The DOI matches its Crossref record; 221 of the 223 references it cites were found in a registry or an index.
Multiple instance learning: A survey of problem characteristics and applications
Open accessOverview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“With over 30 billion sequences in total, this represents one of the largest multiple instance learning experiments ever conducted (Carbonneau et al., 2018).”
From Hopfield Networks is All You Need (2020) · cites this paper as Carbonneau et al., 2018
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet Large Scale Visual Recognition Challenge
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy, A., Khosla, A., Bernstein, M., et al. (2015) · International Journal of Computer Vision · doi:10.1007/s11263-015-0816-y · Cited by 39,380
SMOTE synthetic minority over-sampling technique
Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002) · Journal of Artificial Intelligence Research · doi:10.1613/jair.953 · Cited by 32,865 · cited 2×
“There are external methods like SMOTE [193] and RUSBoost [194] that necessitate accurate labels to perform over or under sampling.”
MethodGreedy function approximation: A gradient boosting machine
Friedman, J. H. (2001) · The Annals of Statistics · doi:10.1214/aos/1013203451 · Cited by 30,731 · cited 1×
“The method is essentially the same as gradient boosting [143] except that the loss function is based on bag classification error.”
MethodThe pascal visual object classes (VOC) challenge
Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., & Zisserman, A. (2010) · International Journal of Computer Vision · doi:10.1007/s11263-009-0275-4 · Cited by 19,599
The WEKA data mining software: an update
Hall, M., Frank, E., Holmes, G., Pfahringer, B., Reutemann, P., & Witten, I. H. (2009) · ACM SIGKDD Explorations Newsletter · doi:10.1145/1656274.1656278 · Cited by 17,750
Representation Learning: A Review and New Perspectives
Bengio, Y., Courville, A., & Vincent, P. (2013) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2013.50 · Cited by 13,001 · cited 1×
“Recently, methods for representation learning [160] have gained in popularity because they usually yield a high level of accuracy.”
MethodStatistical Comparisons of Classifiers over Multiple Data Sets
Demšar, J. (2006) · doi:10.5555/1248547.1248548 · Cited by 12,051
Cross validatory choice and assessment of statistical predictions
Stone, M. (1974) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1974.tb00994.x · Cited by 10,881 · cited 1×
“The experiments were conducted using a nested cross-fold validation protocol [149].”
MethodBeyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories
Lazebnik, S., Schmid, C., & Ponce, J. (2006) · 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - Volume 2 (CVPR'06) · doi:10.1109/cvpr.2006.68 · Cited by 7,916 · cited 1×
“In BoW models for computer vision, this can be achieved by dividing the images [102, 103] or videos [80] into different spatial and/or temporal zones.”
The Earth Mover's Distance as a Metric for Image Retrieval
Rubner, Y., Tomasi, C., & Guibas, L. (2000) · International Journal of Computer Vision · doi:10.1023/a:1026543900054 · Cited by 5,547
Research works printed but not linked: 1
- Babenko, 2011, Robust object tracking with online multiple instance learning, IEEE Trans. Pattern Anal. Mach. Intell., 33, 1619, 10.1109/TPAMI.2010.226 · DOI · Paper page
Cited by
Sorted by citations
Items 201–300 of 734
Yang, Z., Wang, X., Xiang, J., Zhang, J., Yang, S., Wang, X., Yang, W., Li, Z., Han, X., & Liu, Y. (2023) · Archiv für Pathologische Anatomie und Physiologie und für Klinische Medicin · doi:10.1007/s00428-023-03502-z · Cited by 18
“These weakly supervised methods are developed based on the multiple instance learning (MIL) paradigm [23].”
Wu, Y., Qi, S., Feng, J., Chang, R., Pang, H., Hou, J., Li, M., Wang, Y., Xia, S., & Qian, W. (2023) · Journal of Applied Biomedicine · doi:10.1016/j.bbe.2023.06.004 · Cited by 18
Visual Correspondences for Unsupervised Domain Adaptation on Electron Microscopy Images
Bermúdez-Chacón, R., Altingövde, O., Becker, C., Salzmann, M., & Fua, P. (2019) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2019.2946462 · Cited by 18
“MIL is a natural formulation for problems where different versions of objects exist, individual labeling of each version is costly, and a single response per object is sufficient [34].”
