An Adaptive Linear Filter model of procedural category learning

2021 ◽  
Author(s):  
Nicolas Marchant ◽  
Enrique Canessa ◽  
Sergio Chaigneau
Keyword(s):  
Category Learning ◽  
Learning Algorithm ◽  
Current Model ◽  
Corrective Feedback ◽  
Logistic Function ◽  
Prototype Model ◽  
Published Data ◽  
Linear Filter ◽  
Filter Model ◽  
Individual Level

We use a feature-based association model to fit grouped and individual level category learning and transfer data. The model assumes that people use corrective feedback to learn individual feature to categorization-criterion correlations and combine those correlations additively to produce classifications. The current model is an Adaptive Linear Filter (ALF) with logistic output function and Least Mean Squares learning algorithm. Categorization probabilities are computed by a logistic function. Our data spans over 31 published data sets. Both at grouped and individual level analysis levels, the model performs remarkably well, accounting for large amounts of available variances. Notably, the model achieves its fits with a very minimal number of free parameters. We discuss the ALF’s advantages as a model of procedural categorization, in terms of its simplicity, its ability to capture empirical trends and its ability to solve longstanding challenges to similar models. In particular, we discuss why the model is not equivalent to a prototype model, as previously thought.

Global magnesium supply in the food chain

2015 ◽  
Vol 66 (12) ◽  
pp. 1278 ◽  
Author(s):  
Diriba B. Kumssa ◽  
Edward J. M. Joy ◽  
E. Louise Ander ◽  
Michael J. Watts ◽  
Scott D. Young ◽  
...  
Keyword(s):  
Food Composition ◽  
Published Data ◽  
Mg Deficiency ◽  
Individual Level ◽  
Household Food ◽  
Dietary Surveys ◽  
Food Composition Data ◽  
Per Capita ◽  
Potential Inhibitors ◽  
Dietary Deficiency

Magnesium (Mg) is an essential mineral micronutrient in humans. Risks of dietary Mg deficiency are affected by the quantity of Mg ingested and its bioavailability, which is influenced by the consumption of other nutrients and ‘anti-nutrients’. Here, we assess global dietary Mg supplies and risks of dietary deficiency, including the influence of other nutrients. Food supply and food composition data were used to derive the amount of Mg available per capita at national levels. Supplies of Mg were compared with estimated national per capita average requirement ‘cut points’. In 2011, global weighted mean Mg supply was 613 ± 69 mg person–1 day–1 compared with a weighted estimated average requirement for Mg of 173 mg person–1 day–1. This indicates a low risk of dietary Mg deficiency of 0.26% based on supply. This contrasts with published data from national individual-level dietary surveys, which indicate greater Mg deficiency risks. However, individuals in high-income countries are likely to under-report food consumption, which could lead to overestimation of deficiency risks. Furthermore, estimates of deficiency risk based on supply do not account for potential inhibitors of Mg absorption, including calcium, phytic acid and oxalate, and do not consider household food wastage.

Bayesian Prediction in Hybrid Conjoint Analysis

2002 ◽  
Vol 39 (2) ◽  
pp. 253-261 ◽  
Author(s):  
Frenkel Ter Hofstede ◽  
Youngchan Kim ◽  
Michel Wedel
Keyword(s):  
Prior Information ◽  
Current Model ◽  
The Self ◽  
Attribute Importance ◽  
Individual Level ◽  
Special Cases ◽  
Full Profile ◽  
Credible Intervals ◽  
Competing Models ◽  
The Individual

The authors propose a general model that includes the effects of discrete and continuous heterogeneity as well as self-stated and derived attribute importance in hybrid conjoint studies. Rather than use the self-stated importances as prior information, as has been done in several previous approaches, the authors consider them data and therefore include them in the formulation of the likelihood, which helps investigate the relationship of self-stated and derived importances at the individual level. The authors formulate several special cases of the model and estimate them using the Gibbs sampler. The authors reanalyze Srinivasan and Park's (1997) data and show that the current model predicts real choices better than competing models do. The posterior credible intervals of the predictions of models with the different heterogeneity specifications overlap, so there is no clear superior specification of heterogeneity. However, when different sources of data are used—that is, full profile evaluations, self-stated importances, or both—clear differences arise in the accuracy of predictions. Moreover, the authors find that including the self-stated importances in the likelihood leads to much better predictions than does considering them prior information.

