffractive optics

et al. 2023

View full text Add to dashboard Cite

Machine learning technologies have been extensively applied in high-performance information-processing fields. However, the computation rate of existing hardware is severely circumscribed by conventional Von Neumann architecture. Photonic approaches have demonstrated extraordinary potential for executing deep learning processes that involve complex calculations. In this work, an on-chip diffractive optical neural network (DONN) based on a silicon-on-insulator platform is proposed to perform machine learning tasks with high integration and low power consumption characteristics. To validate the proposed DONN, we fabricated 1-hidden-layer and 3-hidden-layer on-chip DONNs with footprints of 0.15 mm2 and 0.3 mm2 and experimentally verified their performance on the classification task of the Iris plants dataset, yielding accuracies of 86.7% and 90%, respectively. Furthermore, a 3-hidden-layer on-chip DONN is fabricated to classify the Modified National Institute of Standards and Technology handwritten digit images. The proposed passive on-chip DONN provides a potential solution for accelerating future artificial intelligence hardware with enhanced performance.

show abstract

Section: Introductionmentioning

confidence: 99%

Photonic machine learning with on-chip diffractive optics

et al. 2023

View full text Add to dashboard Cite

show abstract

“…Artificial neural networks (ANNs) have received significant attention in many fields, including computer vision [1] , natural language processing [2] , decision-making [3,4] , etc. Faced with complex tasks, the requirements of ANNs for computing power are more stringent, causing a heavy computation burden in existing electronic computing hardware [5][6][7][8][9][10][11][12][13] [e.g., the central processing unit (CPU), the graphical processing unit (GPU), the field-programmable gate array (FPGA), and the application-specific integrated circuit (ASIC)].…”

Section: Introductionmentioning

confidence: 99%

Integrated diffractive optical neural network with space-time interleaving

Fu,

Huang,

Sun

et al. 2023

Chin. Opt. Lett.

View full text Add to dashboard Cite

Integrated diffractive optical neural networks (DONNs) have significant potential for complex machine learning tasks with high speed and ultralow energy consumption. However, the on-chip implementation of a high-performance optical neural network is limited by input dimensions. In contrast to existing photonic neural networks, a space-time interleaving technology based on arrayed waveguides is designed to realize an on-chip DONN with high-speed, high-dimensional, and all-optical input signal modulation. To demonstrate the performance of the on-chip DONN with high-speed space-time interleaving modulation, an on-chip DONN with a designed footprint of 0.0945 mm 2 is proposed to resolve the vowel recognition task, reaching a computation speed of about 1.4 × 10 13 operations per second and yielding an accuracy of 98.3% in numerical calculation. In addition, the function of the specially designed arrayed waveguides for realizing parallel signal inputs using space-time conversion has been verified experimentally. This method can realize the on-chip DONN with higher input dimension and lower energy consumption.

show abstract

“…This RL is formed in accordance with the practical application of the search means. Let us suppose that necessary pattern recognition algorithms [8,9], for example neural network algorithms [10][11][12], are used in this stage. In the second stage, the human operator, who is a specialist in the area of application, for instance a criminalist or an analyst of pictures from space, joins the search operation.…”

Section: B Introductionmentioning

confidence: 99%

Model for Evaluating the Effectiveness of Search Operations

Kulik

2015

J. ICT Res. Appl.

View full text Add to dashboard Cite

Abstract. An automated search with human involvement consisting of two stages is given detailed consideration in this paper. In the first stage, a search without direct human involvement is implemented. In the second stage, the search assumes human involvement. Evaluations of the search operations' effectiveness are presented. These operations are implemented for searching one object exclusively among a variety of similar objects. The average number of similar objects recommended for further analysis was used as effectiveness indicator. A set of numerical evaluation criteria for search effectiveness is introduced. The basis of the search block is a pattern recognition algorithm characterized by two probabilities: 1) probability of missing a target, and 2) false alarm probability. An analytical model of the search block was developed. In this paper particular attention is given to the average length of the recommendatory list as an effectiveness indicator. Four properties of this indicator were determined. Keywords: analytical model; false alarm probability; operations research; pattern recognition; probability of missing a target; search operations. 1 B IntroductionDifferent models of operations research [1] make it possible to implement the necessary evaluation of the effectiveness of different operations, including search operations. In practice these operations can be implemented in different fields. For instance, they can be used in criminalistics for searching a fingerprint in a database (DB) of similar fingerprints, in medicine for searching for people with similar illnesses, on the internet for searching factographical data in response to a user's enquiry, in the context of search operations on land and at sea looking for the image of a lost object, such as a ship or a plane, in a database of set images, and other different fields of practical activity. Operations research is an important stage in the evaluation of a search operation's effectiveness and decision-making. In analyzing operations, different tasks should be decided. Some of these tasks have been reviewed in [2][3][4][5]. In some cases, search operations are an essential part of a decision making process. Informative search [2,4,6,7] plays a great role in information systems. Effective search is very

show abstract

Quantile based decision making rule of the neural networks committee for ill-posed approximation problems

Cited by 7 publications

References 19 publications

Photonic machine learning with on-chip diffractive optics

Photonic machine learning with on-chip diffractive optics

Integrated diffractive optical neural network with space-time interleaving

Model for Evaluating the Effectiveness of Search Operations

Contact Info

Product

Resources

About