constructive algorithms
Dotaz
Zobrazit nápovědu
Artificial neural networks are limited in the number of patterns that they can store and accurately recall, with capacity constraints arising from factors such as network size, architectural structure, pattern sparsity, and pattern dissimilarity. Exceeding these limits leads to recall errors, eventually leading to catastrophic forgetting, which is a major challenge in continual learning. In this study, we characterize the theoretical maximum memory capacity of single-layer feedforward networks as a function of these parameters. We derive analytical expressions for maximum theoretical memory capacity and introduce a grid-based construction and sub-sampling method for pattern generation that takes advantage of the full storage potential of the network. Our findings indicate that maximum capacity scales as (N/S) S , where N is the number of input/output units and S the pattern sparsity, under threshold constraints related to minimum pattern differentiability. Simulation results validate these theoretical predictions and show that the optimal pattern set can be constructed deterministically for any given network size and pattern sparsity, systematically outperforming random pattern generation in terms of storage capacity. This work offers a foundational framework for maximizing storage efficiency in neural network systems and supports the development of data-efficient, sustainable AI.
- Klíčová slova
- constructive algorithms, data-efficient AI, memory capacity, neural network, sustainable AI,
- Publikační typ
- časopisecké články MeSH
This paper presents the construction of an innovative high-temperature sensor based on the optical principle. The sensor is designed especially for the measurement of exhaust gases with a temperature range of up to +850 °C. The methodology is based on two principles-luminescence and dark body radiation. The core of this study is the description of sensing element construction together with electronics and the system of photodiode dark current compensation. An advantage of this optical-based system is its immunity to strong magnetic fields. This study also discusses results achieved and further steps. The solution is covered by a European Patent.
- Klíčová slova
- dark current, high temperature, optical fiber, probe construction, ruby crystal, temperature measurement,
- Publikační typ
- časopisecké články MeSH
Bayesian networks have become one of the most popular probabilistic techniques in AI, largely due to the development of several efficient inference algorithms. In this paper we describe a heuristic method for constructing Bayesian networks. Our construction method relies on the relationship between Bayesian networks and decomposable models, a special kind of graphical model. We explain this relationship and then show how it can be used to facilitate model construction. Finally, we describe an implemented computer program that illustrates these ideas.
- MeSH
- algoritmy MeSH
- Bayesova věta MeSH
- lidé MeSH
- neuronové sítě * MeSH
- software MeSH
- systémy řízení databází MeSH
- teorie pravděpodobnosti MeSH
- Check Tag
- lidé MeSH
- Publikační typ
- časopisecké články MeSH
- práce podpořená grantem MeSH
In wireless sensor networks (WSNs), existing routing protocols mainly consider energy efficiency or security separately. However, these protocols must be more comprehensive because many applications should guarantee security and energy efficiency, simultaneously. Due to the limited energy of sensor nodes, these protocols should make a trade-off between network lifetime and security. This paper proposes a cluster-tree-based trusted routing method using the grasshopper optimization algorithm (GOA) called CTTRG in WSNs. This routing scheme includes a distributed time-variant trust (TVT) model to analyze the behavior of sensor nodes according to three trust criteria, including the black hole, sink hole, and gray hole probability, the wormhole probability, and the flooding probability. Furthermore, CTTRG suggests a GOA-based trusted routing tree (GTRT) to construct secure and stable communication paths between sensor nodes and base station. To evaluate each GTRT, a multi-objective fitness function is designed based on three parameters, namely the distance between cluster heads and their parent node, the trust level, and the energy of cluster heads. The evaluation results prove that CTTRG has a suitable and successful performance in terms of the detection speed of malicious nodes, packet loss rate, and end-to-end delay.
