PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “generative artificial intelligence”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 811 records · Page 45Linked to original sources

Approaches for creating computer-interpretable guidelines that facilitate decision support.

During the last decade, studies have shown the benefits of using clinical guidelines in the practice of medicine. Although the importance of these guidelines is widely recognized, health care organizations typically pay more attention to guideline development than to guideline implementation for routine use in daily care. However, studies have shown that clinicians are often not familiar with written guidelines and do not apply them appropriately during the actual care process. Implementing guidelines in computer-based decision support systems promises to improve the acceptance and application of guidelines in daily practice because the actions and observations of health care workers are monitored and advice is generated whenever a guideline is not followed. Such implementations are increasingly applied in diverse areas such as policy development, utilization management, education, clinical trials, and workflow facilitation. Many parties are developing computer-based guidelines as well as decision support systems that incorporate these guidelines. This paper reviews generic approaches for developing and implementing computer-based guidelines that facilitate decision support. It addresses guideline representation, acquisition, verification and execution aspects. The paper describes five approaches (the Arden Syntax, GuideLine Interchange Format (GLIF), PROforma, Asbru and EON), after the approaches are compared and discussed.

Artificial Intelligence↗

Genetic algorithm based system for patient scheduling in highly constrained situations.

In medicine and health care there are a lot of situations when patients have to be scheduled on different devices and/or with different physicians or therapists. It may concern preventive examinations, laboratory tests or convalescent therapies, therefore we are always looking for an optimal schedule that would result in finishing all the activities scheduled as soon as possible, with the least patient waiting time and maximum device utilization. Since patient scheduling is a highly complex problem, it is impossible to make a qualitative schedule by hand or even with exact heuristic methods. Therefore we developed a powerful automated scheduling method for highly constrained situations based on genetic algorithms and machine learning. In this paper we present the method, together with the whole process of schedule generation, the important parameters to direct the evolution and how the algorithm is guaranteed to produce only feasible solutions, not breaking any of the required constraints. We applied the described method to a problem of scheduling patients with different therapy needs to a limited number of therapeutic devices, but the algorithm can be easily modified for use in similar situations. The results are quite encouraging and since all the solutions are feasible, the method can be easily incorporated into an interactive user interface, which can be of major importance when scheduling patients, and human resources in general, is considered.

Algorithms↗

An SVM scorer for more sensitive and reliable peptide identification via tandem mass spectrometry.

Tandem mass spectrometry (MS/MS) has become increasingly important and indispensable in high-throughput proteomics for identifying complex protein mixtures. Database searching is the standard method to accomplish this purpose. A key sub-routine, peptide identification, is used to generate a list of candidate peptides from a protein database according to an experimental MS/MS spectrum, and then validate these candidate peptides for protein identification. Although currently there are many algorithms for peptide identification, most of them either lack an effective validation module or only validate the first-ranked peptide, thus leading to a low identification reliability or sensitivity. This paper proposes a new algorithm, named pepReap, to overcome the above drawbacks. It consists of a two-layered scoring scheme based on machine learning. The first layer is a rough scoring function which uses some simple and heuristic factors to measure the degree of the matches between an experimental MS/MS spectrum and the candidate peptides; thus a ranked list of candidate peptides is generated at a relatively low computational cost. The second layer is a fine scoring function which re-ranks the candidate peptides generated in the first layer and determines which one among them is the true positive. The fine scoring function was designed based on support vector machines (SVMs) using more comprehensive factors, such as the correlations between ions, the mass matching errors of fragment and peptide ions, etc. Consequently, the SVM classifier serves as not only a scorer but also a validation module. Experimental comparison with the popular SEQUEST algorithm coupled with threshold validation criteria on a reported dataset demonstrates that the pepReap algorithm achieves higher performance in terms of identification sensitivity with comparable precision.

Algorithms↗

Validation of a thermographic guidewire for endoluminal mapping of atherosclerotic disease: an in vitro study.