A FILTER PROPOSAL FOR INCLUDING FEATURE CONSTRUCTION IN A GENETIC LEARNING ALGORITHM

2012 ◽  
Vol 20 (supp02) ◽  
pp. 31-49 ◽  
Author(s):  
DAVID GARCIA ◽  
ANTONIO GONZALEZ ◽  
RAUL PEREZ
Keyword(s):  
Learning Algorithm ◽  
A Priori ◽  
Fuzzy Rule ◽  
Feature Construction ◽  
Rule Based ◽  
Filter Model ◽  
Information Measures ◽  
Construction Techniques ◽  
Basic Learning ◽  
Current Proposal

In system identification process often a predetermined set of features is used. However, in many cases it is difficult to know a priori whether the selected features were really the more appropriate ones. This is the reason why the feature construction techniques have been very interesting in many applications. Thus, the current proposal introduces the use of these techniques in order to improve the description of fuzzy rule-based systems. In particular, the idea is to include feature construction in a genetic learning algorithm. The construction of attributes in this study will be restricted to the inclusion of functions defined on the initial attributes of the system. Since the number of functions and the number of attributes can be very large, a filter model, based on the use of information measures, is introduced. In this way, the genetic algorithm only needs to explore the particular new features that may be of greater interest to the final identification of the system. In order to manage the knowledge provided by the new attributes based on the use of functions we propose a new model of rule by extending a basic learning fuzzy rule-based model. Finally, we show the experimental study associated with this work.

scholarly journals Considerations for an Individual-Level Population Notification System for Pandemic Response: A Review and Prototype (Preprint)

2020 ◽  
Author(s):  
Mohammad Nazmus Sakib ◽  
Zahid A Butt ◽  
Plinio Pelegrini Morita ◽  
Mark Oremus ◽  
Geoffrey T Fong ◽  
...  
Keyword(s):  
Text Message ◽  
Unintended Consequences ◽  
Prototype Model ◽  
Health Crisis ◽  
Notification System ◽  
Individual Level ◽  
The Public ◽  
Pandemic Response ◽  
Level 1 ◽  
Threat Levels

UNSTRUCTURED The outbreak of the coronavirus disease (COVID-19), caused by severe acute respiratory syndrome coronavirus 2, spread worldwide after its emergence in China. Whether rich or poor, all nations are struggling to cope with this new global health crisis. The speed of the threat’s emergence and the quick response required from public health authorities and the public itself makes evident the need for a major reform in pandemic surveillance and notification systems. The development and implementation of a graded, individual-level pandemic notification system could be an effective tool to combat future threats of epidemics. This paper describes a prototype model of such a notification system and its potential advantages and challenges for implementation. Similar to other emergency alerts, this system would include a number of threat levels (level 1-5) with a higher level indicating increasing severity and intensity of safety measures (eg, level 1: general hygiene, level 2: enhanced hygiene, level 3: physical distancing, level 4: shelter in place, and level 5: lockdown). The notifications would be transmitted to cellular devices via text message (for lower threat levels) or push notification (for higher threat levels). The notification system would allow the public to be informed about the threat level in real time and act accordingly in an organized manner. New Zealand and the United Kingdom have recently launched similar alert systems designed to coordinate the ongoing COVID-19 pandemic response more efficiently. Implementing such a system, however, faces multiple challenges. Extensive preparation and coordination among all levels of government and relevant sectors are required. Additionally, such systems may be effective primarily in countries where there exists at least moderate trust in government. Advance and ongoing public education about the nature of the system and its steps would be an essential part of the system, such that all members of the public understand the meaning of each step in advance, similar to what has been established in systems for other emergency responses. This educational component is of utmost importance to minimize adverse public reaction and unintended consequences. The use of mass media and local communities could be considered where mobile phone penetration is low. The implementation of such a notification system would be more challenging in developing countries for several reasons, including inadequate technology, limited use of data plans, high population density, poverty, mistrust in government, and tendency to ignore or failure to understand the warning messages. Despite the challenges, an individual-level pandemic notification system could provide added benefits by giving an additional route for notification that would be complementary to existing platforms.

