Multi-instance Multi-label Learning Research Articles

Advancing Automatic Gastritis Diagnosis: An Interpretable Multilabel Deep Learning Framework for the Simultaneous Assessment of Multiple Indicators.

The evaluation of morphological features such as inflammation, gastric atrophy, and intestinal metaplasia is crucial for diagnosing gastritis. However, artificial intelligence (AI) analysis for nontumor diseases like gastritis is limited. Previous deep learning models have omitted important morphological indicators and cannot simultaneously diagnose gastritis indicators or provide interpretable labels. To address this, an attention-based multi-instance multilabel learning network (AMMNet) was developed to simultaneously achieve the multilabel diagnosis of activity, atrophy, and intestinal metaplasia with only slide-level weak labels. To evaluate AMMNet's real-world performance, a diagnostic test was designed to observe improvements in junior pathologists' diagnostic accuracy and efficiency with and without AMMNet assistance. In this study of 1,096 patients from 7 independent medical centers, AMMNet performed well in assessing activity (area under the curve (AUC): 0.93), atrophy (AUC: 0.97), and intestinal metaplasia (AUC: 0.93). The false-negative rates (FNRs) of these indicators were only 0.04, 0.08, and 0.18, respectively, and junior pathologists had lower FNRs with model assistance (0.15 vs. 0.10). Furthermore, AMMNet reduced the time required per whole-slide image (WSI) from 5.46 minutes to only 2.85 minutes, enhancing diagnostic efficiency. In block-level clustering analysis, AMMNet effectively visualized task-related patches within WSIs, improving interpretability. These findings highlight AMMNet's effectiveness in accurately evaluating gastritis morphological indicators on multicenter datasets. Using multi-instance multilabel learning strategies to support routine diagnostic pathology deserves further evaluation.

Single-Stage Broad Multi-Instance Multi-Label Learning (BMIML) With Diverse Inter-Correlations and Its Application to Medical Image Classification

Single-Stage Broad Multi-Instance Multi-Label Learning (BMIML) With Diverse Inter-Correlations and Its Application to Medical Image Classification

Hierarchical multi-instance multi-label learning for Chinese patent text classification

To further enhance the accuracy of the Chinese patent classification, this paper proposes a model, based on the patent structure and takes the patent claim as subjects, with multi-instance multi-label learning as the main method. Firstly, the patent claims are divided into multiple independent texts using the sequence number as the splitting token. For each patent, multiple claims are regarded as multiple instances, and the corresponding IPCs serve as its multiple labels. Next, the concept of secondary_label is introduced following the composition rules of IPC, and the relationships between instances and multiple secondary_labels are mined through the construction of fully-connected layers. To capture more comprehensive semantic information of instances, BIGRU and self-attention are employed to enhance semantics and reduce information loss during the training process. Finally, the max-pooling operations are utilised to obtain the predicted categories of patents based on capturing the relationships between instances and different hierarchical labels. Experimental results on the '2017 Chinese patent dataset' demonstrate that the multi-instance multi-label approach can effectively mine deeper relationships between patents and labels in classification tasks. As a result, our model significantly improves the accuracy of patent text classification.

Fast Broad Multiview Multi-Instance Multilabel Learning (FBM3L) With Viewwise Intercorrelation.

Multiview multi-instance multilabel learning (M3L) is a popular research topic during the past few years in modeling complex real-world objects such as medical images and subtitled video. However, existing M3L methods suffer from relatively low accuracy and training efficiency for large datasets due to several issues: 1) the viewwise intercorrelation (i.e., the correlations of instances and/or bags between different views) are neglected; 2) the diverse correlations (e.g., viewwise intercorrelation, interinstance correlation, and interlabel correlation) are not jointly considered; and 3) high computation burden for training process over bags, instances, and labels across different views. To resolve these issues, a novel framework called fast broad M3L (FBM3L) is proposed with three innovations: 1) utilization of viewwise intercorrelation for better modeling of M3L tasks while existing M3L methods have not considered; 2) based on graph convolutional network (GCN) and broad learning system (BLS), a viewwise subnetwork is newly designed to achieve joint learning among the diverse correlations; and 3) under BLS platform, FBM3L can learn multiple subnetworks jointly across all views with significantly less training time. Experiments show that FBM3L is highly competitive (or even better than) in all evaluation metrics up to 64% in average precision (AP) and much faster than most M3L (or MIML) methods (up to 1030 times), especially on large multiview datasets ( ≥ 260 K objects).

