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 1,027 records · Page 57Linked to original sources

PubMed2004

The power of an integrated informatic and molecular approach to type 1 diabetes research.

Recent years have witnessed an explosive growth in available biological data. This includes a tremendous quantity of sequence data (e.g., biological structures, genetic and physical maps, pathways) generated by genome and transcriptome projects focused on humans, mice, and a multitude of other species. Diabetes research stands to greatly benefit from this data, which is distributed across public and private databases and the scientific literature. The increasing quantity and complexity of this biological data necessitates use of novel bioinformatics strategies for its efficient retrieval, analysis, and interpretation. Bioinformatic capability is becoming increasingly indispensable for fast and comprehensive analysis of biological data by diabetes researchers. There is great potential for diabetes scientists and clinicians to take advantage of recent bioinformatics and knowledge discovery developments to radically transform and advance this field of research. This paper will review advances in the field of bioinformatics relevant to diabetes research and preview a new specialty diabetes database, Diabetaeta, that we are creating to serve as a central bioinformatic portal for type 1 diabetes research, as well as serving as a public repository for beta cell gene and protein expression data.

Animals↗

PubMed1994

[Computer-assisted validation system applied to hematology: Valab-hemato].

Validation of laboratory reports is the ultimate step before transmission of results to the clinician. The biologist checks the intrinsic consistency of the data as well as their possible medical value that is liable to lead to other investigations. Such a policy, when performed on all the data, is time-consuming, boring and uncertain. This step may be simplified by the use of a computerized expert system. The computer assisted validation system presented here concerns routine haematology data (Valab-haemato). Like its predecessor devoted to clinical chemistry (Valab-Biochem) it is based on the performance of a powerful inference engine which generates a decision-making tree for each report according to the data. This adaptability gives the system a capacity very close to human reasoning. In its haematology version the system deals with many variables including sex, age, origin of the patient (hospital ward), and the haematological data (blood cell count, differential, reticulocyte count, various information drawn from microscope examination of the blood smear as well as any report concerning the blood sample, erythrocyte sedimentation rate). Previous data are also taken into account, as well as the normal ranges, the values beyond which no result can be automatically validated and the delta-check. Some information definitely prevents validation of the results, others can be validated if they have been previously approved. Whereas the method of reasoning is fixed, all items are changeable in order to adapt the system to the type of activity of the laboratory.(ABSTRACT TRUNCATED AT 250 WORDS)

Artificial Intelligence↗

PubMed1996

Temporal reasoning for diagnosis in a causal probabilistic knowledge base.

We have added temporal reasoning to the Heart Disease Program (HDP) to take advantage of the temporal constraints inherent in cardiovascular reasoning. Some processes take place over minutes while others take place over months or years and a strictly probabilistic formalism can generate hypotheses that are impossible given the temporal relationships involved. The HDP has temporal constraints on the causal relations specified in the knowledge base and temporal properties on the patient input provided by the user. These are used in two ways. First, they are used to constrain the generation of the pre-computed causal pathways through the model that speed the generation of hypotheses. Second, they are used to generate time intervals for the instantiated nodes in the hypotheses, which are matched and adjusted as nodes are added to each evolving hypothesis. This domain offers a number of challenges for temporal reasoning. Since the nature of diagnostic reasoning is inferring a causal explanation from the evidence, many of the temporal intervals have few constraints and the reasoning has to make maximum use of those that exist. Thus, the HDP uses a temporal interval representation that includes the earliest and latest beginning and ending specified by the constraints. Some of the disease states can be corrected but some of the manifestations may remain. For example, a valve disease such as aortic stenosis produces hypertrophy that remains long after the valve has been replaced. This requires multiple time intervals to account for the existing findings. This paper discusses the issues and solutions that have been developed for temporal reasoning integrated with a pseudo-Bayesian probabilistic network in this challenging domain for diagnosis.

Artificial Intelligence↗

PubMed2002

Automatic generation of a metamodel from an existing knowledge base to assist the development of a new analogous knowledge base.

Knowledge acquisition is a key step in the development of knowledge-based systems and methods have been proposed to help elicitating a domain-specific task model from a generic task model. We explored how an existing validated knowledge base (KB) represented by a decision tree could be automatically processed to infer a higher level domain-specific task model. On-codoc is a guideline-based decision support system applied to breast cancer therapy. Assuming task identity and ontological proximity between breast and lung cancer domains, the generalization of the breast can-cer KB should allow to build a metamodel to serve as a guide for the elaboration of a new specific KB on lung cancer. Two types of parametrized generalization methods based on tree structure simplification and ontological abstraction were used. We defined a similarity distance and a generalization coefficient to select the best metamodel identified as the closest to the original decision tree of the most generalized metamodels.