scholarly journals MetaLight: Value-Based Meta-Reinforcement Learning for Traffic Signal Control

2020 ◽  
Vol 34 (01) ◽  
pp. 1153-1160 ◽  
Author(s):  
Xinshi Zang ◽  
Huaxiu Yao ◽  
Guanjie Zheng ◽  
Nan Xu ◽  
Kai Xu ◽  
...  
Keyword(s):  
Reinforcement Learning ◽  
Real World ◽  
Learning Algorithm ◽  
Traffic Signal ◽  
Training Data ◽  
Signal Control ◽  
Traffic Signal Control ◽  
Individual Level ◽  
Real World Datasets ◽  
Reinforcement Learning Models

Using reinforcement learning for traffic signal control has attracted increasing interests recently. Various value-based reinforcement learning methods have been proposed to deal with this classical transportation problem and achieved better performances compared with traditional transportation methods. However, current reinforcement learning models rely on tremendous training data and computational resources, which may have bad consequences (e.g., traffic jams or accidents) in the real world. In traffic signal control, some algorithms have been proposed to empower quick learning from scratch, but little attention is paid to learning by transferring and reusing learned experience. In this paper, we propose a novel framework, named as MetaLight, to speed up the learning process in new scenarios by leveraging the knowledge learned from existing scenarios. MetaLight is a value-based meta-reinforcement learning workflow based on the representative gradient-based meta-learning algorithm (MAML), which includes periodically alternate individual-level adaptation and global-level adaptation. Moreover, MetaLight improves the-state-of-the-art reinforcement learning model FRAP in traffic signal control by optimizing its model structure and updating paradigm. The experiments on four real-world datasets show that our proposed MetaLight not only adapts more quickly and stably in new traffic scenarios, but also achieves better performance.

Evaluation and Analysis of Cardiovascular Function in Intensive Care Unit Patients by Ultrasound Image Segmentation Based on Deep Learning

2020 ◽  
Vol 10 (8) ◽  
pp. 1892-1898
Author(s):  
Jiaqi Shen ◽  
Fangfang Huang ◽  
Myers Ulrich
Keyword(s):  
Image Segmentation ◽  
Deep Learning ◽  
Carotid Artery ◽  
Blood Vessel ◽  
Nerve Cell ◽  
Clinical Application ◽  
Learning Algorithm ◽  
Linear Filter ◽  
Multi Scale ◽  
Deep Learning Algorithm

Many studies have shown that cardiovascular disease has become one of the major diseases leading to death in the world. Therefore, it is a very meaningful topic to use image segmentation technology to segment blood vessels for clinical application. In order to automatically extract the features of blood vessel images in the process of segmentation, the deep learning algorithm is combined with image segmentation technology to segment the nerve cell membrane and carotid artery images of ICU patients, and to segment the blood vessel images from a multi-dimensional perspective. The relevant data are collected to observe the effect of this model. The results show that the three-dimensional multi-scale linear filter has a good effect on carotid artery segmentation in the image segmentation of nerve cell membranes and carotid artery. When analyzing the accuracy of vascular image segmentation from network parameters and training parameters, it is found that the accuracy of the threedimensional multi-scale linear filter can reach about 85%. Therefore, it can be found that the combination of deep learning algorithm and image segmentation technology has a good segmentation effect, and the segmentation accuracy is also high. The experiment achieves the desired effect, which provides experimental basis for the clinical application of the vascular image segmentation technology.

