AMIA Annu Symp Proc - Stochastic Gradient Descent and the Prediction of MeSH for PubMed Records.

Tópicos

{ problem(2511) optim(1539) algorithm(950) }
{ model(2656) set(1616) predict(1553) }
{ result(1111) use(1088) new(759) }
{ imag(1057) registr(996) error(939) }
{ featur(3375) classif(2383) classifi(1994) }
{ search(2224) databas(1162) retriev(909) }
{ howev(809) still(633) remain(590) }
{ drug(1928) target(777) effect(648) }
{ learn(2355) train(1041) set(1003) }
{ perform(999) metric(946) measur(919) }
{ data(3008) multipl(1320) sourc(1022) }
{ data(1714) softwar(1251) tool(1186) }
{ activ(1452) weight(1219) physic(1104) }
{ model(3404) distribut(989) bayesian(671) }
{ can(774) often(719) complex(702) }
{ assess(1506) score(1403) qualiti(1306) }
{ method(1969) cluster(1462) data(1082) }
{ data(1737) use(1416) pattern(1282) }
{ method(1219) similar(1157) match(930) }
{ chang(1828) time(1643) increas(1301) }
{ algorithm(1844) comput(1787) effici(935) }
{ extract(1171) text(1153) clinic(932) }
{ method(1557) propos(1049) approach(1037) }
{ studi(1410) differ(1259) use(1210) }
{ compound(1573) activ(1297) structur(1058) }
{ ehr(2073) health(1662) electron(1139) }
{ age(1611) year(1155) adult(843) }
{ use(976) code(926) identifi(902) }
{ survey(1388) particip(1329) question(1065) }
{ decis(3086) make(1611) patient(1517) }
{ detect(2391) sensit(1101) algorithm(908) }
{ imag(1947) propos(1133) code(1026) }
{ inform(2794) health(2639) internet(1427) }
{ system(1976) rule(880) can(841) }
{ measur(2081) correl(1212) valu(896) }
{ bind(1733) structur(1185) ligand(1036) }
{ sequenc(1873) structur(1644) protein(1328) }
{ imag(2830) propos(1344) filter(1198) }
{ network(2748) neural(1063) input(814) }
{ imag(2675) segment(2577) method(1081) }
{ patient(2315) diseas(1263) diabet(1191) }
{ take(945) account(800) differ(722) }
{ studi(2440) review(1878) systemat(933) }
{ motion(1329) object(1292) video(1091) }
{ treatment(1704) effect(941) patient(846) }
{ surgeri(1148) surgic(1085) robot(1054) }
{ framework(1458) process(801) describ(734) }
{ error(1145) method(1030) estim(1020) }
{ concept(1167) ontolog(924) domain(897) }
{ clinic(1479) use(1117) guidelin(835) }
{ design(1359) user(1324) use(1319) }
{ control(1307) perform(991) simul(935) }
{ model(2220) cell(1177) simul(1124) }
{ care(1570) inform(1187) nurs(1089) }
{ general(901) number(790) one(736) }
{ method(984) reconstruct(947) comput(926) }
{ featur(1941) imag(1645) propos(1176) }
{ case(1353) use(1143) diagnosi(1136) }
{ data(3963) clinic(1234) research(1004) }
{ risk(3053) factor(974) diseas(938) }
{ research(1085) discuss(1038) issu(1018) }
{ system(1050) medic(1026) inform(1018) }
{ import(1318) role(1303) understand(862) }
{ model(2341) predict(2261) use(1141) }
{ visual(1396) interact(850) tool(830) }
{ perform(1367) use(1326) method(1137) }
{ studi(1119) effect(1106) posit(819) }
{ blood(1257) pressur(1144) flow(957) }
{ spatial(1525) area(1432) region(1030) }
{ record(1888) medic(1808) patient(1693) }
{ health(3367) inform(1360) care(1135) }
{ model(3480) simul(1196) paramet(876) }
{ monitor(1329) mobil(1314) devic(1160) }
{ state(1844) use(1261) util(961) }
{ research(1218) medic(880) student(794) }
{ patient(2837) hospit(1953) medic(668) }
{ data(2317) use(1299) case(1017) }
{ medic(1828) order(1363) alert(1069) }
{ signal(2180) analysi(812) frequenc(800) }
{ cost(1906) reduc(1198) effect(832) }
{ group(2977) signific(1463) compar(1072) }
{ sampl(1606) size(1419) use(1276) }
{ gene(2352) biolog(1181) express(1162) }
{ first(2504) two(1366) second(1323) }
{ intervent(3218) particip(2042) group(1664) }
{ activ(1138) subject(705) human(624) }
{ time(1939) patient(1703) rate(768) }
{ patient(1821) servic(1111) care(1106) }
{ use(2086) technolog(871) perceiv(783) }
{ can(981) present(881) function(850) }
{ analysi(2126) use(1163) compon(1037) }
