SOMH: A self-organizing map based topology preserving hashing method

Xu, Xin-Shun; Xiao-long, Liang; Yang, Guo-Jiao; Wang, Xiaolin; Guo, Shanqing; Shi, Yuliang

doi:10.1016/j.neucom.2016.08.102

Cited by 5 publications

(3 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The self-organizing map [14] is one of the most popular unsupervised learning algorithms for non-linear data projection for visualization. The self-organization map and its variants have many applications [31,32,33,34,35]. Inspired by V1 lateral influences we also design a modified version of the Self-organizing map algorithm.…”

Section: V1 Inspired Som Modelmentioning

confidence: 99%

An anatomy-based V1 model: Extraction of Low-level Features, Reduction of distortion and a V1-inspired SOM

Roy¹,

Pal²

2023

Preprint

View full text Add to dashboard Cite

We present a model of the primary visual cortex V1, guided by anatomical experiments. Unlike most machine learning systems our goal is not to maximize accuracy but to realize a system more aligned to biological systems. Our model consists of the V1 layers 4, 2/3, and 5, with inter-layer connections between them in accordance with the anatomy. We further include the orientation selectivity of the V1 neurons and lateral influences in each layer. Our V1 model, when applied to the BSDS500 ground truth images (indicating LGN contour detection before V1), can extract low-level features from the images and perform a significant amount of distortion reduction. As a follow-up to our V1 model, we propose a V1-inspired self-organizing map algorithm (V1-SOM), where the weight update of each neuron gets influenced by its neighbors. V1-SOM can tolerate noisy inputs as well as noise in the weight updates better than SOM and shows a similar level of performance when trained with high dimensional data such as the MNIST dataset. Finally, when we applied V1 processing to the MNIST dataset to extract low-level features and trained V1-SOM with the modified MNIST dataset, the quantization error was significantly reduced. Our results support the hypothesis that the ventral stream performs gradual untangling of input spaces.

show abstract

Section: V1 Inspired Som Modelmentioning

confidence: 99%

An anatomy-based V1 model: Extraction of Low-level Features, Reduction of distortion and a V1-inspired SOM

Roy¹,

Pal²

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Since Openshaw (1994), a geographical interest in dimensionality reduction using the Self-Organizing Map (Kohonen 1990) has sought to reduce high-dimensional structures in geographical processes and proven extensively useful. In an exploratory mode of analysis, the Self-Organizing Map and its variants (Bação et al 2004;Xu et al 2017;Clark et al 2017) have been consistently used in the analysis of complex, non-linear demographic relationships (Skupin and Fabrikant 2003;Agarwal and Skupin 2008;Pearson and Cooper 2012;Arribas-Bel, Nijkamp, et al 2011;Delmelle et al 2013;Spielman and Logan 2013;Psyllidis et al 2018). Another consistent interest is the use of Self-Organizing map for data-driven map reprojection (Skupin 2003;Henriques, Bação, et al 2009;Skupin and Esperbé 2011), which exploits the Self-Organizing Map's distinctive properties in order to build new or better map projections and transformations.…”

Section: Past Explorations and Prior Concerns For Geographic Dimension mentioning

confidence: 99%

Learning Geographical Manifolds: A Kernel Trick for Geographical Machine Learning

Wolf¹,

Knaap²

2019

Preprint

View full text Add to dashboard Cite

Dimension reduction is one of the oldest concerns in geographical analysis. Despite significant, longstanding attention in geographical problems, recent advances in non-linear techniques for dimension reduction, called manifold learning, have not been adopted in classic data-intensive geographical problems. More generally, machine learning methods for geographical problems often focus more on applying standard machine learning algorithms to geographic data, rather than applying true "spatially-correlated learning," in the words of Kohonen. As such, we suggest a general way to incentivize geographical learning in machine learning algorithms, and link it to many past methods that introduced geography into statistical techniques. We develop a specific instance of this by specifying two geographical variants of Isomap, a non-linear dimension reduction, or "manifold learning," technique. We also provide a method for assessing what is added by incorporating geography and estimate the manifold's intrinsic geographic scale. To illustrate the concepts and provide interpretable results, we conducting a dimension reduction on geographical and high-dimensional structure of social and economic data on Brooklyn, New York. Overall, this paper's main endeavor--defining and explaining a way to "geographize" many machine learning methods--yields interesting and novel results for manifold learning the estimation of intrinsic geographical scale in unsupervised learning.

show abstract

“…SOM principle has been used extensively as an analytical and visualization tool in exploratory data analysis [20]. The SOM models are associated with the nodes of a regular, usually two-dimensional grid, as presented in [21], who upgraded the latest version which is available in [22]. Since SOM application is new in this field of study, there is a lack of reference for SOM implementation relative to the LM strategy.…”

Section: Introductionmentioning

confidence: 99%

Determination of the Optimum Load Profile Under Enhanced of Use Tariff (Etou) Scheme Using Combination of Optimization Algorithms and Self Organizing Mapping

Sulaima

Dahlan

Razak

et al. 2022

AEJ

View full text Add to dashboard Cite

Demand side management (DSM) has been conventionally adopted in many ways to efficiently managing the appropriate electricity loads. However, with the sophisticated design of the Time of Use (TOU) tariff to reflect electricity cost reduction, implementing proper Load Management (LM) strategies is challenging. To date, consumers still struggle to define a figure for the LM percentage to be involved in the demand response program. Due to that reason, this study proposes a method to find the best load profile reflecting the new tariff offered by using a combination of optimization algorithms such as Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Evolutionary PSO (EPSO), and Self-Organizing Mapping (SOM). The evaluation has been made to the manufacturing operation with the existing flat tariff to be transferred to the Enhanced Time of Use (ETOU). The test results show that the ability of the proposed combination method to define the optimal outputs such as energy consumption cost, maximum demand cost, load factor index, and building electricity economic responsive index. Meanwhile, the SOM algorithm has been used to classify the enormous numbers of those simulation results produced by algorithms while defining the best LM weightage. As the test results for the case study, it was found that the practical 6% LM weightage was able to reflect the optimal required load profile shifting to be applied by manufacturing operation. Thus, by determining the optimal load profile that suits the ETOU scheme, the consumers can enjoy cost benefits while supporting the demand response program concurrently.

show abstract

SOMH: A self-organizing map based topology preserving hashing method

Cited by 5 publications

References 28 publications

An anatomy-based V1 model: Extraction of Low-level Features, Reduction of distortion and a V1-inspired SOM

An anatomy-based V1 model: Extraction of Low-level Features, Reduction of distortion and a V1-inspired SOM

Learning Geographical Manifolds: A Kernel Trick for Geographical Machine Learning

Determination of the Optimum Load Profile Under Enhanced of Use Tariff (Etou) Scheme Using Combination of Optimization Algorithms and Self Organizing Mapping

Contact Info

Product

Resources

About