Knowledge of soft tissue fiber structure is necessary for accurate characterization and modeling of their mechanical response. Fiber configuration and structure informs both our understanding of healthy tissue physiology and of pathological processes resulting from diseased states. This study develops an automatic algorithm to simultaneously estimate fiber global orientation, abundance, and waviness in an investigated image. To our best knowledge, this is the first validated algorithm which can reliably separate fiber waviness from its global orientation for considerably wavy fibers. This is much needed feature for biological tissue characterization. The algorithm is based on incremental movement of local regions of interest (ROI) and analyzes two-dimensional images. Pixels belonging to the fiber are identified in the ROI, and ROI movement is determined according to local orientation of fiber within the ROI. The algorithm is validated with artificial images and ten images of porcine trachea containing wavy fibers. In each image, 80-120 fibers were tracked manually to serve as verification. The coefficient of determination R2 between curve lengths and histograms documenting the fiber waviness and global orientation were used as metrics for analysis. Verification-confirmed results were independent of image rotation and degree of fiber waviness, with curve length accuracy demonstrated to be below 1% of fiber curved length. Validation-confirmed median and interquartile range of R2, respectively, were 0.90 and 0.05 for curved length, 0.92 and 0.07 for waviness, and 0.96 and 0.04 for global orientation histograms. Software constructed from the proposed algorithm was able to track one fiber in about 1.1 s using a typical office computer. The proposed algorithm can reliably and accurately estimate fiber waviness, curve length, and global orientation simultaneously, moving beyond the limitations of prior methods.
- Klíčová slova
- automated algorithm, collagen structure, fiber orientation, fiber waviness, image analysis, soft tissue analysis,
- MeSH
- algoritmy * MeSH
- kolagen MeSH
- prasata MeSH
- software * MeSH
- zvířata MeSH
- Check Tag
- zvířata MeSH
- Publikační typ
- časopisecké články MeSH
- Názvy látek
- kolagen MeSH
We describe a method for simulating biomembranes of arbitrary shape. In contrast to other dynamically triangulated surface (DTS) algorithms, our method provides a rich, quasi-tangent-continuous, yet local description of the surface. We use curved Nagata triangles, which we generalize to cubic order to achieve the requisite flexibility. The resulting interpolation can be constructed locally without iterations, at the cost of having only approximate tangent continuity away from the vertices. This allows us to provide a parallelized and fine-tuned Monte Carlo implementation. As a first example of the potential benefits of the enhanced description, our method supports inhomogeneous lipid distributions as well as lipid mixing. It also supports restraints and constraints of various types and is constructed to be as easily extensible as possible. We validate the approach by testing its numerical accuracy, followed by reproducing the known Helfrich solutions for shapes with rotational symmetry. Finally, we present some example applications, including curvature-driven demixing and stylized effects of proteins. Input files for these examples, as well as the implementation itself, are freely available for researchers under the name OrganL (https://zenodo.org/doi/10.5281/zenodo.11204709).
Technical details, algorithms, and MATLAB implementation for a method advanced in the paper ``Wavelet Based Dictionaries for Dimensionality Reduction of ECG Signals'', are presented. This work aims to be the companion of that publication, in which an adaptive mathematical model for a given ECG record is proposed. The method comprises the following building blocks.(i)Construction of a suitable redundant set, called 'dictionary', for decomposing an ECG signal as a superposition of elementary components, called 'atoms', selected from that dictionary.(ii)Implementation of the greedy strategy Optimized Orthogonal Matching Pursuit (OOMP) for selecting the atoms intervening in the signal decomposition.This paper gives the details of the algorithms for implementing stage (i), which is not fully elaborated in the previous publication. The proposed dictionaries are constructed from known wavelet families, but translating the prototypes with a shorter step than that corresponding to a wavelet basis. Stage (ii) is readily implementable by the available function OOMP.•The use of the software and the power of the technique is illustrated by reducing the dimensionality of ECG records taken from the MIT-BIH Arrhythmia Database.•The MATLAB software has been made publicly available on a dedicated website.•We provide the explanations, algorithms and software for the construction of scaling functions and wavelet prototypes for 17 different wavelet families. The procedure is designed to allow for straightforward extension of the software by the inclusion of additional options for the wavelet families.
- Klíčová slova
- Dictionary, Dimensionality reduction, ECG signal, Optimized orthogonal matching pursuit, Sarse representation, Wavelet,
- Publikační typ
- časopisecké články MeSH
Sum fraction terms can approximate multi-variable functions on the basis of discrete observations, replacing a partial differential equation definition with polynomial elementary data relation descriptions. Artificial neural networks commonly transform the weighted sum of inputs to describe overall similarity relationships of trained and new testing input patterns. Differential polynomial neural networks form a new class of neural networks, which construct and solve an unknown general partial differential equation of a function of interest with selected substitution relative terms using non-linear multi-variable composite polynomials. The layers of the network generate simple and composite relative substitution terms whose convergent series combinations can describe partial dependent derivative changes of the input variables. This regression is based on trained generalized partial derivative data relations, decomposed into a multi-layer polynomial network structure. The sigmoidal function, commonly used as a nonlinear activation of artificial neurons, may transform some polynomial items together with the parameters with the aim to improve the polynomial derivative term series ability to approximate complicated periodic functions, as simple low order polynomials are not able to fully make up for the complete cycles. The similarity analysis facilitates substitutions for differential equations or can form dimensional units from data samples to describe real-world problems.