{ health(1844) social(1437) communiti(874) }
{ structur(1116) can(940) graph(676) }
{ high(1669) rate(1365) level(1280) }
{ cancer(2502) breast(956) screen(824) }
{ use(1733) differ(960) four(931) }
{ implement(1333) system(1263) develop(1122) }
{ estim(2440) model(1874) function(577) }
{ process(1125) use(805) approach(778) }
{ method(2212) result(1239) propos(1039) }

Resumo

Stochastic Gradient Descent (SGD) has gained popularity for solving large scale supervised machine learning problems. It provides a rapid method for minimizing a number of loss functions and is applicable to Support Vector Machine (SVM) and Logistic optimizations. However SGD does not provide a convenient stopping criterion. Generally an optimal number of iterations over the data may be determined using held out data. Here we compare stopping predictions based on held out data with simply stopping at a fixed number of iterations and show that the latter works as well as the former for a number of commonly studied text classification problems. In particular fixed stopping works well for MeSH(?) predictions on PubMed(?) records. We also surveyed the published algorithms for SVM learning on large data sets, and chose three for comparison: PROBE, SVMperf, and Liblinear and compared them with SGD with a fixed number of iterations. We find SGD with a fixed number of iterations performs as well as these alternative methods and is much faster to compute. As an application we made SGD-SVM predictions for all MeSH terms and used the Pool Adjacent Violators (PAV) algorithm to convert these predictions to probabilities. Such probabilistic predictions lead to ranked MeSH term predictions superior to previously published results on two test sets.

Resumo Limpo

stochast gradient descent sgd gain popular solv larg scale supervis machin learn problem provid rapid method minim number loss function applic support vector machin svm logist optim howev sgd provid conveni stop criterion general optim number iter data may determin use held data compar stop predict base held data simpli stop fix number iter show latter work well former number common studi text classif problem particular fix stop work well mesh predict pubm record also survey publish algorithm svm learn larg data set chose three comparison probe svmperf liblinear compar sgd fix number iter find sgd fix number iter perform well altern method much faster comput applic made sgdsvm predict mesh term use pool adjac violat pav algorithm convert predict probabl probabilist predict lead rank mesh term predict superior previous publish result two test set

Resumos Similares

IEEE Trans Image Process - Recovering missing slices of the discrete Fourier transform using Ghosts. ( 0,678672247861368 )
IEEE Trans Image Process - Multiview deblurring for 3-D images from light-sheet-based fluorescence microscopy. ( 0,65259353172423 )
IEEE Trans Image Process - Second-order optimization of mutual information for real-time image registration. ( 0,648975279956154 )
IEEE Trans Image Process - Modified gradient search for level set based image segmentation. ( 0,647778657398386 )
IEEE Trans Image Process - Efficient rate-distortion optimal packetization of embedded bitstreams into independent source packets. ( 0,641718496429828 )
IEEE Trans Pattern Anal Mach Intell - Hashing Hyperplane Queries to Near Points with Applications to Large-scale Active Learning. ( 0,641091312901208 )
IEEE Trans Neural Netw Learn Syst - Incremental Support Vector Learning for Ordinal Regression. ( 0,639387787797184 )
IEEE Trans Image Process - A generalized accelerated proximal gradient approach for total-variation-based image restoration. ( 0,624156373724458 )
Neural Comput - Alternating proximal regularized dictionary learning. ( 0,616096745361534 )