Artificial Intelligence↗

PubMed2006

The dynamic neural field approach to cognitive robotics.

This tutorial presents an architecture for autonomous robots to generate behavior in joint action tasks. To efficiently interact with another agent in solving a mutual task, a robot should be endowed with cognitive skills such as memory, decision making, action understanding and prediction. The proposed architecture is strongly inspired by our current understanding of the processing principles and the neuronal circuitry underlying these functionalities in the primate brain. As a mathematical framework, we use a coupled system of dynamic neural fields, each representing the basic functionality of neuronal populations in different brain areas. It implements goal-directed behavior in joint action as a continuous process that builds on the interpretation of observed movements in terms of the partner's action goal. We validate the architecture in two experimental paradigms: (1) a joint search task; (2) a reproduction of an observed or inferred end state of a grasping-placing sequence. We also review some of the mathematical results about dynamic neural fields that are important for the implementation work.

Animals↗

PubMed2004

Effects of classifier structures and training regimes on integrated segmentation and recognition of handwritten numeral strings.

In integrated segmentation and recognition of character strings, the underlying classifier is trained to be resistant to noncharacters. We evaluate the performance of state-of-the-art pattern classifiers of this kind. First, we build a baseline numeral string recognition system with simple but effective presegmentation. The classification scores of the candidate patterns generated by presegmentation are combined to evaluate the segmentation paths and the optimal path is found using the beam search strategy. Three neural classifiers, two discriminative density models, and two support vector classifiers are evaluated. Each classifier has some variations depending on the training strategy: maximum likelihood, discriminative learning both with and without noncharacter samples. The string recognition performances are evaluated on the numeral string images of the NIST Special Database 19 and the zipcode images of the CEDAR CDROM-1. The results show that noncharacter training is crucial for neural classifiers and support vector classifiers, whereas, for the discriminative density models, the regularization of parameters is important. The string recognition results compare favorably to the best ones reported in the literature though we totally ignored the geometric context. The best results were obtained using a support vector classifier, but the neural classifiers and discriminative density models show better trade-off between accuracy and computational overhead.

Algorithms↗

PubMed2006

Stereo using monocular cues within the tensor voting framework.

We address the fundamental problem of matching in two static images. The remaining challenges are related to occlusion and lack of texture. Our approach addresses these difficulties within a perceptual organization framework, considering both binocular and monocular cues. Initially, matching candidates for all pixels are generated by a combination of matching techniques. The matching candidates are then embedded in disparity space, where perceptual organization takes place in 3D neighborhoods and, thus, does not suffer from problems associated with scanline or image neighborhoods. The assumption is that correct matches produce salient, coherent surfaces, while wrong ones do not. Matching candidates that are consistent with the surfaces are kept and grouped into smooth layers. Thus, we achieve surface segmentation based on geometric and not photometric properties. Surface overextensions, which are due to occlusion, can be corrected by removing matches whose projections are not consistent in color with their neighbors of the same surface in both images. Finally, the projections of the refined surfaces on both images are used to obtain disparity hypotheses for unmatched pixels. The final disparities are selected after a second tensor voting stage, during which information is propagated from more reliable pixels to less reliable ones. We present results on widely used benchmark stereo pairs.

Algorithms↗

PubMed2004

Efficient generation of discontinuity-preserving adaptive triangulations from range images.

This paper presents an efficient technique for generating adaptive triangular meshes from range images. The algorithm consists of two stages. First, a user-defined number of points is adaptively sampled from the given range image. Those points are chosen by taking into account the surface shapes represented in the range image in such a way that points tend to group in areas of high curvature and to disperse in low-variation regions. This selection process is done through a noniterative, inherently parallel algorithm in order to gain efficiency. Once the image has been subsampled, the second stage applies a two and one half-dimensional Delaunay triangulation to obtain an initial triangular mesh. To favor the preservation of surface and orientation discontinuities (jump and crease edges) present in the original range image, the aforementioned triangular mesh is iteratively modified by applying an efficient edge flipping technique. Results with real range images show accurate triangular approximations of the given range images with low processing times.

Algorithms↗

PubMedDate not supplied

Scanning scene tunnel for city traversing.

