IEEE J Biomed Health Inform - Systematic Poisoning Attacks on and Defenses for Machine Learning in Healthcare.

Tópicos

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

Resumo

Machine learning is being used in a wide range of application domains to discover patterns in large datasets. Increasingly, the results of machine learning drive critical decisions in applications related to healthcare and biomedicine. Such health-related applications are often sensitive and, thus, any security breach would be catastrophic. Naturally, the integrity of the results computed by machine learning is of great importance. Recent research has shown that some machine learning algorithms can be compromised by augmenting their training datasets with malicious data, leading to a new class of attacks called poisoning attacks. Hindrance of a diagnosis may have lifethreatening consequences and could cause distrust. On the other hand, not only may a false diagnosis prompt users to distrust the machine learning algorithm and even abandon the entire system but also such a false positive classification may cause patient distress. In this paper, we present a systematic, algorithmindependent approach for mounting poisoning attacks across a wide range of machine learning algorithms and healthcare datasets. The proposed attack procedure generates input data, which, when added to the training set, can either cause the results of machine learning to have targeted errors (e.g., increase the likelihood of classification into a specific class), or simply introduce arbitrary errors (incorrect classification). These attacks may be applied to both fixed and evolving datasets. They can be applied even when only statistics of the training dataset are available or, in some cases, even without access to the training dataset, although at a lower efficacy. We establish the effectiveness of the proposed attacks using a suite of six machine learning algorithms and five healthcare datasets. Finally, we present countermeasures against the proposed generic attacks that are based on tracking and detecting deviations in various accuracy metrics, and benchmark their effectiveness.

Resumo Limpo

machin learn use wide rang applic domain discov pattern larg dataset increas result machin learn drive critic decis applic relat healthcar biomedicin healthrel applic often sensit thus secur breach catastroph natur integr result comput machin learn great import recent research shown machin learn algorithm can compromis augment train dataset malici data lead new class attack call poison attack hindranc diagnosi may lifethreaten consequ caus distrust hand may fals diagnosi prompt user distrust machin learn algorithm even abandon entir system also fals posit classif may caus patient distress paper present systemat algorithmindepend approach mount poison attack across wide rang machin learn algorithm healthcar dataset propos attack procedur generat input data ad train set can either caus result machin learn target error eg increas likelihood classif specif class simpli introduc arbitrari error incorrect classif attack may appli fix evolv dataset can appli even statist train dataset avail case even without access train dataset although lower efficaci establish effect propos attack use suit six machin learn algorithm five healthcar dataset final present countermeasur propos generic attack base track detect deviat various accuraci metric benchmark effect