IEEE Trans Image Process - Saliency detection by multitask sparsity pursuit. ( 0,614014628066896 )
IEEE Trans Neural Netw Learn Syst - A one-class kernel fisher criterion for outlier detection. ( 0,607836959881912 )
IEEE Trans Image Process - Efficient algorithms for robust recovery of images from compressed data. ( 0,603296704658158 )
Neural Comput - Nondegenerate piecewise linear systems: a finite Newton algorithm and applications in machine learning. ( 0,602534273530943 )
IEEE Trans Neural Netw Learn Syst - Comparison of l1-Norm SVR and Sparse Coding Algorithms for Linear Regression. ( 0,601516737260138 )
Neural Comput - A proof of convergence of the concave-convex procedure using Zangwill's theory. ( 0,599509239146711 )
IEEE Trans Image Process - Fast image recovery using variable splitting and constrained optimization. ( 0,597505439686848 )
Neural Comput - High-dimensional cluster analysis with the masked EM algorithm. ( 0,595568463328027 )
IEEE Trans Image Process - Nonlocal regularization of inverse problems: a unified variational framework. ( 0,590784612343989 )
IEEE Trans Image Process - Efficient image classification via multiple rank regression. ( 0,58922908214736 )
IEEE Trans Neural Netw Learn Syst - On recursive edit distance kernels with application to time series classification. ( 0,58745028125588 )
J. Comput. Biol. - Border length minimization problem on a square array. ( 0,587280136888413 )
IEEE Trans Pattern Anal Mach Intell - Optimized Product Quantization. ( 0,58544203311272 )
IEEE Trans Image Process - Fast nonconvex nonsmooth minimization methods for image restoration and reconstruction. ( 0,585107727965163 )
IEEE Trans Image Process - Smoothed low rank and sparse matrix recovery by iteratively reweighted least squares minimization. ( 0,583441223658893 )
Med Biol Eng Comput - Genetic algorithm-based regularization parameter estimation for the inverse electrocardiography problem using multiple constraints. ( 0,582723317855932 )
IEEE Trans Image Process - Image decomposition with multilabel context: algorithms and applications. ( 0,582419041999793 )
IEEE Trans Neural Netw Learn Syst - A Neurodynamic Optimization Method for Recovery of Compressive Sensed Signals With Globally Converged Solution Approximating to l0 Minimization. ( 0,582320039470196 )
IEEE Trans Image Process - Restoration of Poissonian images using alternating direction optimization. ( 0,581025567275331 )
IEEE Trans Image Process - A unified feature and instance selection framework using optimum experimental design. ( 0,578882083725857 )
Neural Comput - Guaranteed classification via regularized similarity learning. ( 0,578382763899497 )
Comput Biol Chem - A hyper-heuristic for the Longest Common Subsequence problem. ( 0,577352138263148 )
Brief. Bioinformatics - Pattern recognition in bioinformatics. ( 0,577286961805968 )
IEEE Trans Pattern Anal Mach Intell - Linear Scale and Rotation Invariant Matching. ( 0,574511673191411 )
IEEE Trans Image Process - An alternating direction algorithm for total variation reconstruction of distributed parameters. ( 0,574278648592208 )
IEEE Trans Image Process - Improved image recovery from compressed data contaminated with impulsive noise. ( 0,57407372803394 )
IEEE Trans Neural Netw Learn Syst - Missile Guidance Law Based on Robust Model Predictive Control Using Neural-Network Optimization. ( 0,57405613773675 )
IEEE Trans Pattern Anal Mach Intell - Constrained Nonnegative Matrix Factorization for Image Representation. ( 0,573842369156724 )
IEEE Trans Pattern Anal Mach Intell - Reidentification by Relative Distance Comparison. ( 0,573594972443443 )
IEEE Trans Pattern Anal Mach Intell - Polynomial Eigenvalue Solutions to Minimal Problems in Computer Vision. ( 0,573255750005038 )
IEEE Trans Pattern Anal Mach Intell - A Variance Minimization Criterion to Feature Selection Using Laplacian Regularization. ( 0,57306350564722 )