This paper proposes a visual representation named scene tunnel for capturing urban scenes along routes and visualizing them on the Internet. We scan scenes with multiple cameras or a fish-eye camera on a moving vehicle, which generates a real scene archive along streets that is more complete than previously proposed route panoramas. Using a translating spherical eye, properly set planes of scanning, and unique parallel-central projection, we explore the image acquisition of the scene tunnel from camera selection and alignment, slit calculation, scene scanning, to image integration. The scene tunnels cover high buildings, ground, and various viewing directions and have uniformed resolutions along the street. The sequentially organized scene tunnel benefits texture mapping onto the urban models. We analyze the shape characteristics in the scene tunnels for designing visualization algorithms. After combining this with a global panorama and forward image caps, the capped scene tunnels can provide continuous views directly for virtual or real navigation in a city. We render scene tunnel dynamically by view warping, fast transmission, and flexible interaction. The compact and continuous scene tunnel facilitates model construction, data streaming, and seamless route traversing on the Internet and mobile devices.

Algorithms↗

PubMed1988

Self-organizing system obtaining communication ability. Primitive model for language generation.

As one of the synthetic approaches to brain functions, the possibility is discussed that two intellectual robots could make "words" for themselves and come to communicate with each other. Associatron, a model for associative memory with a neural network structure, is used as the memory in the brain, and information from the outer world is accumulated in it. Concepts regarding objects and attributes are extracted from the stored information, where the Associatron properties are utilized. The robots give specific names to the concepts and interchange them with each other. When the robots have the same experience at the same time, they modify their own words according to what the other robot says. As they repeat common experiences, the differences between the words become smaller and smaller, and finally the robots agree on a word for each object. At this stage, the robots can exchange information with their words and each can act accordingly on the word from the other. This study shows that the self-organizing system with the above function can be constructed as a neural network model.

Artificial Intelligence↗

PubMed2006

Localizing calcifications in cardiac CT data sets using a new vessel segmentation approach.

The new generation of multislice computed tomography (CT) scanners allows for the acquisition of high-resolution images of the heart. Based on that image data, the heart can be analyzed in a noninvasive way-improving the diagnosis of cardiovascular malfunctions on one hand, and the planning of an eventually necessary intervention on the other. One important parameter for the evaluation of the severity of a coronary artery disease is the number and localization of calcifications (hard plaques). This work presents a method for localizing these calcifications by employing a newly developed vessel segmentation approach. This extraction technique has been developed for, and tested with, contrast-enhanced CT data sets of the heart. The algorithm provides enough information to compute the vessel diameter along the extracted segment. An approach for automatically detecting calcified regions that combines diameter information and gray value analysis is presented. In addition, specially adapted methods for the visualization of these analysis results are described.

Algorithms↗

PubMed2005

Raman spectroscopy for diagnosis of atherosclerosis: a rapid analysis using neural networks.

Near-infrared Raman spectroscopy (NIRS) is one of the novel techniques that has a potential for in vivo diagnosis of atherosclerosis in human arteries. For such real time clinical applications, a rapid collection and analysis of the data is needed. One of the major problems with the fast data collection is that the noise generated by the detector has the same level as the Raman signal from the tissue, which makes the analysis difficult. In this work, NIRS measurements have been carried out on a total of 60 samples from human coronary arteries. Raman spectral data with the correlated histopathological analysis have been used as a basis to stimulate the cases of severe noise conditions. The main objective of this paper is the comparison of different processing algorithms that have been developed based on either wavelet transformation or principal component analysis for compressing the Raman spectral vectors and a rapid data classification based on different neural network architectures. The developed algorithms found to provide promising diagnosis results with classification errors smaller than 5%, even in the cases of Raman data with collection times as small as 20 ms. It has been concluded that the developed algorithms would be very much useful in the development of Raman spectroscopy systems for in vivo biological applications.

Algorithms↗

PubMed2003

Predicting hepatitis B virus-positive metastatic hepatocellular carcinomas using gene expression profiling and supervised machine learning.

Hepatocellular carcinoma (HCC) is one of the most common and aggressive human malignancies. Its high mortality rate is mainly a result of intra-hepatic metastases. We analyzed the expression profiles of HCC samples without or with intra-hepatic metastases. Using a supervised machine-learning algorithm, we generated for the first time a molecular signature that can classify metastatic HCC patients and identified genes that were relevant to metastasis and patient survival. We found that the gene expression signature of primary HCCs with accompanying metastasis was very similar to that of their corresponding metastases, implying that genes favoring metastasis progression were initiated in the primary tumors. Osteopontin, which was identified as a lead gene in the signature, was over-expressed in metastatic HCC; an osteopontin-specific antibody effectively blocked HCC cell invasion in vitro and inhibited pulmonary metastasis of HCC cells in nude mice. Thus, osteopontin acts as both a diagnostic marker and a potential therapeutic target for metastatic HCC.