Resumos Similares

J. Comput. Biol. - Imbalanced class learning in epigenetics. ( 0,860951922033025 )
Comput Math Methods Med - On multilabel classification methods of incompletely labeled biomedical text data. ( 0,846166142783984 )
J Am Med Inform Assoc - Learning classification models with soft-label information. ( 0,82994739253273 )
IEEE Trans Neural Netw Learn Syst - A Kernel Classification Framework for Metric Learning. ( 0,807602921712611 )
J Med Syst - 3D similarity-dissimilarity plot for high dimensional data visualization in the context of biomedical pattern classification. ( 0,807059536621782 )
IEEE Trans Image Process - A linear support higher-order tensor machine for classification. ( 0,803952452575975 )
J Biomed Inform - Semi-supervised clinical text classification with Laplacian SVMs: an application to cancer case management. ( 0,800494879628803 )
Int J Neural Syst - Span: spike pattern association neuron for learning spatio-temporal spike patterns. ( 0,795415518968898 )
Neural Comput - Adaptive metric learning vector quantization for ordinal classification. ( 0,795247678753653 )
IEEE Trans Image Process - Task-specific image partitioning. ( 0,792639148259509 )
IEEE Trans Image Process - Geodesic propagation for semantic labeling. ( 0,780596212899778 )
IEEE Trans Image Process - Manifold regularized multitask learning for semi-supervised multilabel image classification. ( 0,775151055597945 )
Neural Comput - Computing sparse representations of multidimensional signals using Kronecker bases. ( 0,774669041223998 )
Neural Comput - Reduction from cost-sensitive ordinal ranking to weighted binary classification. ( 0,76660335116507 )
AMIA Annu Symp Proc - Comparison and combination of several MeSH indexing approaches. ( 0,766187140277646 )
Neural Comput - Metacognitive learning in a fully complex-valued radial basis function neural network. ( 0,76449885064755 )
J Biomed Inform - Incremental Gaussian Discriminant Analysis based on Graybill and Deal weighted combination of estimators for brain tumour diagnosis. ( 0,763577008272858 )
Int J Neural Syst - Linear time relational prototype based learning. ( 0,762408203656416 )
J Am Med Inform Assoc - Active learning for clinical text classification: is it better than random sampling? ( 0,759163080073394 )
IEEE Trans Image Process - Hyperspectral image classification through bilayer graph-based learning. ( 0,75906827205469 )
Int J Neural Syst - Aggregation of sparse linear discriminant analyses for event-related potential classification in brain-computer interface. ( 0,752966268722457 )
IEEE Trans Pattern Anal Mach Intell - Distance-Based Image Classification: Generalizing to New Classes at Near Zero Cost. ( 0,752709278843609 )
Comput. Biol. Med. - Robust prediction of protein subcellular localization combining PCA and WSVMs. ( 0,751941479380213 )
IEEE Trans Image Process - Multiview Hessian regularization for image annotation. ( 0,751390615108355 )
BMC Med Inform Decis Mak - Learning to improve medical decision making from imbalanced data without a priori cost. ( 0,749506814750696 )
J Biomed Inform - Class proximity measures--dissimilarity-based classification and display of high-dimensional data. ( 0,749356784079316 )
Neural Comput - Divergence-based vector quantization. ( 0,749315267317929 )
IEEE Trans Pattern Anal Mach Intell - Feature Selection with Conjunctions of Decision Stumps and Learning from Microarray Data. ( 0,747831554223301 )
J Chem Inf Model - Training based on ligand efficiency improves prediction of bioactivities of ligands and drug target proteins in a machine learning approach. ( 0,747268560121959 )
IEEE Trans Image Process - Improving Web image search by bag-based reranking. ( 0,746748567096239 )
IEEE Trans Neural Netw Learn Syst - Adaptive Batch Mode Active Learning. ( 0,746167219544476 )
IEEE Trans Pattern Anal Mach Intell - Weakly Supervised Recognition of Daily Life Activities with Wearable Sensors. ( 0,744303416216303 )
IEEE Trans Image Process - Joint segmentation of images and scanned point cloud in large-scale street scenes with low-annotation cost. ( 0,74390243902439 )
J Biomed Inform - Multi-label classification of chronically ill patients with bag of words and supervised dimensionality reduction algorithms. ( 0,742798784480664 )