J Chem Inf Model - Global optimization of binary Lennard-Jones clusters using three perturbation operators. ( 0,572788700251159 )
IEEE Trans Image Process - A comprehensive framework for image inpainting. ( 0,570743719439788 )
Neural Comput - Regularized variational Bayesian learning of echo state networks with delay&sum readout. ( 0,568967916315075 )
Artif Intell Med - Modeling and solving the dynamic patient admission scheduling problem under uncertainty. ( 0,563245604349628 )
IEEE Trans Image Process - Sparse stochastic processes and discretization of linear inverse problems. ( 0,562798249960381 )
Neural Comput - Sufficient dimension reduction via squared-loss mutual information estimation. ( 0,560889306668874 )
IEEE Trans Image Process - A primal-dual method for total-variation-based wavelet domain inpainting. ( 0,560046492474258 )
IEEE Trans Image Process - An iterative L1-based image restoration algorithm with an adaptive parameter estimation. ( 0,559018822658831 )
IEEE Trans Image Process - A majorize-minimize strategy for subspace optimization applied to image restoration. ( 0,556829802362852 )
IEEE Trans Image Process - Parallel proximal algorithm for image restoration using hybrid regularization. ( 0,556327985416849 )
IEEE Trans Image Process - Parameter selection for total-variation-based image restoration using discrepancy principle. ( 0,553436906156084 )
Comput Math Methods Med - A 3D finite-difference BiCG iterative solver with the Fourier-Jacobi preconditioner for the anisotropic EIT/EEG forward problem. ( 0,549817175655808 )
IEEE Trans Image Process - Discretization of parametrizable signal manifolds. ( 0,548839859987981 )
IEEE Trans Image Process - Shape-based normalized cuts using spectral relaxation for biomedical segmentation. ( 0,548668208388244 )
IEEE Trans Image Process - A fast adaptive parameter estimation for total variation image restoration. ( 0,548057921767938 )
IEEE Trans Image Process - Blind spectral unmixing based on sparse nonnegative matrix factorization. ( 0,54804848586338 )
Neural Comput - Prewhitening high-dimensional FMRI data sets without eigendecomposition. ( 0,547722557505166 )
Neural Comput - Robust subspace discovery via relaxed rank minimization. ( 0,547392009102465 )
Neural Comput - Alternating direction methods for latent variable gaussian graphical model selection. ( 0,546832148561633 )
IEEE Trans Image Process - Enhancement of coupled multichannel images using sparsity constraints. ( 0,544061946252947 )
J Chem Inf Model - GA(M)E-QSAR: a novel, fully automatic genetic-algorithm-(meta)-ensembles approach for binary classification in ligand-based drug design. ( 0,541342731232188 )
IEEE Trans Neural Netw Learn Syst - Learning With Mixed Hard/Soft Pointwise Constraints. ( 0,54087121884347 )
IEEE Trans Pattern Anal Mach Intell - Forward Basis Selection for Pursuing Sparse Representations Over a Dictionary. ( 0,540455212515056 )
IEEE Trans Image Process - Alternating minimization algorithm for speckle reduction with a shifting technique. ( 0,540250004451995 )
Comput. Biol. Med. - Relabeling algorithm for retrieval of noisy instances and improving prediction quality. ( 0,538898267373884 )
IEEE Trans Pattern Anal Mach Intell - Nonnegative Matrix Factorization with Earth Mover's Distance Metric for Image Analysis. ( 0,537995542132365 )
IEEE Trans Image Process - An augmented Lagrangian approach to the constrained optimization formulation of imaging inverse problems. ( 0,537595732331992 )
Neural Comput - Information-maximization clustering based on squared-loss mutual information. ( 0,537306245380066 )
Neural Comput - Input statistics and Hebbian cross-talk effects. ( 0,537168374176054 )
IEEE Trans Neural Netw Learn Syst - Scalable Nonparametric Low-Rank Kernel Learning Using Block Coordinate Descent. ( 0,536884014452145 )