Predictive Data Modeling Using sp-kNN for Risk Factor Evaluation in Urban Demographical Healthcare Data

2021 ◽  
Vol 11 (1) ◽  
pp. 7-14
Author(s):  
Uzair Aslam Bhatti ◽  
Linwang Yuan ◽  
Zhaoyuan Yu ◽  
Saqib Ali Nawaz ◽  
Anum Mehmood ◽  
...  
Keyword(s):  
Machine Learning ◽  
Clinical Decision Making ◽  
Learning Algorithm ◽  
Clinical Decision ◽  
Eye Diseases ◽  
Machine Learning Algorithms ◽  
Medical Resources ◽  
Individual Level ◽  
Healthcare Data ◽  
Increasing Demand

Healthcare diseases are spreading all around the globe day to day. Hospital datasets are full from the data with much information. It's an urgent requirement to use that data perfectly and efficiently. We propose a novel algorithm for predictive model for eye diseases using KNN with machine learning algorithms and artificial intelligence (AI). The aims are to evaluate the connection between the accumulated preoperative risk variables and different eye diseases and to manufacture a model that can anticipate the results on an individual level, thus giving relevance to impactful factors and geographic and demographic features. Risk factors of the desired diseases were calculated and machine learning algorithm applied to provide the prediction of the diseases. Health monitoring is an economic discipline that focuses on the effective allocation of medical resources, mainly to maximize the benefits of society to health through the available resources. With the increasing demand for medical services and the limited allocation of medical resources, the application of health economics in clinical practice has been paid more and more attention, and it has gradually played an important role in clinical decision-making.

scholarly journals Towards reconstruction of the flow duration curve: development of a conceptual framework with a physical basis

2011 ◽  
Vol 8 (2) ◽  
pp. 3961-3992 ◽  
Author(s):  
Y. Yokoo ◽  
M. Sivapalan
Keyword(s):  
Conceptual Framework ◽  
Sensitivity Analyses ◽  
Linear Filter ◽  
Flow Duration Curve ◽  
Rainfall Runoff ◽  
Filter Model ◽  
Flow Component ◽  
Flow Duration ◽  
Rainfall Runoff Model ◽  
Runoff Model

Abstract. In this paper we investigate the climatic and landscape controls on the flow duration curve (FDC) with the use of a physically-based rainfall-runoff model. The FDC is a stochastic representation of within-year variability of runoff, which arises from the transformation, by the catchment, of within-year variability of precipitation that can itself be characterized by a corresponding duration curve for precipitation (PFDC). Numerical simulations are carried out with the rainfall-runoff model under a variety of combinations of climatic inputs (i.e., precipitation, potential evaporation, including their within-year variability) and landscape properties (i.e., soil type and depth). The simulations indicated that the FDC can be disaggregated into two components, with sharply differing characteristics and origins: the FDC for surface (fast) runoff (SFDC) and the FDC for subsurface (slow) runoff (SSFDC). SFDC closely tracked PFDC and can be approximated with the use of a simple, nonlinear (threshold) filter model. On the other hand, SSFDC tracked the FDC that is constructed from the regime curve (ensemble mean within-year variation of streamflow), which can be closely approximated by a linear filter model. Sensitivity analyses were carried out to understand the climate and landscape controls on each component, gaining useful physical insights into their respective shapes. In particular the results suggested that evaporation from dynamic saturated areas, especially in the dry season, can contribute to a sharp dip at the lower tail of the FDCs. Based on these results, we develop a conceptual framework for the reconstruction of FDCs in ungauged basins. This framework partitions the FDC into: (1) a fast flow component, governed by a filtered version of PFDC, (2) a slow flow component governed by the regime curve, and (3) a correction to SSFDC to capture the effects of high evapotranspiration at low flows.