BackgroundAnomaly detection with inexact labels
Iwata, T., Toyoda, M., Tora, S., & Ueda, N. (2020) · Machine Learning · doi:10.1007/s10994-020-05880-w · Cited by 18
“Existing multiple instance learning methods are not robust to class imbalance (Herrera et al., 2016; Carbonneau et al., 2018).”
BackgroundLian, W., Lindblad, J., Stark, C. R., Hirsch, J., & Sladoje, N. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.109498 · Cited by 17
Multi-attention multiple instance learning
Konstantinov, A. V., & Utkin, L. V. (2022) · Neural Computing and Applications · doi:10.1007/s00521-022-07259-5 · Cited by 16
“The most common condition is that negative bags contain only negative instances, and positive bags contain at least one positive instance [9].”
BackgroundDeep Hierarchical Multiple Instance Learning for Whole Slide Image Classification
Zhou, Y., & Lu, Y. (2022) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi52829.2022.9761678 · Cited by 17
Kim, Y., Wang, T., Xiong, D., Wang, X., & Park, S. (2022) · BMC Bioinformatics · doi:10.1186/s12859-022-05012-2 · Cited by 17
“The applications include biology and chemistry, computer vision, document classification, web mining, reinforcement learning, speech recognition, and time series classification [8].”
BackgroundMethodInfluentialLocal Self-attention-based Hybrid Multiple Instance Learning for Partial Spoof Speech Detection
Zhu, Y., Chen, Y., Zhao, Z., Liu, X., & Guo, J. (2023) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3616540 · Cited by 18
“First, MIL [19–22] is a weakly supervised learning algorithm that deals with bags consisting of multiple instances and only requires bags labels during training.”
BackgroundMorales-Álvarez, P., Schmidt, A., Hernández-Lobato, J. M., & Molina, R. (2023) · Pattern Recognition · doi:10.1016/j.patcog.2023.110057 · Cited by 17
“Although it has been applied in many different areas [1], we will focus on the case of computational pathology.”
MethodSenadeera, D. C., Yang, X., Kollias, D., & Slabaugh, G. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00493 · Cited by 17
“Models trained to detect anomalies in this manner might not adequately understand the context in which certain violent actions are taking place and therefore may not be able to generalize well as their main task is not to learn the context-specific features for violent events [3].”
BackgroundHezi, H., Gelber, M., Balabanov, A., Maruvka, Y. E., & Freiman, M. (2024) · Computer Methods and Programs in Biomedicine · doi:10.1016/j.cmpb.2024.108513 · Cited by 18
Safta, W., Farhangi, M. M., Veasey, B. P., Amini, A. A., & Frigui, H. (2019) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi.2019.8759289 · Cited by 17
“The smallest volume V1 was set to [9,9,7] while the largest volume V21 was set to [53,55,45] in each of the x,y and z directions with center at the nodule centroid.”
BackgroundMethodInfluentialChen, S., Zhong, X., Li, L., Liu, W., Gu, C., & Zhong, L. (2020) · Neural Processing Letters · doi:10.1007/s11063-020-10352-2 · Cited by 17
Multiple Instance Learning with Genetic Pooling for medical data analysis
Bhattacharjee, K., Pant, M., Zhang, Y., & Satapathy, S. C. (2020) · Pattern Recognition Letters · doi:10.1016/j.patrec.2020.02.025 · Cited by 17
Matten, P., Scherer, J., Schlegl, T., Nienhaus, J., Stino, H., Niederleithner, M., Schmidt-Erfurth, U. M., Leitgeb, R. A., Drexler, W., Pollreisz, A., et al. (2023) · Scientific Reports · doi:10.1038/s41598-023-35713-4 · Cited by 16
Li, Z., Wang, Z., & Liu, Q. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i2.25237 · Cited by 16
“Multiple instance learning (Carbonneau et al. 2018) is used to train the class-aware branch, where top-k pooling is applied over CAS for each class to make video-level class predictions.”
Weakly-supervised learning based automatic augmentation of aerial insulator images
Jiang, D., Cao, Y., & Yang, Q. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.122739 · Cited by 16
Multiple Instance Learning for Content Feedback Localization without Annotation
Hellman, S., Murray, W., Wiemerslage, A., Rosenstein, M., Foltz, P. W., Becker, L. A., & Derr, M. A. (2020) · Workshop on Innovative Use of NLP for Building Educational Applications · doi:10.18653/v1/2020.bea-1.3 · Cited by 16
“The description provided here is based on Carbonneau et al. (2016).”