Algorithms↗

PubMed2025

scPlantLLM: A Foundation Model for Exploring Single-cell Expression Atlases in Plants.

Single-cell RNA sequencing (scRNA-seq) provides unprecedented insights into plant cellular diversity by enabling high-resolution analyses of gene expression at the single-cell level. However, the complexity of scRNA-seq data, including challenges in batch integration, cell type annotation, and gene regulatory network (GRN) inference, demands advanced computational approaches. To address these challenges, we developed scPlantLLM, a Transformer model trained on millions of plant single-cell data points. Using a sequential pretraining strategy incorporating masked language modeling and cell type annotation tasks, scPlantLLM generates robust and interpretable single-cell data embeddings. When applied to Arabidopsis thaliana datasets, scPlantLLM excels in clustering, cell type annotation, and batch integration, achieving an accuracy of up to 0.91 in zero-shot learning scenarios. Furthermore, the model demonstrates an ability to identify biologically meaningful GRNs and subtle cellular subtypes, showcasing its potential to advance plant biology research. Compared to traditional methods, scPlantLLM outperforms in key metrics such as adjusted rand index (ARI), normalized mutual information (NMI), and silhouette score (SIL), highlighting its superior clustering accuracy and biological relevance. scPlantLLM represents a foundation model for exploring plant single-cell expression atlases, offering unprecedented capabilities to resolve cellular heterogeneity and regulatory dynamics across diverse plant systems. The code used in this study is available at https://github.com/compbioNJU/scPlantLLM.

Single-Cell Analysis↗

PubMed2005

Preliminary results of neural networks and zernike polynomials for classification of videokeratography maps.

PURPOSE: Our main goal in this work was to develop an artificial neural network (NN) that could classify specific types of corneal shapes using Zernike coefficients as input. Other authors have implemented successful NN systems in the past and have demonstrated their efficiency using different parameters. Our claim is that, given the increasing popularity of Zernike polynomials among the eye care community, this may be an interesting choice to add complementing value and precision to existing methods. By using a simple and well-documented corneal surface representation scheme, which relies on corneal elevation information, one can generate simple NN input parameters that are independent of curvature definition and that are also efficient. METHODS: We have used the Matlab Neural Network Toolbox (MathWorks, Natick, MA) to implement a three-layer feed-forward NN with 15 inputs and 5 outputs. A database from an EyeSys System 2000 (EyeSys Vision, Houston, TX) videokeratograph installed at the Escola Paulista de Medicina-Sao Paulo was used. This database contained an unknown number of corneal types. From this database, two specialists selected 80 corneas that could be clearly classified into five distinct categories: (1) normal, (2) with-the-rule astigmatism, (3) against-the-rule astigmatism, (4) keratoconus, and (5) post-laser-assisted in situ keratomileusis. The corneal height (SAG) information of the 80 data files was fit with the first 15 Vision Science and it Applications (VSIA) standard Zernike coefficients, which were individually used to feed the 15 neurons of the input layer. The five output neurons were associated with the five typical corneal shapes. A group of 40 cases was randomly selected from the larger group of 80 corneas and used as the training set. RESULTS: The NN responses were statistically analyzed in terms of sensitivity [true positive/(true positive + false negative)], specificity [true negative/(true negative + false positive)], and precision [(true positive + true negative)/total number of cases]. The mean values for these parameters were, respectively, 78.75, 97.81, and 94%. CONCLUSION: Although we have used a relatively small training and testing set, results presented here should be considered promising. They are certainly an indication of the potential of Zernike polynomials as reliable parameters, at least in the cases presented here, as input data for artificial intelligence automation of the diagnosis process of videokeratography examinations. This technique should facilitate the implementation and add value to the classification methods already available. We also discuss briefly certain special properties of Zernike polynomials that are what we think make them suitable as NN inputs for this type of application.

Artificial Intelligence↗

PubMed2004

Knowledge for better health: a conceptual framework and foundation for health research systems.