A multi-instance multi-label learning algorithm based on radial basis functions and multi-objective particle swarm optimization

Radial basis function (RBF) neural networks for Multi-Instance Multi-Label (MIML) directly can exploit the connections between instances and labels so that they can preserve useful prior information, but they only adopt Gaussian radial basis function as their RBF whose parameters are difficult to determine. In this paper, parameters can be obtained by multi-objective optimization methods with multi performance measures treated as objectives, specifically, parameter estimation of different RBFs by an improved multi-objective particle swarm optimization (MOPSO) is proposed where Recall rate and Precision rate are chosen to obtain the most desirable Pareto optimal solution set. Furthermore, share-learning factor is proposed to modify the particle velocity in standard MOPSO to improve the global search ability and group cooperative ability. It is experimentally demonstrated that the proposed method can estimate the reliable parameters of different RBFs, and it is also very competitive with the state of art MIML methods.

Nearest neighbor-based approaches for multi-instance multi-label classification

Nearest neighbor-based methods are classic techniques that, due to their efficiency, still are widely used today. However, they have not been broadly applied to solve the multi-instance multi-label (MIML) problem, a supervised learning paradigm that combines multi-instance (MI) and multi-label (ML) learning. This work presents new neighbor-based approaches for solving MIML problems. On the one hand, MIML data are transformed into ML data and ML nearest neighbor algorithms are used. On the other hand, algorithms that directly address MIML data and use a bag-based distance are proposed.A comprehensive study and an overall comparison have been conducted to study the performance of these methods using different configurations. Experiments included 16 datasets and 8 performance metrics. The results and statistical tests showed that the problem transformation applied and the distance function used impacted the performance and that the approaches that do not transform the problem obtained the best predictive results. Furthermore, most of the proposed algorithms outperformed the MIMLkNN algorithm, the state-of-art algorithm for MIML learning based on nearest-neighbor. Therefore, the relevance and capabilities of neighbor-based approaches to obtain competitive results in MIML learning are shown. Finally, all the algorithms developed in this paper have been included in the MIML library to facilitate the comparison with other future proposals.

Meta Multi-Instance Multi-Label learning by heterogeneous network fusion

Multi-Instance Multi-Label Learning (MIML) models complex objects (bags), each of which is composed with a set of instances and associated with a set of labels. Current MIML solutions still focus on a single-type of bags and assume an independent and identically distribution (IID) of training data. But these bags are linked with objects of other types, which also encode the semantics of bags. In addition, they generally need abundant labeled data for training. To effectively mine MIML objects linked with objects of other types, we propose a heterogeneous network embedding and meta learning based approach (MetaMIML). MetaMIML introduces the context learner with network embedding to learn context representations of bags for structure information extraction, the task learner to extract the meta knowledge for fast adapting to new tasks with scarce training objects, and finally fuses the structural and attribute information to predict the labels of bags/instances. In this way, MetaMIML can not only naturally deal with MIML objects at data level improving, but also exploit the power of meta-learning at the model enhancing. Experiments on benchmark datasets demonstrate that MetaMIML achieves a significantly better performance than state-of-the-art algorithms.

MIML-GAN: A GAN-Based Algorithm for Multi-Instance Multi-Label Learning on Overlapping Signal Waveform Recognition

Existing studies for automatic waveform recognition of overlapping signals have mostly been conducted in a supervised manner. Although demonstrating superior performance in recent years, supervised methods rely heavily on sufficient labeled samples, rely but the acquisition of annotated data is expensive, time-consuming, and sometimes rely infeasible. This shortage drives the need for semi-supervised learning methods, where the unlabeled samples can be fully exploited in the training stage. In addition, multi-instance multi-label (MIML) learning is essentially another weakly supervised learning protocol, and precisely fits the form of the time-frequency images TFIs obtained from the transformation of overlapping signals. In this paper, delving into the MIML learning problem, we leverage the advantage of adversarial training to formulate an effective algorithm MIML-GAN, which is tailored to the MIML problem of overlapping signal waveform recognition. After feeding the TFIs into the network, MIML-GAN approximates the distribution of the training data using the adversarial learning principle. Subsequently, the baglevel prediction can be derived from the instance-level prediction upon the MIML discriminator through adaptive threshold calibration. Specifically, we elaborately studied the global optimality of the MIML-GAN objective function, and extensive simulations are carried out with overlapping signal dataset, validating the ascendancy of the proposed method. Comparative experiments demonstrate that the proposed algorithm possesses promising feature representation capability, and outperforms the existing semi-supervised and supervised signal waveform recognition approaches.