BackgroundCost Effective Multi-label Active Learning via Querying Subexamples
Chen, X., Yu, G., Domeniconi, C., Wang, J., Li, Z., & Zhang, Z. (2018) · Proceedings · doi:10.1109/icdm.2018.00109 · Cited by 16
“In multi-instance learning, a bag is positive with respect to a particular label if at least one of its instances is positive, while the bag is negative if all its instances are negative for that label [18], [19].”
BackgroundInjecting Text Clues for Improving Anomalous Event Detection From Weakly Labeled Videos
Liu, T., Lam, K., & Bao, B. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3477351 · Cited by 15
“Under the supervision of video-level annotations, existing WS-VAD methods typically employ a multiple instance learning (MIL) [15], [16] framework to train a classifier, for generating a sequence of frame-level or snippet-level anomalous scores.”
MethodHistopathological image classification based on self-supervised vision transformer and weak labels
Gul, A. G., Çetin, Ö., Reich, C., Flinner, N., Prangemeier, T., & Koeppl, H. (2022) · SPIE Medical Imaging: Digital and Computational Pathology · doi:10.1117/12.2624609 · Cited by 16
“Class labels are only provided for the bags and instance labels are not known a priori [12].”
Empowering digital pathology applications through explainable knowledge extraction tools
Marchesin, S., Giachelle, F., Marini, N., Atzori, M., Boytcheva, S., Buttafuoco, G., Ciompi, F., Nunzio, G. M. D., Fraggetta, F., Irrera, O., et al. (2022) · Journal of Pathology Informatics · doi:10.1016/j.jpi.2022.100139 · Cited by 16
“In this regard, the pathological concepts extracted from diagnostic reports can serve as weak labels to train prediction models for image classi fi cation tasks, 22,23 or asnodes tobuildreport-levelknowledgegraphsfor informationretrieval tasks.”
BackgroundSong, H., Yang, H. L., & Jung, J. (2022) · IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing · doi:10.1109/jstars.2022.3230625 · Cited by 15
“For weakly supervised object detection (WSOD), many effective algorithms have been developed based on the multiple instance learning paradigm [37].”
MethodHu, H., Ye, R., Thiyagalingam, J., Coenen, F., & Su, J. (2023) · Applied Intelligence · doi:10.1007/s10489-023-04458-y · Cited by 16
“As most medical images have relatively high resolution and weakly labeled small datasets, the MIL method is a common method for medical image analysis [2].”
MethodExplaining Anomalous Events in Flight Data of UAV With Deep Attention-Based Multi-Instance Learning
Yang, J. R., Tang, D., Yu, J., Zhang, J., & Liu, H. (2023) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3301678 · Cited by 15
“This approach effectively identifies key instances to explain predictions and decision functions [16] and has been widely used in diverse areas, such as computer vision and document classification [17].”
BackgroundWu, R., Liu, H., Li, H., Chen, L., Wei, L., Huang, X., Liu, X., Men, X., Li, X., Han, L., et al. (2023) · BioMedical Engineering OnLine · doi:10.1186/s12938-023-01164-1 · Cited by 16
“Next, a multi-instance learning (MIL) network was applied to classify CSVD sequence-wise [35].”
MethodMIL-Mixer: A Robust Bag Encoding Strategy for Multiple Instance Learning (MIL) using MLP-Mixer
Waqas, M., Khan, Z., Ahmed, S. U., Tahir, M. A., & Raza, A. (2023) · International Conference on Emerging Technologies · doi:10.1109/icet59753.2023.10374927 · Cited by 15
Fu, Z., Chen, Q., Wang, M., & Huang, C. (2024) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2024.106970 · Cited by 18
A Stochastic View of Varying Styles in Art Paintings
Sargentis, G.-F., Dimitriadis, P., Iliopoulou, T., & Koutsoyiannis, D. (2021) · Heritage · doi:10.3390/heritage4010021 · Cited by 15
Vocaturo, E., Zumpano, E., Giallombardo, G., & Miglionico, G. (2020) · International Database Engineering and Applications Symposium (IDEAS) · doi:10.1145/3410566.3410611 · Cited by 15
Imbalance deep multi‐instance learning for predicting isoform–isoform interactions
Yu, G., Zeng, J., Wang, J., Zhang, H., Zhang, X., & Guo, M. (2021) · International Journal of Intelligent Systems · doi:10.1002/int.22402 · Cited by 15
“1,2,15,17 Particularly, a gene pair G 1 and G 2 is regarded as a “ bag, ” an isoform pair, respectively, spliced from the two genes is considered as an “ instance.”