Int J Neural Syst - Structurally enhanced incremental neural learning for image classification with subgraph extraction. ( 0,741549420830729 )
Neural Comput - Online learning with (multiple) kernels: a review. ( 0,740650725839511 )
Int J Neural Syst - Online semi-supervised growing neural gas. ( 0,739350170714108 )
AMIA Annu Symp Proc - Sample-efficient learning with auxiliary class-label information. ( 0,738961948128388 )
IEEE Trans Image Process - Structured max-margin learning for inter-related classifier training and multilabel image annotation. ( 0,737207900398004 )
IEEE Trans Pattern Anal Mach Intell - Representation Learning: A Review and New Perspectives. ( 0,734711194461544 )
IEEE Trans Image Process - Active learning for solving the incomplete data problem in facial age classification by the furthest nearest-neighbor criterion. ( 0,732102606940591 )
AMIA Annu Symp Proc - Outlier Detection with One-Class SVMs: An Application to Melanoma Prognosis. ( 0,73187366561542 )
IEEE Trans Neural Netw Learn Syst - ML-Tree: a tree-structure-based approach to multilabel learning. ( 0,729167728186261 )
Neural Comput - Multiple spectral kernel learning and a gaussian complexity computation. ( 0,72883195606086 )
IEEE Trans Image Process - Design of non-linear kernel dictionaries for object recognition. ( 0,728395181813778 )
Comput. Biol. Med. - Sparse Manifold Clustering and Embedding to discriminate gene expression profiles of glioblastoma and meningioma tumors. ( 0,721895863427393 )
J Biomed Inform - Learning classification models from multiple experts. ( 0,719043195076851 )
IEEE Trans Image Process - Incremental training of a detector using online sparse eigendecomposition. ( 0,716393192632989 )
BMC Med Inform Decis Mak - Towards case-based medical learning in radiological decision making using content-based image retrieval. ( 0,715699881646413 )
IEEE Trans Pattern Anal Mach Intell - Facial Age Estimation by Learning from Label Distributions. ( 0,71558129823374 )
IEEE Trans Image Process - Learning discriminative dictionary for group sparse representation. ( 0,715227548337545 )
IEEE Trans Neural Netw Learn Syst - An efficient topological distance-based tree kernel. ( 0,712950414152649 )
Neural Comput - Representing objects, relations, and sequences. ( 0,710872831368081 )
J Biomed Inform - Portable automatic text classification for adverse drug reaction detection via multi-corpus training. ( 0,709454189069523 )
IEEE Trans Image Process - Self-supervised online metric learning with low rank constraint for scene categorization. ( 0,704703362161443 )
IEEE Trans Pattern Anal Mach Intell - Label Consistent K-SVD: Learning A Discriminative Dictionary for Recognition. ( 0,703120158989823 )
Comput Math Methods Med - Correlation kernels for support vector machines classification with applications in cancer data. ( 0,701729543373877 )
IEEE Trans Image Process - Unsupervised amplitude and texture classification of SAR images with multinomial latent model. ( 0,700451393960455 )
Artif Intell Med - Multi-objective evolutionary algorithms for fuzzy classification in survival prediction. ( 0,699509398758518 )
J Biomed Inform - Applying active learning to assertion classification of concepts in clinical text. ( 0,693399975671487 )
J. Comput. Biol. - Locally learning biomedical data using diffusion frames. ( 0,693011743654063 )
Comput Methods Programs Biomed - Modified CC-LR algorithm with three diverse feature sets for motor imagery tasks classification in EEG based brain-computer interface. ( 0,691242742660517 )
Artif Intell Med - Exploiting the systematic review protocol for classification of medical abstracts. ( 0,690567001856715 )
IEEE Trans Image Process - Supervised ordering in IRp: application to morphological processing of hyperspectral images. ( 0,688338373723118 )
AMIA Annu Symp Proc - Classification of medication status change in clinical narratives. ( 0,688290023027288 )
J Chem Inf Model - Note on naive Bayes based on binary descriptors in cheminformatics. ( 0,688112046811606 )
Neural Comput - Incremental learning by message passing in hierarchical temporal memory. ( 0,687827739924227 )
IEEE Trans Image Process - Multiple-kernel, multiple-instance similarity features for efficient visual object detection. ( 0,68658713810347 )