Predicting the function of rice proteins through Multi-instance Multi-label Learning based on multiple features fusion.

There are a large number of unannotated proteins with unknown functions in rice, which are difficult to be verified by biological experiments. Therefore, computational method is one of the mainstream methods for rice proteins function prediction. Two representative rice proteins, indica protein and japonica protein, are selected as the experimental dataset. In this paper, two feature extraction methods (the residue couple model method and the pseudo amino acid composition method) and the Principal Component Analysis method are combined to design protein descriptive features. Moreover, based on the state-of-the-art MIML algorithm EnMIMLNN, a novel MIML learning framework MK-EnMIMLNN is proposed. And the MK-EnMIMLNN algorithm is designed by learning multiple kernel fusion function neural network. The experimental results show that the hybrid feature extraction method is better than the single feature extraction method. More importantly, the MK-EnMIMLNN algorithm is superior to most classic MIML learning algorithms, which proves the effectiveness of the MK-EnMIMLNN algorithm in rice proteins function prediction.

W44. OXTR DNA METHYLATION DIFFERENTIATES MEN ON THE OBESITY SPECTRUM WITH AND WITHOUT AN EATING DISORDER

The facts show that multi-instance multi-label (MIML) learning plays a pivotal role in Artificial Intelligence studies. Evidently, the MIML learning introduces a framework in which data is described by a bag of instances associated with a set of labels. In this framework, the modeling of the connection is the challenging problem for MIML. The RBF neural network can explain the complex relations between the instances and labels in the MIMLRBF. The parameters estimation of the RBF network is a difficult task. In this paper, the computational convergence and the modeling accuracy of the RBF network has been improved. The present study aimed to investigate the impact of a novel hybrid algorithm consisting of Gases Brownian Motion optimization (GBMO) algorithm and the gradient based fast converging parameter estimation method on multi-instance multi-label learning. In the current study, a hybrid algorithm was developed to estimate the RBF neural network parameters (the weights, widths and centers of the hidden units) simultaneously. The algorithm uses the robustness of the GBMO to search the parameter space and the efficiency of the gradient. For this purpose, two real-world MIML tasks and a Corel dataset were utilized within a two-step experimental design. In the first step, the GBMO algorithm was used to determine the widths and centers of the network nodes. In the second step, for each molecule with fixed inputs and number of hidden nodes, the parameters were optimized by a structured nonlinear parameter optimization method (SNPOM). The findings demonstrated the superior performance of the hybrid algorithmic method. Additionally, the results for training and testing the dataset revealed that the hybrid method enhances RBF network learning more efficiently in comparison with other conventional RBF approaches. The results obtain better modeling accuracy than some other algorithms.

Towards Discovering What Patterns Trigger What Labels

In many real applications, especially those involving data objects with complicated semantics, it is generally desirable to discover the relation between patterns in the input space and labels corresponding to different semantics in the output space. This task becomes feasible with MIML (Multi-Instance Multi-Label learning), a recently developed learning framework, where each data object is represented by multiple instances and is allowed to be associated with multiple labels simultaneously. In this paper, we propose KISAR, an MIML algorithm that is able to discover what instances trigger what labels. By considering the fact that highly relevant labels usually share some patterns, we develop a convex optimization formulation and provide an alternating optimization solution. Experiments show that KISAR is able to discover reasonable relations between input patterns and output labels, and achieves performances that are highly competitive with many state-of-the-art MIML algorithms.

A Multi-instance Multi-label Dual Learning Approach for Video Captioning

Video captioning is a challenging task in the field of multimedia processing, which aims to generate informative natural language descriptions/captions to describe video contents. Previous video captioning approaches mainly focused on capturing visual information in videos using an encoder-decoder structure to generate video captions. Recently, a new encoder-decoder-reconstructor structure was proposed for video captioning, which captured the information in both videos and captions. Based on this, this article proposes a novel multi-instance multi-label dual learning approach (MIMLDL) to generate video captions based on the encoder-decoder-reconstructor structure. Specifically, MIMLDL contains two modules: caption generation and video reconstruction modules. The caption generation module utilizes a lexical fully convolutional neural network (Lexical FCN) with a weakly supervised multi-instance multi-label learning mechanism to learn a translatable mapping between video regions and lexical labels to generate video captions. Then the video reconstruction module synthesizes visual sequences to reproduce raw videos using the outputs of the caption generation module. A dual learning mechanism fine-tunes the two modules according to the gap between the raw and the reproduced videos. Thus, our approach can minimize the semantic gap between raw videos and the generated captions by minimizing the differences between the reproduced and the raw visual sequences. Experimental results on a benchmark dataset demonstrate that MIMLDL can improve the accuracy of video captioning.

