Capturing significant events with neural networks

Szu, Harold; Hsu, Charles C.; Jenkins, Jeffrey L.; Willey, Jeff; Landa, Joseph S.

doi:10.1016/j.neunet.2012.01.003

Cited by 13 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Applied Computational Intelligence and Soft Computing Analysis [3]. Our observation is consistent with the mathematics of Compressive Video Sensing, but suggests an organized sparseness video sampling [4]. In other words, these sensory change response vectors have a naturally structured sparseness spanning a pseudoorthogonal subspace that can equally yield a minimum representation of a set of underdetermined linear measurements.…”

Section: Compressive Sensing Ecgsupporting

confidence: 70%

Smartphone Homecare Monitoring of Hearts

Szu

Hsu

Moon

et al. 2013

Applied Computational Intelligence and Soft Computing

Self Cite

View full text Add to dashboard Cite

Homecare monitoring blood pressures and heartbeats are commercially available using dedicated devices, for example, wrist watch, pulse oximetry. With the advent of Smartphone and compressive sensing technology, we wish to monitor precisely the electrical waveforms of heartbeats called the electrocardiography (ECG) for an aging global villager biomedical wellness homecare system. Our design separates into 3 innovative modules within the size-weight and power-cost bandwidth (Swap-CB) limitation. We develop each separately but in concert with one another: (i) Smart Electrode (adopting a low-power-mixed signal embedded with modern compressive sensing firmware and applying the nanotechnology to improve the electrodes’ contact impedance as well as novel transduction mechanism, between ECG and electronics, e.g., a pressure mattress coupling, or fiber-optics coupling); (ii) Learnable Database (utilizing adaptive wavelets transforms for systolic and diastolic P-QRS-T-U features extraction Aided Target Recognition and adopting Sequential Query Language for a relational database allowing distant monitoring and retrievable); (iii) Smartphone (inheriting a large touch screen interface display with powerful computation capability and assisting caretaker reporting system with GPS and ID and two-way interaction with patient panic button for programmable emergence reporting procedure). While (i) is novel, (ii) and (iii) are mature. Together, they can eventually provide a supplementary home screening system for the post- or the prediagnosis care at home with a built-in database searchable with the time, the place, and the degree of urgency happened, using in situ screening.

show abstract

Section: Compressive Sensing Ecgsupporting

confidence: 70%

Smartphone Homecare Monitoring of Hearts

Szu

Hsu

Moon

et al. 2013

Applied Computational Intelligence and Soft Computing

Self Cite

View full text Add to dashboard Cite

show abstract

“…This energy is equivalent to H. Then, we assume that J / bond = J / photon since 1 photon will be absorbed by 1 bond to break the bond. Next, we observe that = ħ (5) where E = energy per photon, ħ = Planck's reduced constant = 6.626x10^-34 J s, c = speed of light = 3.00x10^8 m/s, and λ = wavelength. Rearranging, we solve for the unknown wavelength with the known bond energy and obtain λ = ħc / E. Given a uniform plane wave, we derive the complex de Broglie matter wave for a particle in terms of the chemical properties that influence the dynamics of a photochemical reaction.…”

Section: -Realistic Simulation Of Physics and Chemical Materials Propementioning

confidence: 93%

“…We assume that the measurable parts of multi-resolution interactions of, or between systems, is an imprecise representation of real-world data. Using TMR-AR and VR we could capture the mechanisms of interaction over time to enhance data representation and visualization possibilities [5].…”