Temperature heterogeneity along the inner surface of an artery may be a surrogate marker of impending plaque rupture and has been associated with an increased likelihood of future coronary events. Initial studies using catheter-based thermographic devices have demonstrated that the changes in temperature are subtle, while the effects of coronary flow on measured temperature have not yet been examined. A novel guidewire-based system (ThermoCoil, Imetrx) designed to measure surface temperature in coronary arteries was used to study the effects of heat source intensity and flow on measured temperature. An in vitro model of a focal, eccentric, heat-generating lesion demonstrated that a guidewire-based system can detect changes in surface temperature with a precision of less than 0.08 degrees C. In this model, temperature measurements increased linearly with source temperature and decreased with increases in flow by an exponent of -0.33 (P < 0.001 for both). Flow rates and heat source properties can significantly influence the measurement and interpretation of thermographic data. The incorporation of 2D thermographic images may contribute further to the characterization of metabolically active plaques likely to cause acute coronary syndromes.

Artificial Intelligence↗

Validation of partial tissue segmentation of single-channel magnetic resonance images of the brain.

We describe and evaluate a practical, automated algorithm based on local statistical mixture modeling for segmenting single-channel, T1-weighted volumetric magnetic resonance images of the brain into gray matter, white matter, and cerebrospinal fluid. We employed a stereological sampling method to assess, prospectively, the performance of the method with respect to human experts on 10 normal T1-weighted brain scans acquired with a three-dimensional gradient echo pulse sequence. The overall kappa statistic for the concordance of the algorithm with the human experts was 0.806, while that among raters, excluding the algorithm, was 0.802. The algorithm had better agreement with the modal expert decision (kappa = 0.878). The algorithm could not be distinguished from the experts by this measure. We also validated the algorithm on a simulated MR scan of a digital brain phantom with known tissue composition. Global gray matter and white matter errors were 1% and <1%, respectively, and correlation coefficients with the underlying tissue model were 0.95 for gray matter, 0.98 for white matter, and 0.95 for cerebrospinal fluid. In both approaches to validation, we evaluated both local and global performance of the algorithm. Human experts generated slightly higher global gray matter proportion estimates on the test brain scans relative to the algorithm (3.7%) and on the simulated MR scan relative to the true tissue model (4.4%). The algorithm underestimated gray in some subcortical nuclei which contain admixed gray and white matter. We demonstrate the reliability of the method on individual 1 NEX data sets of the test subjects, and its insensitivity to the precise values of initial model parameters. The output of this algorithm is suitable for quantifying cerebral cortical tissue, using a commonly performed commercial pulse sequence.

Algorithms↗

BIOESTIM: software for automatic design of estimators in bioprocess engineering.

This paper describes BIOESTIM, a software package devoted to on-line estimation in bioprocess engineering. BIOESTIM enables bioengineers automatically to design state and parameter estimators from a minimal knowledge of the process kinetics. Such estimators allow development of software sensors capable of coping with the lack of reliable instrumentation suited to real-time monitoring. The estimator building procedure through BIOESTIM starts up from a dynamical material balance model of the bioprocess. This model, supplied by the user, is next completed by other information with no requirement for numerical values: the user has only to specify available measurements, coupled reactions and the known yield coefficients. On the base of this knowledge, BIOESTIM proceeds to symbolic algebraic manipulations on the model in order to study estimation possibilities and check identifiability of yield coefficients. When the design of an estimator is possible, the corresponding equations are automatically generated. Moreover, these estimators are stored in a user-specified file which is automatically interfaced with a specialized simulation software including data treatment and numerical integration packages. Thus, the user can simulate the estimator performances under various operational conditions using available experimental measurements. A typical example dealing with microbial growth and biosynthesis reactions is given in order to illustrate the main functional capabilities of BIOESTIM. BIOESTIM has been designed and written in a modular fashion. The module dealing with estimators design makes use of symbolic computation; it is written in Mathematica and runs on every computer on which this language is available.