scholarly journals Learning to Personalize Medicine from Aggregate Data

2020 ◽  
Author(s):  
Rich Colbaugh ◽  
Kristin Glass
Keyword(s):  
Clinical Trials ◽  
Personalized Medicine ◽  
Prediction Models ◽  
Predictive Analytics ◽  
Learning Algorithm ◽  
Individual Characteristics ◽  
Training Data ◽  
Experimental Therapy ◽  
Individual Level ◽  
The Individual

AbstractThere is great interest in personalized medicine, in which treatment is tailored to the individual characteristics of patients. Achieving the objectives of precision healthcare will require clinically-grounded, evidence-based approaches, which in turn demands rigorous, scalable predictive analytics. Standard strategies for deriving prediction models for medicine involve acquiring ‘training’ data for large numbers of patients, labeling each patient according to the outcome of interest, and then using the labeled examples to learn to predict the outcome for new patients. Unfortunately, labeling individuals is time-consuming and expertise-intensive in medical applications and thus represents a major impediment to practical personalized medicine. We overcome this obstacle with a novel machine learning algorithm that enables individual-level prediction models to be induced from aggregate-level labeled data, which is readily-available in many health domains. The utility of the proposed learning methodology is demonstrated by: i.) leveraging US county-level mental health statistics to create a screening tool which detects individuals suffering from depression based upon their Twitter activity; ii.) designing a decision-support system that exploits aggregate clinical trials data on multiple sclerosis (MS) treatment to predict which therapy would work best for the presenting patient; iii.) employing group-level clinical trials data to induce a model able to find those MS patients likely to be helped by an experimental therapy.

scholarly journals A summary-statistics-based approach to examine the role of serotonin transporter promoter tandem repeat polymorphism in psychiatric phenotypes

2021 ◽  
Author(s):  
Arunabha Majumdar ◽  
Preksha Patel ◽  
Bogdan Pasaniuc ◽  
Roel A. Ophoff
Keyword(s):  
Serotonin Transporter ◽  
Tandem Repeat ◽  
Large Scale ◽  
Disease Risk ◽  
Learning Algorithm ◽  
Summary Statistics ◽  
Significant Evidence ◽  
Individual Level ◽  
Genome Wide

AbstractIn genetic studies of psychiatric disorders in the pre-genome-wide association study (GWAS) era, one of the most commonly studied loci is the serotonin transporter (SLC6A4) promoter polymorphism, a 43-base-pair insertion/deletion polymorphism in the promoter region (5-HTTLPR). The genetic association signals between 5-HTTLPR and psychiatric phenotypes, however, have been inconsistent across many studies. Since the polymorphism cannot be tested via available SNP arrays, we had previously proposed an efficient machine learning algorithm to predict the genotypes of 5-HTTLPR based on the genotypes of eight nearby SNPs, which requires access to individual-level genotype and phenotype data. To utilize the advantage of publicly available GWAS summary statistics obtained from studies with very large sample sizes, we develop a GWAS summary-statistics-based approach for testing the variable number of tandem repeat (VNTR) associations with various phenotypes. We first cross-verify the accuracy of the summary-statistics-based approach for 61 phenotypes in the UK Biobank. Since we observed a strong similarity between the predicted individual-level 5-HTTLPR genotype-based approach and the summary-statistics-based approach, we applied our method to the available neurobehavioral GWAS summary statistics data obtained from large-scale GWAS. We found no genome-wide significant evidence for association between 5-HTTLPR and any of the neurobehavioral traits. We did observe, however, genome-wide significant evidence for association between this locus and human adult height, BMI, and total cholesterol. Our summary-statistics-based approach provides a systematic way to examine the role of VNTRs and related types of genetic polymorphisms in disease risk and trait susceptibility of phenotypes for which large-scale GWAS summary statistics data are available.

Sign in / Sign up

Export Citation Format

Share Document