BackgroundDraelos, R. L., Kesty, C. E., & Kesty, K. R. (2025) · Journal of Cosmetic Dermatology · doi:10.1111/jocd.70050 · Cited by 16
Qu, J., Wei, X., & Qian, X. (2023) · Medical Image Analysis · doi:10.1016/j.media.2023.102774 · Cited by 14
Multiple-Instance Metric Learning Network for Hyperspectral Target Detection
Yang, B., He, Y., Jiao, C., Pan, X., Wang, G., Wang, L., & Wu, J. (2023) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2023.3291439 · Cited by 14
“Many scholars considered choosing a suitable permutation-invariant MIL pooling function [19], [20] to locate and aggregate the prime prototypes in weakly labeled bags, e.g., mean pooling, max pooling, and log-sum-exp pooling.”
BackgroundWeakly supervised image segmentation beyond tight bounding box annotations
Wang, J., & Xia, B. (2023) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2023.107913 · Cited by 15
Yi, X., Chen, H., Wu, P., Wang, G., Mo, L., Wu, B., Yi, Y., Fu, X., & Qian, P. (2024) · Agronomy · doi:10.3390/agronomy14061285 · Cited by 14
Deep multiple instance learning on heterogeneous graph for drug–disease association prediction
Gu, Y., Zheng, S., Zhang, B., Kang, H., Jiang, R., & Li, J. (2024) · Computers in Biology and Medicine · doi:10.1016/j.compbiomed.2024.109403 · Cited by 14
Wibawa, M. S., Lo, K. W., Young, L. S., & Rajpoot, N. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-25082-8_43 · Cited by 13
Generating Bounding Box Supervision for Semantic Segmentation with Deep Learning
Bonechi, S., Andreini, P., Bianchini, M., & Scarselli, F. (2018) · Lecture notes in computer science · doi:10.1007/978-3-319-99978-4_15 · Cited by 14
Bleu-Laine, M.-H., Puranik, T. G., Mavris, D. N., & Matthews, B. L. (2021) · AIAA SciTech Forum · doi:10.2514/6.2021-0776 · Cited by 14
Steganographer detection via a similarity accumulation graph convolutional network
Zhang, Z., Zheng, M., Zhong, S.-H., & Liu, Y. (2021) · Neural Networks · doi:10.1016/j.neunet.2020.12.026 · Cited by 14
Weak Supervision: A Survey on Predictive Maintenance
Martínez‐Heredia, A. M., & Ventura, S. (2025) · Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery · doi:10.1002/widm.70022 · Cited by 16
ANEDL: Adaptive Negative Evidential Deep Learning for Open-Set Semi-supervised Learning
Yu, Y., Deng, D., Liu, F., Dou, Q., Jin, Y., Chen, G., & Heng, P. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i15.29597 · Cited by 13
Tang, X., Guo, R., Zhang, C., & Qian, X. (2024) · Medical Image Analysis · doi:10.1016/j.media.2024.103266 · Cited by 13
Deep Learning in Automating Breast Cancer Diagnosis from Microscopy Images
Gu, Q., Prodduturi, N., & Hart, S. N. (2024) · Design of Medical Devices Conference (DMD) · doi:10.1115/dmd2024-1017 · Cited by 15
ViCap-AD: video caption-based weakly supervised video anomaly detection
Lim, J., Lee, J., Kim, H., & Park, E. (2025) · Machine Vision and Applications · doi:10.1007/s00138-025-01676-x · Cited by 13
Lin, F., Zhang, Z., Wang, J., Liang, C., Xu, J., Zeng, X., Zeng, Q., Chen, H., Zhuang, J., Ma, Y., et al. (2025) · EClinicalMedicine · doi:10.1016/j.eclinm.2025.103166 · Cited by 13
“42 Nevertheless, these models have yet to provide clear indications for clinical decision-making.”
Weakly-Supervised Multi-view Multi-instance Multi-label Learning
Xing, Y., Yu, G., Wang, J., Domeniconi, C., & Zhang, X. (2020) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2020/432 · Cited by 13
“Multi-instance learning typically uses the bag-instance relations to predict the labels of bags; some approaches can also identify the labels of instances [Carbonneau et al., 2018].”