Algorithms↗

The ontological basis of strong artificial life.

This article concerns the claim that it is possible to create living organisms, not merely models that represent organisms, simply by programming computers ("virtual" strong alife). I ask what sort of things these computer-generated organisms are supposed to be (where are they, and what are they made of?). I consider four possible answers to this question: (a) The organisms are abstract complexes of pure information; (b) they are material objects made of bits of computer hardware; (c) they are physical processes going on inside the computer; and (d) they are denizens of an entire artificial world, different from our own, that the programmer creates. I argue that (a) could not be right, that (c) collapses into (b) and that (d) would make strong alife either absurd or uninteresting. Thus, "virtual" strong alife amounts to the claim that, by programming a computer, one can literally bring bits of its hardware to life.

Artificial Intelligence↗

Multifacet structure of observed reconstructed integral images.

Three-dimensional images generated by an integral imaging system suffer from degradations in the form of grid of multiple facets. This multifacet structure breaks the continuity of the observed image and therefore reduces its visual quality. We perform an analysis of this effect and present the guidelines in the design of lenslet imaging parameters for optimization of viewing conditions with respect to the multifacet degradation. We consider the optimization of the system in terms of field of view, observer position and pupil function, lenslet parameters, and type of reconstruction. Numerical tests are presented to verify the theoretical analysis.

Algorithms↗

Protein secondary structure prediction using nearest-neighbor methods.

We have studied the use of nearest-neighbor classifiers to predict the secondary structure of proteins. The nearest-neighbor rule states that a test instance is classified according to the classifications of "nearby" training examples from a database of known structures. In the context of secondary structure prediction, the test instances are windows of n consecutive residues, and the label is the secondary structure type (alpha-helix, beta-strand, or coil) of the center position of the window. To define the neighborhood of a test instance, we employed a novel similarity metric based on the local structural environment scoring scheme of Bowie et al. In this manner, we have attempted to exploit the underlying structural similarity between segments of different proteins to aid in the prediction of secondary structure. Furthermore, in addition to using neighborhoods of fixed radius, we explored a modification of the standard nearest-neighbor algorithm that involved defining an "effective radius" for each exemplar by measuring its performance on a training set. Using these ideas, we achieved a peak prediction accuracy of 68%. Finally, we sought to improve the biological utility of secondary structure prediction by identifying the subset of the predictions that are most likely to be correct. Toward this end, we developed a nearest-neighbor estimator that produced not the traditional "one-state" prediction (alpha-helix, beta-strand, or coil) but rather a probability distribution over the three states. It should be emphasized that this scheme estimates true probability values and that the resulting numbers are not pseudo-probability scores generated by simple normalization of the raw output of the predictor. Applying the mutual information statistic, we found that these probability triplets possess 58% more information than the one-state predictions. Furthermore, the probability estimates allow one to assign an a priori confidence level to the prediction at each residue. Using this approach, we found that the top 28% of the predictions were 86% accurate and the top 43% of the predictions were 81% accurate. These results indicate that, notwithstanding the limitations on overall accuracy of secondary structure prediction, a substantial proportion of a protein can be predicted with considerable accuracy.

Algorithms↗

Improved discrimination of breast lesions using selective sampling of segmented MR images.