Health research generates knowledge that can be utilized to improve health system performance and, ultimately, health and health equity. We propose a conceptual framework for health research systems (HRSs) that defines their boundaries, components, goals, and functions. The framework adopts a systems perspective towards HRSs and serves as a foundation for constructing a practical approach to describe and analyse HRSs. The analysis of HRSs should, in turn, provide a better understanding of how research contributes to gains in health and health equity. In this framework, the intrinsic goals of the HRS are the advancement of scientific knowledge and the utilization of knowledge to improve health and health equity. Its four principal functions are stewardship, financing, creating and sustaining resources, and producing and using research. The framework, as it is applied in consultation with countries, will provide countries and donor agencies with relevant inputs to policies and strategies for strengthening HRSs and using knowledge for better health.

Artificial Intelligence↗

PubMed2006

Fuzzy theory approach for temporal model-based diagnosis: An application to medical domains.

OBJECTIVE: The aim of this work is to provide a theoretical framework which is sufficiently expressive to describe temporal evolution of diseases, and also to propose a diagnostic process for building explanations of patient's observed temporal evolution based on these disease descriptions. BACKGROUND: Model-based diagnosis (MBD) tackles the problem of troubleshooting systems by starting from a description of their structure and function (or behaviour). It is in this area where the use of deep causal models, as part of MBD systems, has shown its greater efficiency over classical rule based systems. From its beginnings, the temporal dimension was considered as an important component in MBD, since it makes it possible to define the dynamic behaviour. Several approaches have been proposed to represent time in MBD, enabling the representation of temporal concepts and relations, as well as the use of temporal reasoning mechanisms. METHODOLOGY: We first propose a temporal behavioural model (TBM), which allows us to capture the dynamics underlying temporal evolution of diseases and to include contextual information. Contextual information is required to model how contextual factors change the temporal evolution of diseases. The temporal component is modelled by fuzzy temporal constraints networks (FTCN), which makes the representation of quantitative and qualitative imprecise temporal information possible. We also provide a diagnostic process, which is based on a temporal adaptation of classical cover and differentiate method. RESULTS: The TBM and diagnostic process proposed provides a unique framework which addresses three problems not dealt with together so far: (a) the inclusion of contextual information, (b) the expressivity of the solution provided, and (c) the evaluation of the diagnostic hypotheses. This proposal demonstrates that the FTCN formalism provides mechanisms sufficiently expressive to cope with the intrinsic imprecision in the description of diseases' temporal evolution. The explanation generated provides the user with a complete picture of the temporal evolution of diseases and its causal links, thus allowing the appearance of repeated instances of the same disease through time. Mechanisms are provided which evaluate the credibility of alternative hypotheses, based on possibility theory. A prototype is presented along with a knowledge acquisition tool that guides medical experts in the model building process. CONCLUSIONS: In this paper, we propose a model that tightly couples methods from MBD area with constraint-based temporal reasoning techniques. The proposed model allows us to model complex contextual relationships in a compact way as well as providing solutions expressive enough to be used for decision support purposes. The solution provided conforms a causal network entailing the abnormal observations, including pathophysiological and etiological states. Furthermore, different instances of the same diagnostic hypotheses, located at different time instants, are also possible in the final solution. Finally, we provide an analysis of related and future works.

Artificial Intelligence↗

PubMed2002

A feed-forward network for input that is both categorical and quantitative.

The data on which a multi-layer perceptron (MLP) is to be trained to approximate a continuous function may have inputs that are categorical rather than numeric or quantitative such as color, gender, race, etc. A categorical variable causes a discontinuous relationship between an input variable and the output. A MLP, with connection matrices that multiply input values and sigmoid functions that further transform values, represents a continuous mapping in all input variables. A MLP therefore requires that all inputs correspond to numeric, continuously valued variables and represents a continuous function in all input variables. The way that this problem is usually dealt with is to replace the categorical values by numeric ones and treat them as if they were continuously valued. However, there is no meaningful correspondence between the continuous quantities generated this way and the original categorical values. Another approach is to encode the categorical portion of the input using 1-out-of-n encoding and include this code as input to the MLP. The approach in this paper is to segregate categorical variables from the continuous independent variables completely. The MLP is trained with multiple outputs; a separate output unit for each of the allowed combination of values of the categorical independent variables. During training the categorical value or combination of categorical values determines which of the output units should have the target value on it, with the remaining outputs being 'do not care'. Three data sets were used for comparison of methods. Results show that this approach is much more effective than the conventional approach of assigning continuous variables to the categorical features. In case of the data set where there were several categorical variables the method proposed here is also more effective than the 1-out-of-n input method.

Algorithms↗