Multiview Multi-Instance Multilabel Active Learning.

Multiview multi-instance multilabel learning (M3L) is a framework for modeling complex objects. In this framework, each object (or bag) contains one or more instances, is represented with different feature views, and simultaneously annotated with a set of nonexclusive semantic labels. Given the multiplicity of the studied objects, traditional M3L methods generally demand a large number of labeled bags to train a predictive model to annotate bags (or instances) with semantic labels. However, annotating sufficient bags is very expensive and often impractical. In this article, we present an active learning-based M3L approach (M3AL) to reduce the labeling costs of bags and to improve the performance as much as possible. M3AL first adapts the multiview self-representation learning to evacuate the shared and individual information of bags and to learn the shared/individual similarities between bags across/within views. Next, to avoid scrutinizing all the possible labels, M3AL introduces a new query strategy that leverages the shared and individual information, and the diverse instance distribution of bags across views, to select the most informative bag-label pair for the query. Experimental studies on benchmark data sets show that M3AL can significantly reduce the query costs while achieving a better performance than other related competitive methods at the same cost.

ESRE: handling repeated entities in distant supervised relation extraction

Distant supervised relation extraction has been widely used to find novel relational facts from unstructured text. As far as we know, nearly all existing relation extraction models assume that each sentence contains precisely one entity pair, i.e., two entities. However, in reality, the datasets constructed by distant supervision have lots of sentences which contain repeated entities. In other words, there may be more than two entities in a sentence. This phenomenon breaks the assumption of existing models and makes them inevitably encounter the attention bias problem; that is, the model focuses on the wrong entities during relation extraction. To alleviate this problem, in this paper, we utilize the idea of ensemble learning and propose a novel distant supervised relation extraction model. The proposed model follows the multi-instance multi-label learning mechanism and conducts relation extraction based on the sentence-bag representations. Specifically, it first tries to identify the most critical entity and keywords, and then it uses voting mechanism to determine the sentence-level and bag-level features. Experimental results show that our proposed model outperforms the state-of-the-art baselines in relation extraction on a popular benchmark dataset, and it also indicates that the proposed model can indeed alleviate the problem caused by repeated-entity phenomenon.

Partial Multi-Label Learning With Noisy Label Identification.

Partial multi-label learning (PML) deals with problems where each instance is assigned with a candidate label set, which contains multiple relevant labels and some noisy labels. Recent studies usually solve PML problems with the disambiguation strategy, which recovers ground-truth labels from the candidate label set by simply assuming that the noisy labels are generated randomly. In real applications, however, noisy labels are usually caused by some ambiguous contents of the example. Based on this observation, we propose a partial multi-label learning approach to simultaneously recover the ground-truth information and identify the noisy labels. The two objectives are formalized in a unified framework with trace norm and l1 norm regularizers. Under the supervision of the observed noise-corrupted label matrix, the multi-label classifier and noisy label identifier are jointly optimized by incorporating the label correlation exploitation and feature-induced noise model. Furthermore, by mapping each bag to a feature vector, we extend PML-NI method into multi-instance multi-label learning by identifying noisy labels based on ambiguous instances. A theoretical analysis of generalization bound and extensive experiments on multiple data sets from various real-world tasks demonstrate the effectiveness of the proposed approach.

Smart Handheld Based Human Activity Recognition Using Multiple Instance Multiple Label Learning

Human activity recognition (HAR) and monitoring is beneficial for many medical applications, such as eldercare and post-trauma rehabilitation after surgery. HAR models based on smartphone’s accelerometer data could provide a convenient and ubiquitous solution to this problem. However, such models are mostly concerned with identifying basic activities such as ‘stand’/‘walk’ and thus the high-level context such as ‘walk in a queue’ for which a set of specific activities is performed remain unnoticed. Consequently, in this paper, we design a HAR framework that can identify a group of activities (rather than a single basic activity) being performed in a time window, thus, enables us to extract more meaningful information about the subject’s overall context. An algorithm is designed to formulate HAR as a multi-instance multi-label (MIML) learning problem. The procedure of generating feature bags of consecutive activity traces having multiple labels is formulated. In this work, the temporal relationship among activities is exploited to obtain a more comprehensive HAR model. Interestingly, the framework is found to completely/partially identify activity sequences that may not even be present in the training dataset. The framework is implemented and found to be working adequately when tested with real dataset collected from 8 users for 12 different activity combinations. MIML-kNN is found to provide maximum average precision (around 90%) even for an unseen test data-set.