OBJECTIVE: The aim of this work is to examine if the specificity of differentiation between malignant and benign tumours can be improved by retrospectively examining lesion-extracted distributions. A semi-automated method for selecting a region-of-interest (ROI) is described. A new histogram segmentation approach for sampling pharmacokinetic breast maps of transfer uptake is defined in order to assign classification variables for the lesion. METHOD: Fifty exchange rate parameter maps were extracted from 49 subjects and retrospectively analysed. Distributions obtained from semi-automatically delineated ROIs were subdivided into ten overlapping segments. Parameters were extracted from each segment which effectively presents a new pixel intensity sampling strategy. Mann-Whitney non-parametric tests and ROC curves were generated. RESULTS: Correlation exists between mean parameter values drawn from semi-automatically or manually drawn ROIs. However, the former yield higher specificity values as applied to this subset of enhancing benign lesions. Segmenting the exchange rate parameter histogram allows the identification of which part of the distribution correlates most with tumour type. Significant improvement in specificity is obtained when using half the pixels within the ROI. CONCLUSION: Improved specificity values are obtained by a new method of selecting the differentiation parameters which relies on intensity rather than spatial segmentation. Only half the pixels available within the ROI contributed to the measured classification parameters.

Adult↗

Velocity-image model for online signature verification.

In general, online signature capturing devices provide outputs in the form of shape and velocity signals. In the past, strokes have been extracted while tracking velocity signal minimas. However, the resulting strokes are larger and complicated in shape and thus make the subsequent job of generating a discriminative template difficult. We propose a new stroke-based algorithm that splits velocity signal into various bands. Based on these bands, strokes are extracted which are smaller and more simpler in nature. Training of our proposed system revealed that low- and high-velocity bands of the signal are unstable, whereas the medium-velocity band can be used for discrimination purposes. Euclidean distances of strokes extracted on the basis of medium velocity band are used for verification purpose. The experiments conducted show improvement in discriminative capability of the proposed stroke-based system.

Algorithms↗

Hierarchical planning for a surface mounting machine placement.

For a surface mounting machine (SMM) in printed circuit board (PCB) assembly line, there are four problems, e.g. CAD data conversion, nozzle selection, feeder assignment and placement sequence determination. A hierarchical planning for them to maximize the throughput rate of an SMM is presented here. To minimize set-up time, a CAD data conversion system was first applied that could automatically generate the data for machine placement from CAD design data files. Then an effective nozzle selection approach implemented to minimize the time of nozzle changing. And then, to minimize picking time, an algorithm for feeder assignment was used to make picking multiple components simultaneously as much as possible. Finally, in order to shorten pick-and-place time, a heuristic algorithm was used to determine optimal component placement sequence according to the decided feeder positions. Experiments were conducted on a four head SMM. The experimental results were used to analyse the assembly line performance.

Algorithms↗

Approach to a hospital-based application of a medical expert system.

CADIAG-2 is a data-driven fuzzy medical expert system built for computer-based consultation in internal medicine. It has been integrated into the medical information system WAMIS of the Vienna General Hospital. Through the integration, CADIAG-2 is able to access patient data and laboratory test results already collected in the central patient database of WAMIS. CADIAG-2 operates in two subsequent phases: (1) as an automatic screening procedure for detecting pathological states in the patient, for generating diagnostic hypotheses, and for proposing further useful examinations; and (2) as an on-line consultation system for the clinician to assist him in clarifying patients' disorders completely and in great detail. At present, CADIAG-2 is subject to extended clinical trials. There are four application areas: rheumatic diseases, pancreatic diseases, gall bladder and bile duct diseases and colon diseases. First results on the performance of CADIAG-2 based on the evaluation of about 500 cases were obtained. In this paper, the principal goals, main components and concepts of CADIAG-2, and the experience gained until now are discussed.

Artificial Intelligence↗

The potential of latent semantic analysis for machine grading of clinical case summaries.

OBJECTIVE: This paper introduces latent semantic analysis (LSA), a machine learning method for representing the meaning of words, sentences, and texts. LSA induces a high-dimensional semantic space from reading a very large amount of texts. The meaning of words and texts can be represented as vectors in this space and hence can be compared automatically and objectively. PSYCHOLOGICAL THEORY: A generative theory of the mental lexicon based on LSA is described. The word vectors LSA constructs are context free, and each word, irrespective of how many meanings or senses it has, is represented by a single vector. However, when a word is used in different contexts, context appropriate word senses emerge. CURRENT APPLICATIONS: Several applications of LSA to educational software are described, involving the ability of LSA to quickly compare the content of texts, such as an essay written by a student and a target essay. POTENTIAL MEDICAL APPLICATIONS: An LSA-based software tool is sketched for machine grading of clinical case summaries written by medical students.