BackgroundWeakly-supervised temporal action localization: a survey
Baraka, A. M., & Noor, M. H. M. (2022) · Neural Computing and Applications · doi:10.1007/s00521-022-07102-x · Cited by 13
“In general, MIL can model several types of tasks: classification, regression, ranking, and clustering [19].”
BackgroundFuhrman, J. D., Yip, R., Zhu, Y., Jirapatnakul, A. C., Li, F., Henschke, C. I., Yankelevitz, D. F., & Giger, M. L. (2023) · Scientific Reports · doi:10.1038/s41598-023-27549-9 · Cited by 13
Deep joint learning of pathological region localization and Alzheimer’s disease diagnosis
Park, C., Jung, W., & Suk, H. (2023) · Scientific Reports · doi:10.1038/s41598-023-38240-4 · Cited by 13
SVM-Based Multiple Instance Classification via DC Optimization
Astorino, A., Fuduli, A., Giallombardo, G., & Miglionico, G. (2019) · Algorithms · doi:10.3390/a12120249 · Cited by 13
Saberian, M., Moriarty, K. P., Olmstead, A. D., Hallgrimson, C., Jean, F., Nabi, I. R., Libbrecht, M. W., & Hamarneh, G. (2022) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2022.3178523 · Cited by 14
“dependent on laborious human annotations [10].”
MethodFreyre, C. A. C., Spiegel, S., Keller, C. G., Vandemeulebroecke, M., Hoefling, H., Dubost, V., Çörek, E., Moulin, P., & Hossain, I. (2021) · Toxicologic Pathology · doi:10.1177/0192623320987202 · Cited by 12
“27-32 The present MIL models are trained over multiple examples of slides with different labels (eg, lesion, no lesion) to learn about differences between the tiles and also to combine them together into a slide-level prediction.”
BackgroundChristley, S., Ostmeyer, J., Quirk, L., Zhang, W., Sirak, B. A., Giuliano, A. R., Zhang, S., Monson, N. L., Tiro, J. A., Lucas, E., et al. (2021) · Frontiers in Immunology · doi:10.3389/fimmu.2021.624230 · Cited by 12
Zhou, Y., & Lu, Y. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9747121 · Cited by 12
Movies2Scenes: Using Movie Metadata to Learn Scene Representation
Chen, S., Liu, C.-H., Hao, X., Nie, X., Arap, M., & Hamid, R. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00632 · Cited by 16
“A useful way to interpret our movie similarity learning step is to view it from a multiple instance learning (MIL) perspective [5].”
BackgroundYao, H., Wu, Z., Gao, S. S., Guymer, R. H., Steffen, V., Chen, H., Hejrati, M., & Zhang, M. (2023) · Ophthalmology Science · doi:10.1016/j.xops.2023.100428 · Cited by 12
Photovoltaic Cell Defect Detection Based on Weakly Supervised Learning With Module-Level Annotations
Kang, H., Hong, J., Lee, J., & Kang, S. (2024) · IEEE Access · doi:10.1109/access.2024.3349975 · Cited by 12
“The multiple-instance learning approach [16], [17] uses a set of labeled bags, each containing many instances, as a training dataset to build a model for instance-level prediction.”
MethodWang, X., Ma, Y., Xu, Y., Huang, Q., Yang, Z., & Zhang, Z. (2025) · International Journal of Remote Sensing · doi:10.1080/01431161.2025.2459215 · Cited by 16
“Multiple instance learning (MIL) (Dietterich et al., 1997; Carbonneau et al., 2018) is a variant of a supervised machine learning method.”
Multiple-Instance Learning from Similar and Dissimilar Bags
Feng, L., Shu, S.-L., Cao, Y.-Z., Tao, L., Wei, H., Xiang, T., An, B., & Niu, G. (2021) · Knowledge Discovery and Data Mining · doi:10.1145/3447548.3467318 · Cited by 13
Multiple Instance Neuroimage Transformer
Singla, A., Zhao, Q., Daniel, K., Zhou, Y., Pohl, K. M., & Adeli, E. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-16919-9_4 · Cited by 13
“Inspired by some prior MIL deep learning models applied to medical images [7,24], we next develop a convolution-free transformer-based architecture inspired by the MIL paradigm, called MINiT.”
MethodWeakly-Supervised Temporal Action Detection for Fine-Grained Videos with Hierarchical Atomic Actions
Li, Z., He, L., & Xu, H. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-20080-9_33 · Cited by 12
On Detection of Diabetic Retinopathy via Multiple Instance Learning
Avolio, M., Fuduli, A., Vocaturo, E., & Zumpano, E. (2023) · International Database Engineering and Applications Symposium (IDEAS) · doi:10.1145/3589462.3589490 · Cited by 12
“For such problems, there are mainly three types of approaches (see for example [1, 8]).”