Saliency based multiple object cosegmentation by ensemble MIML learning

As an interesting and emerging topic, multiple foreground cosegmentation (MFC) aims at extracting a finite number of common objects from an image collection, which is useful to variety of visual media applications. Although a number of approaches have been proposed to address this problem, many of them are designed with the misleading consistent information, suboptimal image representation, or inefficient segmentation assist and thus still suffer from certain limitations, which reduces their capability in the real-world scenarios. To alleviate these limitations, we propose a novel unsupervised MFC framework, which is composed of three components: unsupervised label generation, saliency based pseudo-annotation and cosegmentation by MIML learning. Specifically, we combine the high-level and low-level feature to represent the proposal objects, and adopt a novel SPAP clustering scheme to obtain more accurate consistent information of common objects. Then the saliency based pseudo-annotation help us reformulate the MFC problem as a Multi-Instance Multi-Label (MIML) learning problem by label propagation. Finally, by introducing a novel ensemble MIML learning scheme, the consistent information of common objects can more efficiently assist the segmentation of the images and get the more accurate segmentation results. We evaluate our framework on widely used public databases including the ICoseg dataset, MSRC dataset and FlickrMFC dataset for single and multiple common object cosegmentation respectively. Comparison results show that the proposed methods reach advanced and efficient performance.

Multi-Instance Multi-Label Learning for Gene Mutation Prediction in Hepatocellular Carcinoma.

Gene mutation prediction in hepatocellular carcinoma (HCC) is of great diagnostic and prognostic value for personalized treatments and precision medicine. In this paper, we tackle this problem with multi-instance multi-label learning to address the difficulties on label correlations, label representations, etc. Furthermore, an effective oversampling strategy is applied for data imbalance. Experimental results have shown the superiority of the proposed approach.

Query-Driven Multi-Instance Learning

We introduce a query-driven approach (qMIL) to multi-instance learning where the queries aim to uncover the class labels embodied in a given bag of instances. Specifically, it solves a multi-instance multi-label learning (MIML) problem with a more challenging setting than the conventional one. Each MIML bag in our formulation is annotated only with a binary label indicating whether the bag contains the instance of a certain class and the query is specified by the word2vec of a class label/name. To learn a deep-net model for qMIL, we construct a network component that achieves a generalized compatibility measure for query-visual co-embedding and yields proper instance attentions to the given query. The bag representation is then formed as the attention-weighted sum of the instances' weights, and passed to the classification layer at the end of the network. In addition, the qMIL formulation is flexible for extending the network to classify unseen class labels, leading to a new technique to solve the zero-shot MIML task through an iterative querying process. Experimental results on action classification over video clips and three MIML datasets from MNIST, CIFAR10 and Scene are provided to demonstrate the effectiveness of our method.

Multi-Instance Multi-Label Action Recognition and Localization Based on Spatio-Temporal Pre-Trimming for Untrimmed Videos

Weakly supervised action recognition and localization for untrimmed videos is a challenging problem with extensive applications. The overwhelming irrelevant background contents in untrimmed videos severely hamper effective identification of actions of interest. In this paper, we propose a novel multi-instance multi-label modeling network based on spatio-temporal pre-trimming to recognize actions and locate corresponding frames in untrimmed videos. Motivated by the fact that person is the key factor in a human action, we spatially and temporally segment each untrimmed video into person-centric clips with pose estimation and tracking techniques. Given the bag-of-instances structure associated with video-level labels, action recognition is naturally formulated as a multi-instance multi-label learning problem. The network is optimized iteratively with selective coarse-to-fine pre-trimming based on instance-label activation. After convergence, temporal localization is further achieved with local-global temporal class activation map. Extensive experiments are conducted on two benchmark datasets, i.e. THUMOS14 and ActivityNet1.3, and experimental results clearly corroborate the efficacy of our method when compared with the state-of-the-arts.