Artificial Intelligence↗

Artificial neural network analysis for evaluation of peptide MS/MS spectra in proteomics.

The aim of the work was to explore usefulness of artificial neural network (ANN) analysis for the evaluation of proteomics data. The analysis was applied to the data generated by the widely used protein identification program Sequest, completed with several structural parameters readily calculated from peptide molecular formulas. Proteins from yeast cells were identified based on the MS/MS spectra of peptides. The constructed ANN was demonstrated to classify automatically as either "good" or "bad" the peptide MS/MS spectra otherwise classified manually. An appropriately trained ANN proves to be a high-throughput tool facilitating examination of Sequest's results. ANNs are recommended as a means of automatic processing of large amounts of MS/MS data, which normally must be considered in the analysis of complex mixtures of proteins in proteomics.

Artificial Intelligence↗

Multidimensional support vector machines for visualization of gene expression data.

MOTIVATION: Since DNA microarray experiments provide us with huge amount of gene expression data, they should be analyzed with statistical methods to extract the meanings of experimental results. Some dimensionality reduction methods such as Principal Component Analysis (PCA) are used to roughly visualize the distribution of high dimensional gene expression data. However, in the case of binary classification of gene expression data, PCA does not utilize class information when choosing axes. Thus clearly separable data in the original space may not be so in the reduced space used in PCA. RESULTS: For visualization and class prediction of gene expression data, we have developed a new SVM-based method called multidimensional SVMs, that generate multiple orthogonal axes. This method projects high dimensional data into lower dimensional space to exhibit properties of the data clearly and to visualize a distribution of the data roughly. Furthermore, the multiple axes can be used for class prediction. The basic properties of conventional SVMs are retained in our method: solutions of mathematical programming are sparse, and nonlinear classification is implemented implicitly through the use of kernel functions. The application of our method to the experimentally obtained gene expression datasets for patients' samples indicates that our algorithm is efficient and useful for visualization and class prediction. CONTACT: komura@hal.rcast.u-tokyo.ac.jp.

Algorithms↗

Sparse Bayesian learning for efficient visual tracking.

This paper extends the use of statistical learning algorithms for object localization. It has been shown that object recognizers using kernel-SVMs can be elegantly adapted to localization by means of spatial perturbation of the SVM. While this SVM applies to each frame of a video independently of other frames, the benefits of temporal fusion of data are well-known. This is addressed here by using a fully probabilistic Relevance Vector Machine (RVM) to generate observations with Gaussian distributions that can be fused over time. Rather than adapting a recognizer, we build a displacement expert which directly estimates displacement from the target region. An object detector is used in tandem, for object verification, providing the capability for automatic initialization and recovery. This approach is demonstrated in real-time tracking systems where the sparsity of the RVM means that only a fraction of CPU time is required to track at frame rate. An experimental evaluation compares this approach to the state of the art showing it to be a viable method for long-term region tracking.

Algorithms↗

Learning and alignment methods applied to protein structure prediction.

Learning techniques are able to extract structural knowledge specific to a selected set of proteins. We describe two algorithms that optimize scores expressing the propensity of a polypeptide sequence to adopt a local fold. The first algorithm generates secondary structure prediction rules based on a dictionary of geometrical patterns frequently found in the learning database. The second algorithm leads to scores that indicate the fit between an amino acid and a given local structural environment. Dynamic programming is then used to align structural information profiles by modifying the local mutation cost with the above learned functions. The main features of the system are exemplified on the structural prediction of the N-terminal domain of the CD4 antigen. Then the usefulness of additional 3-D information in the alignment is benchmarked on eight pairs of weakly homologous proteins.

Algorithms↗