IEEE Trans Image Process - Saliency and gist features for target detection in satellite images. ( 0,686193561078639 )
IEEE Trans Pattern Anal Mach Intell - Learning Categories from Few Examples with Multi Model Knowledge Transfer. ( 0,679518112875817 )
Neural Comput - Mismatched training and test distributions can outperform matched ones. ( 0,679064097696593 )
Comput Methods Programs Biomed - Multistage approach for clustering and classification of ECG data. ( 0,678905027644092 )
Int J Comput Assist Radiol Surg - Statistical shape model of a liver for autopsy imaging. ( 0,677534084976935 )
IEEE Trans Pattern Anal Mach Intell - A Bag-of-Features Framework to Classify Time Series. ( 0,676100807257033 )
IEEE Trans Pattern Anal Mach Intell - Exemplar-Based Colour Constancy and Multiple Illumination. ( 0,673958092999888 )
IEEE Trans Neural Netw Learn Syst - An Experimentation Platform for On-Chip Integration of Analog Neural Networks: A Pathway to Trusted and Robust Analog/RF ICs. ( 0,673914428252809 )
IEEE Trans Image Process - Contextual kernel and spectral methods for learning the semantics of images. ( 0,673805358552167 )
Neural Comput - Unsupervised learning of generative and discriminative weights encoding elementary image components in a predictive coding model of cortical function. ( 0,672143411417068 )
Artif Intell Med - A classifier ensemble approach for the missing feature problem. ( 0,671936850813724 )
J Biomed Inform - Reducing systematic review workload through certainty-based screening. ( 0,669861949321053 )
J Biomed Inform - Supervised methods for symptom name recognition in free-text clinical records of traditional Chinese medicine: an empirical study. ( 0,669720616356579 )
IEEE Trans Pattern Anal Mach Intell - Scene-Specific Pedestrian Detection for Static Video Surveillance. ( 0,667975337864478 )
Artif Intell Med - Screening nonrandomized studies for medical systematic reviews: a comparative study of classifiers. ( 0,667589311380717 )
Neural Comput - Adaptive multiclass classification for brain computer interfaces. ( 0,667247386759582 )
IEEE Trans Image Process - Artistic image analysis using graph-based learning approaches. ( 0,664749709248583 )
Comput Methods Programs Biomed - Biomedical system based on the Discrete Hidden Markov Model using the Rocchio-Genetic approach for the classification of internal carotid artery Doppler signals. ( 0,663945532805054 )
Comput Methods Programs Biomed - Auto-adaptive robot-aided therapy using machine learning techniques. ( 0,663746795775519 )
IEEE Trans Image Process - Real-time object tracking via online discriminative feature selection. ( 0,663502765249966 )
Comput Methods Programs Biomed - Machine learning algorithms and forced oscillation measurements applied to the automatic identification of chronic obstructive pulmonary disease. ( 0,66213887829402 )
J Chem Inf Model - Atom environment kernels on molecules. ( 0,660499435499719 )
IEEE Trans Pattern Anal Mach Intell - The Effect of Model Misspecification on Semi-Supervised Classification. ( 0,658034883676734 )
Med Decis Making - The Impact of Oversampling with SMOTE on the Performance of 3 Classifiers in Prediction of Type 2 Diabetes. ( 0,656174336648144 )
Comput. Biol. Med. - Identification of epilepsy stages from ECoG using genetic programming classifiers. ( 0,653605795658364 )
Neural Comput - Enhanced gradient for training restricted Boltzmann machines. ( 0,652956461607666 )
J Chem Inf Model - Modeling and benchmark data set for the inhibition of c-Jun N-terminal kinase-3. ( 0,652192754101344 )
BMC Med Inform Decis Mak - Decision tree-based learning to predict patient controlled analgesia consumption and readjustment. ( 0,651131419481152 )
IEEE Trans Image Process - Cross-Device Automated Prostate Cancer Localization With Multiparametric MRI. ( 0,649154306107951 )
J Med Syst - A study on determining the perception of learning organisation applications by health sector workers. ( 0,643964452332329 )
J Biomed Inform - Classifying temporal relations in clinical data: a hybrid, knowledge-rich approach. ( 0,641278150101831 )
IEEE Trans Neural Netw Learn Syst - Application of Reinforcement Learning Algorithms for the Adaptive Computation of the Smoothing Parameter for Probabilistic Neural Network. ( 0,639840794759375 )