BackgroundDouble similarities weighted multi-instance learning kernel and its application
Zhang, J., Wu, Y., Fang, H., Liu, X., Li, M., Zhou, D., & Zheng, W. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.121900 · Cited by 12
Predicting Social Network Users with Depression from Simulated Temporal Data
Wongkoblap, A., Vadillo, M. A., & Ćurčin, V. (2019) · IEEE EUROCON - International Conference on Computer as a Tool (EUROCON) · doi:10.1109/eurocon.2019.8861514 · Cited by 13
“supervised learning techniques that need to know the labels of all instances [11].”
BackgroundIns-ATP: Deep estimation of ATP for organoid based on high throughput microscope images
Bian, X., Chen, S., & Liu, W. (2025) · Methods · doi:10.1016/j.ymeth.2025.01.012 · Cited by 11
Pérez-Cano, J., Wu, Y., Schmidt, A., López-Pérez, M., Morales-Álvarez, P., Molina, R., & Katsaggelos, A. K. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.122296 · Cited by 11
Majumdar, A., Lad, J., Tumanova, K., Serra, S., Quereshy, F. A., Khorasani, M., & Vitkin, I. A. (2023) · Journal of Biomedical Optics · doi:10.1117/1.jbo.29.5.052915 · Cited by 12
“…(measure of asymmetry) or kurtosis (measure of “ tailedness ” ) may provide additional insights at the ROI-level that can improve predictive power; 55 (3) multiple-instance learning — a weakly supervised learning algorithm that deals with data of similar form to this study (i.e., numerous data…”
ResultTrusted multi-scale classification framework for whole slide image
Ming, F., Xu, K., Wu, N., Huang, W., Bai, Y., Wang, Y., Wang, C., & Wang, H. (2023) · Biomedical Signal Processing and Control · doi:10.1016/j.bspc.2023.105790 · Cited by 13
Yeom, J. C., Kim, J., Kim, Y. J., Kim, K. G., & Kim, K. G. (2024) · Journal of Imaging Informatics in Medicine · doi:10.1007/s10278-024-01020-1 · Cited by 15
“Adnan et al. [11] divided large pathological images into patches, and multiple instance learning (MIL) [12] was applied to classify a bag of patches.”
Verma, R., Alban, T. J., Parthasarathy, P. B., Mokhtari, M., Toro, P., Cohen, M. L., Lathia, J. D., Ahluwalia, M. S., & Tiwari, P. (2024) · Science Advances · doi:10.1126/sciadv.adi0302 · Cited by 11
“In future, we will explore multiple-instance learning approaches ( 34 , 35 ), to address the challenges with patch selection, and further improve robustness of our prognostic models.”
BackgroundWeakly-Supervised Depression Detection in Speech Through Self-Learning Based Label Correction
Sun, Y., Zhou, Y., Xu, X., Qi, J., Xu, F., Ren, Z. X., & Schuller, B. W. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3533370 · Cited by 13
“In order to address the inexact supervision, Multi-Instance Learning (MIL)-based approaches treat each sample as a bag containing multiple instances [36], [37], employed in a variety of tasks (e. g., medical diagnosis and action localisation) [38], [39], [40].”
A heuristic approach for multiple instance learning by linear separation
Fuduli, A., Gaudioso, M., Khalaf, W., & Vocaturo, E. (2022) · Soft Computing · doi:10.1007/s00500-021-06713-1 · Cited by 11
Kalanadhabhatta, M., Santana, A. M., Zhang, Z., Ganesan, D., Grabell, A. S., & Rahman, T. (2022) · Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies · doi:10.1145/3534583 · Cited by 11
“Carbonneau et al. [28] provide a detailed review of MIL frameworks under different assumptions.”
BackgroundYu, F., Wang, X., Sali, R., & Li, R. (2023) · IEEE Journal of Biomedical and Health Informatics · doi:10.1109/jbhi.2023.3262454 · Cited by 12
“As the number of cancer cells varies among whole slide images, feature pooling is a typical tool in MIL [34] for generating holistic feature representations with a fixed di-mension.”