- Klíčová slova
- Differential polynomial neural network, General partial differential equation composition, Multi-variable function approximation, Sum derivative term substitution,
- MeSH
- algoritmy MeSH
- interpretace statistických dat MeSH
- matematika * MeSH
- nelineární dynamika MeSH
- neuronové sítě * MeSH
- počasí MeSH
- počítačová simulace MeSH
- strojové učení MeSH
- Publikační typ
- časopisecké články MeSH
- práce podpořená grantem MeSH
Machine learning can be used to define subtypes of psychiatric conditions based on shared biological foundations of mental disorders. Here we analyzed cross-sectional brain images from 4,222 individuals with schizophrenia and 7038 healthy subjects pooled across 41 international cohorts from the ENIGMA, non-ENIGMA cohorts and public datasets. Using the Subtype and Stage Inference (SuStaIn) algorithm, we identify two distinct neurostructural subgroups by mapping the spatial and temporal 'trajectory' of gray matter change in schizophrenia. Subgroup 1 was characterized by an early cortical-predominant loss with enlarged striatum, whereas subgroup 2 displayed an early subcortical-predominant loss in the hippocampus, striatum and other subcortical regions. We confirmed the reproducibility of the two neurostructural subtypes across various sample sites, including Europe, North America and East Asia. This imaging-based taxonomy holds the potential to identify individuals with shared neurobiological attributes, thereby suggesting the viability of redefining existing disorder constructs based on biological factors.
- MeSH
- algoritmy * MeSH
- dospělí MeSH
- hipokampus diagnostické zobrazování patologie MeSH
- lidé středního věku MeSH
- lidé MeSH
- magnetická rezonanční tomografie * MeSH
- mozek diagnostické zobrazování patologie MeSH
- neurozobrazování MeSH
- průřezové studie MeSH
- reprodukovatelnost výsledků MeSH
- schizofrenie * diagnostické zobrazování patologie MeSH
- šedá hmota * diagnostické zobrazování patologie MeSH
- strojové učení MeSH
- Check Tag
- dospělí MeSH
- lidé středního věku MeSH
- lidé MeSH
- mužské pohlaví MeSH
- ženské pohlaví MeSH
- Publikační typ
- časopisecké články MeSH
- Geografické názvy
- Evropa MeSH
- Severní Amerika MeSH
BACKGROUND AND OBJECTIVE: The geodesic ray-tracing method has shown its effectiveness for the reconstruction of fibers in white matter structure. Based on reasonable metrics on the spaces of the diffusion tensors, it can provide multiple solutions and get robust to noise and curvatures of fibers. The choice of the metric on the spaces of diffusion tensors has a significant impact on the outcome of this method. Our objective is to suggest metrics and modifications of the algorithms leading to more satisfactory results in the construction of white matter tracts as geodesics. METHODS: Starting with the DTI modality, we propose to rescale the initially chosen metric on the space of diffusion tensors to increase the geodetic cost in the isotropic regions. This change should be conformal in order to preserve the angles between crossing fibers. We also suggest to enhance the methods to be more robust to noise and to employ the fourth order tensor data in order to handle the fiber crossings properly. RESULTS: We propose a way to choose the appropriate conformal class of metrics where the metric gets scaled according to tensor anisotropy. We use the logistic functions, which are commonly used in statistics as cumulative distribution functions. To prevent deviation of geodesics from the actual paths, we propose a hybrid ray-tracing approach. Furthermore, we suggest how to employ diagonal projections of 4th order tensors to perform fiber tracking in crossing regions. CONCLUSIONS: The algorithms based on the newly suggested methods were succesfuly implemented, their performance was tested on both synthetic and real data, and compared to some of the previously known approaches.