Section: -Real and Synthetic Multi-resolution Imaging And Analysis Enmentioning

confidence: 99%

Telescopic multi-resolution augmented reality

2014

Self Cite

View full text Add to dashboard Cite

To ensure a self-consistent scaling approximation, the underlying microscopic fluctuation components can naturally influence macroscopic means, which may give rise to emergent observable phenomena. In this paper, we describe a consistent macroscopic (cm-scale), mesoscopic (micron-scale), and microscopic (nano-scale) approach to introduce Telescopic Multi-Resolution (TMR) into current Augmented Reality (AR) visualization technology. We propose to couple TMR-AR by introducing an energy-matter interaction engine framework that is based on known Physics, Biology, Chemistry principles. An immediate payoff of TMR-AR is a self-consistent approximation of the interaction between microscopic observables and their direct effect on the macroscopic system that is driven by real-world measurements. Such an interdisciplinary approach enables us to achieve more than multiple scale, telescopic visualization of real and virtual information but also conducting thought experiments through AR. As a result of the consistency, this framework allows us to explore a large dimensionality parameter space of measured and unmeasured regions. Towards this direction, we explore how to build learnable libraries of biological, physical, and chemical mechanisms. Fusing analytical sensors with TMR-AR libraries provides a robust framework to optimize testing and evaluation through data-driven or virtual synthetic simulations. Visualizing mechanisms of interactions requires identification of observable image features that can indicate the presence of information in multiple spatial and temporal scales of analog data. The AR methodology was originally developed to enhance pilot-training as well as 'make believe' entertainment industries in a user-friendly digital environment We believe TMR-AR can someday help us conduct thought experiments scientifically, to be pedagogically visualized in a zoom-in-and-out, consistent, multi-scale approximations.

show abstract

“…We wish to emphasize an important nearest neighbor classifier strategy, especially that the change storage tends to be ortho-normal to the original memory, and that it readily permits a Fault Tolerant recall, after the direction cosine of angles less than 45 o, or values bigger than 0.7 [9]. This allows us to validate archived information given new stories as well as guess missing information from past history when a full hypothesis or storyline is not yet available.…”

Section: Associative Memory (Am) Can Gain Fault Tolerance (Ft) and Genmentioning

confidence: 99%

“…AM matrix representation showing basic recognition of facial features (for full explanation, see ref[9]) .…”

mentioning

confidence: 99%

e-IQ and IQ knowledge mining for generalized LDA

Jenkins

Bergem

Sweet³

et al. 2015

SPIE Proceedings

Self Cite

View full text Add to dashboard Cite

How can the human brain uncover patterns, associations and features in real-time, real-world data? There must be a general strategy used to transform raw signals into useful features, but representing this generalization in the context of our information extraction tool set is lacking. In contrast to Big Data (BD), Large Data Analysis (LDA) has become a reachable multi-disciplinary goal in recent years due in part to high performance computers and algorithm development, as well as the availability of large data sets. However, the experience of Machine Learning (ML) and information communities has not been generalized into an intuitive framework that is useful to researchers across disciplines. The data exploration phase of data mining is a prime example of this unspoken, ad-hoc nature of ML -the Computer Scientist works with a Subject Matter Expert (SME) to understand the data, and then build tools (i.e. classifiers, etc.) which can benefit the SME and the rest of the researchers in that field. We ask, why is there not a tool to represent information in a meaningful way to the researcher asking the question? Meaning is subjective and contextual across disciplines, so to ensure robustness, we draw examples from several disciplines and propose a generalized LDA framework for independent data understanding of heterogeneous sources which contribute to Knowledge Discovery in Databases (KDD). Then, we explore the concept of adaptive Information resolution through a 6W unsupervised learning methodology feedback system. In this paper, we will describe the general process of man-machine interaction in terms of an asymmetric directed graph theory (digging for embedded knowledge), and model the inverse machine-man feedback (digging for tacit knowledge) as an ANN unsupervised learning methodology. Finally, we propose a collective learning framework which utilizes a 6W semantic topology to organize heterogeneous knowledge and diffuse information to entities within a society in a personalized way.

show abstract

Capturing significant events with neural networks

Cited by 13 publications

References 20 publications

Smartphone Homecare Monitoring of Hearts

Smartphone Homecare Monitoring of Hearts

Telescopic multi-resolution augmented reality

e-IQ and IQ knowledge mining for generalized LDA

Contact Info

Product

Resources

About