MethodA Domain Adaptation Multi-instance Learning for Diabetic Retinopathy Grading on Retinal Images
Song, R., Cao, P., Yang, J., Zhao, D., & Zai͏̈ane, O. R. (2020) · IEEE International Conference on Bioinformatics and Biomedicine (BIBM) · doi:10.1109/bibm49941.2020.9313398 · Cited by 11
“Specifically, a whole image that is uniformly split into many patches as instances, called bag, and a single class label is specified for all instances in a bag [5].”
BackgroundPose Consistency KKT-Loss for Weakly Supervised Learning of Robot-Terrain Interaction Model
Šalanský, V., Zimmermann, K., Petricek, T., & Svoboda, T. (2021) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2021.3076957 · Cited by 11
“One could achieve a similar effect by employing the MIL [22].”
MethodEnsemble-Based Instance Relevance Estimation in Multiple-Instance Learning
Waqas, M., Tahir, M. A., & Qureshi, R. (2021) · European Workshop on Visual Information Processing · doi:10.1109/euvip50544.2021.9483988 · Cited by 11
“Later, MIL-based approaches are also used in object tracking, web mining, and speech localization [8].”
BackgroundMethodHu, J., Yang, Y., Xu, Y., & Shen, H. (2022) · Bioinformatics · doi:10.1093/bioinformatics/btac634 · Cited by 11
RLogist: Fast Observation Strategy on Whole-Slide Images with Deep Reinforcement Learning
Zhao, B., Zhang, J., Ye, D., Cao, J., Han, X., Fu, Q., & Yang, W. P. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i3.25467 · Cited by 12
Yu, B., Nagae, G., Midorikawa, Y., Tatsuno, K., Dasgupta, B., Aburatani, H., & Ueda, H. (2024) · Briefings in Bioinformatics · doi:10.1093/bib/bbae529 · Cited by 10
“MIL is a machine learning technique used for classification tasks where the input data are represented as a collection of unlabeled “instances” [31, 32].”
MethodAction Selection Learning for Multi-label Multi-view Action Recognition
Nguyễn, T. T., Kawanishi, Y., Komamizu, T., & Ide, I. (2024) · ACM International Conference on Multimedia in Asia (MMAsia) · doi:10.1145/3696409.3700211 · Cited by 10
Cheng, H., Liu, X., Zhang, J., Dong, X., Ma, X., Zhang, Y., Meng, H., Chen, X., Yue, G., Li, Y., et al. (2025) · Expert Systems with Applications · doi:10.1016/j.eswa.2025.127425 · Cited by 10
Weakly-Supervised Log-Based Anomaly Detection with Inexact Labels via Multi-Instance Learning
He, M., Jia, T., Duan, C., Cai, H., Li, Y., & Huang, G. (2025) · ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) · doi:10.1109/icse55347.2025.00189 · Cited by 10
Xiang, X., & Wu, X. (2021) · Applied Sciences · doi:10.3390/app112110368 · Cited by 11
“Nevertheless, most existing MIC algorithms including attention-related methods treat instances in each bag as independent and identically distributed without considering the spatial topological relationships of instances corresponding to the original bag [14].”
Multiple instance learning for lung pathophysiological findings detection using CT scans
Frade, J., Pereira, T., Morgado, J., Silva, F., Freitas, C., Mendes, J., Negrão, E., de Lima, B. F., da Silva, M. C., Madureira, A. J., et al. (2022) · Medical & Biological Engineering & Computing · doi:10.1007/s11517-022-02526-y · Cited by 10
“After labeling the instances individually [24], the bag is classified on the underlying MIL assumption.”
BackgroundMethodDeep learning framework for early detection of COVID-19 using X-ray images
Khero, K., Usman, M., & Fong, A. C. M. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-15995-6 · Cited by 11
Machine Learning Approach for Frozen Tuna Freshness Inspection Using Low-Frequency A-Mode Ultrasound
Sakai, A., Yagi, M., Yasutomi, S., Mizuno, K., Suzuki, K., & Goto, K. (2023) · IEEE Access · doi:10.1109/access.2023.3319400 · Cited by 10
“This problem setting falls within the framework of multiple instance learning [83].”
BackgroundYang, Y., Zeng, N., Chen, Z., Li, W., Guo, Y., Wang, S., Duan, W., Liu, Y., Chen, R., & Kang, Y. (2023) · Journal of Healthcare Engineering · doi:10.1155/2023/3715603 · Cited by 10
“Furthermore, compared with experiment 1, multiple-instance learning (MIL) [34], a form of weakly supervised learning, is applied in experiment 2.”
