Robust Self-organizing Maps

Allende, Héctor; Moreno, Sebastián; Rogel, Cristian; Salas, Rodrigo

doi:10.1007/978-3-540-30463-0_22

Cited by 12 publications

(7 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…During training, in each iteration, a small number of inputs/weights are selected randomly to be perturbed by adding some type of noise [15] [31]. The basic idea is to add noise to the inputs or weights of each neuron.…”

Section: Noise and Weight Perturbationsmentioning

confidence: 99%

Fault tolerance of self organizing maps

Torres-Huitzil

Popovych

Girau

2017

2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (W

View full text Add to dashboard Cite

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

show abstract

Section: Noise and Weight Perturbationsmentioning

confidence: 99%

Fault tolerance of self organizing maps

Torres-Huitzil

Popovych

Girau

2017

2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (W

View full text Add to dashboard Cite

show abstract

“…During training, in each iteration, a small number of inputs/weights are selected randomly to be perturbed by adding some type of noise [25] [3]. The basic idea is to add noise to the inputs or weights of each neuron.…”

Section: Noise and Weight Perturbationsmentioning

confidence: 99%

Fault tolerance of self-organizing maps

Girau

Torres-Huitzil

2018

Neural Comput & Applic

View full text Add to dashboard Cite

Bio-inspired computing principles are considered as a source of promising paradigms for fault-tolerant computation. Among bio-inspired approaches, neural networks are potentially capable of absorbing some degrees of vulnerability based on their natural properties. This calls for attention, since beyond energy, the growing number of defects in physical substrates is now a major constraint that affects the design of computing devices. However, studies have shown that most neural networks cannot be considered intrinsically fault tolerant without a proper design. In this paper, the fault tolerance of Self Organizing Maps (SOMs) is investigated, considering implementations targeted onto field programmable gate arrays (FPGAs), where the bit-flip fault model is employed to inject faults in registers. Quantization and distortion measures are used to evaluate performance on synthetic datasets under different fault ratios. Three passive techniques intended to enhance fault tolerance of SOMs during training/learning are also considered in the evaluation. We also evaluate the influence of technological choices on fault tolerance: sequential or parallel implementation, weight storage policies. Experimental results are analyzed through the evolution of neural prototypes during learning and fault injection. We show that SOMs benefit from an already desirable property: graceful degradation. Moreover, depending on some technological choices, SOMs may become very fault tolerant, and their fault tolerance even improves when weights are stored in an individualized way in the implementation.

show abstract

“…Time-serial input data for Kohonen's SOM can be either discrete or continuous (Allende, Moreno, Rogel, & Salas, 2004). When a researcher wishes to conduct the SOM on multivariate data, a value is required for all variables for each measurement point (Germano, 1999).…”

Section: Data Requirements For Sommentioning

confidence: 99%

Capturing moment-to-moment changes in multivariate human experience

Ruiter¹,

Steen²,

Hartigh³

et al. 2016

International Journal of Behavioral Development

View full text Add to dashboard Cite

In this article, we aim to shed light on a technique to study intra-individual variability that spans the time frame of seconds and minutes, i.e., micro-level development. This form of variability is omnipresent in behavioural development and processes of human experience, yet is often ignored in empirical studies, given a lack of proper analysis tools. The current article illustrates that a clustering technique called Kohonen’s Self-Organizing Maps (SOM), which is widely used in fields outside of psychology, is an accessible technique that can be used to capture intra-individual variability of multivariate data. We illustrate this technique with a case study involving self-experience in the context of a parent–adolescent interaction. We show that, with techniques such as SOM, it is possible to reveal how multiple components of an intra-individual process (the adolescent’s self-affect and autonomy) are non-linearly connected across time, and how these relationships transition in accordance with a changing contextual factor (parental connectedness) during a single interaction. We aim to inspire researchers to adopt this technique and explore the intra-individual variability of more developmental processes, across a variety of domains, as deciphering such micro-level processes is crucial for understanding the nature of psychological and behavioural development.

show abstract

Robust Self-organizing Maps

Cited by 12 publications

References 10 publications

Fault tolerance of self organizing maps

Fault tolerance of self organizing maps

Fault tolerance of self-organizing maps

Capturing moment-to-moment changes in multivariate human experience

Contact Info

Product

Resources

About