IEEE Trans Image Process - On the complexity of mumford-shah-type regularization, viewed as a relaxed sparsity constraint. ( 0,535915455186534 )
IEEE Trans Pattern Anal Mach Intell - Maximum Correntropy Criterion for Robust Face Recognition. ( 0,53556899086579 )
IEEE Trans Pattern Anal Mach Intell - Learning with Augmented Features for Supervised and Semi-supervised Heterogeneous Domain Adaptation. ( 0,534974348502415 )
Comput Methods Programs Biomed - General bounds for electrode mislocation on the EEG inverse problem. ( 0,534536317043464 )
IEEE Trans Image Process - Alternating direction method for balanced image restoration. ( 0,534020666852603 )
J Integr Bioinform - Modelling proteolytic enzymes with Support Vector Machines. ( 0,533136065180333 )
IEEE Trans Image Process - An alternating minimization algorithm for binary image restoration. ( 0,532085068098352 )
Neural Comput - Linear coordinate-descent message passing for quadratic optimization. ( 0,531752266201186 )
Comput. Biol. Med. - Nonparametric optimization of constrained total variation for tomography reconstruction. ( 0,531005873728106 )
IEEE Trans Image Process - Inductive robust principal component analysis. ( 0,530810892956835 )
IEEE Trans Image Process - A multiplicative iterative algorithm for box-constrained penalized likelihood image restoration. ( 0,529787331893731 )
Neural Comput - A novel iterative method for computing generalized inverse. ( 0,52972997204394 )
IEEE Trans Neural Netw Learn Syst - Further result on guaranteed H8 performance state estimation of delayed static neural networks. ( 0,528837854626772 )
J Chem Inf Model - Large-scale learning of structure-activity relationships using a linear support vector machine and problem-specific metrics. ( 0,528418150380802 )
IEEE Trans Neural Netw Learn Syst - Sparse multivariate gaussian mixture regression. ( 0,527823032035334 )
Comput. Biol. Med. - Disocclusion-based 2D-3D registration for aortic interventions. ( 0,5256378348616 )
J Chem Inf Model - RS-Predictor models augmented with SMARTCyp reactivities: robust metabolic regioselectivity predictions for nine CYP isozymes. ( 0,524630644955025 )
IEEE Trans Image Process - Filtering in the diffeomorphism group and the registration of point sets. ( 0,524242846895996 )
IEEE Trans Neural Netw Learn Syst - Convergence and rate analysis of neural networks for sparse approximation. ( 0,524104950212566 )
IEEE Trans Image Process - Hessian Schatten-norm regularization for linear inverse problems. ( 0,522935787651056 )
Comput Math Methods Med - Finite element surface registration incorporating curvature, volume preservation, and statistical model information. ( 0,522411999942462 )
IEEE Trans Image Process - An iterative linear expansion of thresholds for l1-based image restoration. ( 0,521958957830274 )
IEEE Trans Image Process - Efficient algorithm for nonconvex minimization and its application to PM regularization. ( 0,52151342193518 )
J. Comput. Biol. - An improved satisfiability algorithm for nested canalyzing functions and its application to determining a singleton attractor of a Boolean network. ( 0,521321949281967 )
Neural Comput - Block clustering based on difference of convex functions (DC) programming and DC algorithms. ( 0,520007849234323 )
Brief. Bioinformatics - Fighting against uncertainty: an essential issue in bioinformatics. ( 0,519434020506051 )
IEEE Trans Neural Netw Learn Syst - Kernel reconstruction ICA for sparse representation. ( 0,519429716180518 )
Comput Math Methods Med - Mixed-norm regularization for brain decoding. ( 0,518974596584922 )
J Chem Inf Model - Ligand and structure-based classification models for prediction of P-glycoprotein inhibitors. ( 0,51894744983513 )
Comput. Biol. Med. - A note on hyper ellipse method for classifying biological and medical data. ( 0,518927920521628 )