Bayesian multiple instance regression for modeling immunogenic neoantigens
Park, S., Wang, X., Lim, J., Xiao, G., Lu, T., & Wang, T. (2020) · Statistical Methods in Medical Research · doi:10.1177/0962280220914321 · Cited by 10
“Because machine learning communities have addressed a similar type of tasks formulated as weakly labeled classification, there is abundant literature in the field of multiple instance classification (see a recent survey by Carbonneau et al. 1 and references therein).”
BackgroundYin, S. W., Peng, Q., Li, H., Zhang, Z., You, X., Fischer, K. M., Furth, S. L., Tasian, G. E., & Yong Fan, A. (2020) · IEEE International Symposium on Biomedical Imaging (ISBI) · doi:10.1109/isbi45749.2020.9098506 · Cited by 10
“By treating all 2D ultrasound images of the same subject as multiple instances of a bag, the MIL methods yield an overall bag-level classification score [7-9].”
MethodCastro-Macías, F. M., Morales-Álvarez, P., Wu, Y., Molina, R., & Katsaggelos, A. K. (2024) · Artificial Intelligence · doi:10.1016/j.artint.2024.104115 · Cited by 10
“The MIL approach has been successfully applied to many different scientific domains [1], being particularly well suited to medical imaging [3].”
BackgroundLuo, X., Wang, J., Tan, C., Dou, Q., Han, Z., Wang, Z., Tasnim, F., Wang, X., Zhan, Q., Li, X., et al. (2024) · Gastroenterology · doi:10.1053/j.gastro.2024.03.039 · Cited by 10
NMGrad: Advancing Histopathological Bladder Cancer Grading with Weakly Supervised Deep Learning
Fuster, S., Kiraz, U., Eftestøl, T. C., Janssen, E. A. M., & Engan, K. (2024) · Bioengineering · doi:10.3390/bioengineering11090909 · Cited by 9
“However, MIL methods can be susceptible to individual instances dominating the weighted aggregation of the WSI representation [27–29].”
BackgroundWeakly supervised localization model for plant disease based on Siamese networks
Chen, J., Guo, J., Zhang, H., Liang, Z., & Wang, S. (2024) · Frontiers in Plant Science · doi:10.3389/fpls.2024.1418201 · Cited by 9
Li, D., Wang, L., Hu, T., Zhang, L., & Lv, Q. (2021) · IEEE Transactions on Cognitive and Developmental Systems · doi:10.1109/tcds.2021.3135947 · Cited by 9
“As a weakly supervised method, multiple-instance learning (MIL) is a generalization of supervised classification in which the training class labels are associated with sets of bags instead of individual patterns [21].”
MethodWang, Z., Li, Z., Li, Z., Xu, Y., Qi, F., & Kong, J. (2023) · Computers & Security · doi:10.1016/j.cose.2023.103362 · Cited by 9
Deep learning-based gastric cancer diagnosis and clinical management
Xie, K., & Peng, J. (2023) · Journal of Radiation Research and Applied Sciences · doi:10.1016/j.jrras.2023.100602 · Cited by 10
Metadata
- Authors
- Carbonneau, M.-A., Cheplygina, V., Granger, E., & Gagnon, G.
- Affiliations
- École de Technologie Supérieure (CA)
- Erasmus MC (NL)
- Corresponding author
- Carbonneau, Marc-André
- Venue
- Pattern Recognition
- Publisher
- Elsevier BV
- Published
- 2018-05
- Volume
- 77
- Pages
- 329-353
- ISSN
- 0031-3203 (ISSN-L)
- OpenAlex
- W2560886373
- Cited by
- 740
- References
- 223
- Citations by year
- 2026: 75 · 2025: 118 · 2024: 91 · 2023: 124 · 2022: 98 · 2021: 92 · 2020: 84 · 2019: 41 · 2018: 14
- FWCI
- 22.10
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- Pattern Recognition (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- TU/e Research Portal (submittedVersion)
- EUR Research Repository (Erasmus University Rotterdam) (submittedVersion)
- Espace ÉTS (ETS) (submittedVersion)
- TU/e Research Portal (submittedVersion, other-oa)
- RePub (Erasmus University Rotterdam) (submittedVersion, other-oa)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Algorithms
- Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Keywords
- multiple instance learning, weakly supervised learning, problem characteristics, data distribution, label ambiguity, computer vision, document classification
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
