Labeled Training Samples Research Articles

Active Learning for Deep Visual Tracking.

Convolutional neural networks (CNNs) have been successfully applied to the single target tracking task in recent years. Generally, training a deep CNN model requires numerous labeled training samples, and the number and quality of these samples directly affect the representational capability of the trained model. However, this approach is restrictive in practice, because manually labeling such a large number of training samples is time-consuming and prohibitively expensive. In this article, we propose an active learning method for deep visual tracking, which selects and annotates the unlabeled samples to train the deep CNN model. Under the guidance of active learning, the tracker based on the trained deep CNN model can achieve competitive tracking performance while reducing the labeling cost. More specifically, to ensure the diversity of selected samples, we propose an active learning method based on multiframe collaboration to select those training samples that should be and need to be annotated. Meanwhile, considering the representativeness of these selected samples, we adopt a nearest-neighbor discrimination method based on the average nearest-neighbor distance to screen isolated samples and low-quality samples. Therefore, the training samples' subset selected based on our method requires only a given budget to maintain the diversity and representativeness of the entire sample set. Furthermore, we adopt a Tversky loss to improve the bounding box estimation of our tracker, which can ensure that the tracker achieves more accurate target states. Extensive experimental results confirm that our active-learning-based tracker (ALT) achieves competitive tracking accuracy and speed compared with state-of-the-art trackers on the seven most challenging evaluation benchmarks. Project website: https://sites.google.com/view/altrack/.

UIR-ES: An unsupervised underwater image restoration framework with equivariance and stein unbiased risk estimator

Underwater imaging faces challenges for enhancing object visibility and restoring true colors due to the absorptive and scattering characteristics of water. Underwater image restoration (UIR) seeks solutions to restore clean images from degraded ones, providing significant utility in downstream tasks. Recently, data-driven UIR has garnered much attention due to the potent expressive capabilities of deep neural networks (DNNs). These DNNs are supervised, relying on a large amount of labeled training samples. However, acquiring such data is expensive or even impossible in real-world underwater scenarios. While recent researches suggest that unsupervised learning is effective in UIR, none of these frameworks consider signal physical priors. In this work, we present a novel physics-inspired unsupervised UIR framework empowered by equivariance and unbiased estimation techniques. Specifically, equivariance stems from the invariance, inherent in natural signals to enhance data-efficient learning. Given that degraded images invariably contain noise, we propose a noise-tolerant loss for unsupervised UIR based on the Stein unbiased risk estimator to achieve an accurate estimation of the data consistency. Extensive experiments on the benchmark UIR datasets, including the UIEB and RUIE datasets, validate the superiority of the proposed method in terms of quantitative scores, visual outcomes, and generalization ability, compared to state-of-the-art counterparts. Moreover, our method demonstrates even comparable performance with the supervised model.

Identifying Mergers in the Legacy Surveys with Few-shot Learning

Galaxy mergers exert a pivotal influence on the evolutionary trajectory of galaxies and the expansive development of cosmic structures. The primary challenge encountered in machine learning–based identification of merging galaxies arises from the scarcity of meticulously labeled data sets specifically dedicated to merging galaxies. In this paper, we propose a novel framework utilizing few-shot learning techniques to identify galaxy mergers in the Legacy Surveys. Few-shot learning enables effective classification of merging galaxies even when confronted with limited labeled training samples. We employ a deep convolutional neural network architecture trained on data sets sampled from Galaxy Zoo Decals to learn essential features and generalize to new instances. Our experimental results demonstrate the efficacy of our approach, achieving high accuracy and precision in identifying galaxy mergers with few labeled training samples. Furthermore, we investigate the impact of various factors, such as the number of training samples and network architectures, on the performance of the few-shot learning model. The proposed methodology offers a promising avenue for automating the identification of galaxy mergers in large-scale surveys, facilitating the comprehensive study of galaxy evolution and structure formation. In pursuit of identifying galaxy mergers, our methodology is applied to analyze the Data Release 9 of the Dark Energy Spectroscopic Instrument Legacy Imaging Surveys. As a result, we have unveiled an extensive catalog encompassing 648,183 galaxy merger candidates. We publicly release the catalog alongside this paper.

Graph neural pre-training based drug-target affinity prediction.

Computational drug-target affinity prediction has the potential to accelerate drug discovery. Currently, pre-training models have achieved significant success in various fields due to their ability to train the model using vast amounts of unlabeled data. However, given the scarcity of drug-target interaction data, pre-training models can only be trained separately on drug and target data, resulting in features that are insufficient for drug-target affinity prediction. To address this issue, in this paper, we design a graph neural pre-training-based drug-target affinity prediction method (GNPDTA). This approach comprises three stages. In the first stage, two pre-training models are utilized to extract low-level features from drug atom graphs and target residue graphs, leveraging a large number of unlabeled training samples. In the second stage, two 2D convolutional neural networks are employed to combine the extracted drug atom features and target residue features into high-level representations of drugs and targets. Finally, in the third stage, a predictor is used to predict the drug-target affinity. This approach fully utilizes both unlabeled and labeled training samples, enhancing the effectiveness of pre-training models for drug-target affinity prediction. In our experiments, GNPDTA outperforms other deep learning methods, validating the efficacy of our approach.

Zero-shot visual grounding via coarse-to-fine representation learning

Visual grounding (VG) locates target objects in visual scenes by understanding given natural language queries. Current methods for VG mainly focus on grounding referring expressions or noun phrases covered in the labeled training samples. Despite their grounding prowess, these approaches struggle in grounding novel query–image pairs excluded from the training data. This shortage is usually caused by the deficiency of discriminative representation learning both in images and queries. To address these issues, we propose a one-stage coarse-to-fine framework for zero-shot VG to ground novel query-image samples. Specifically, in the coarse stage, we mine the global context information in the visual features and query embeddings by employing a multi-head self-attention block, strengthening the intra-modality relations in the visual and textual features. In the fine stage, we first learn the query-aware visual representations based on the acquired global context information via a multi-modal relation-enhanced transformer block, which explores the informative information by modeling the cross-modal interaction among visual and textual domains. We further excavate target-oriented discriminative representations from the acquired query-aware visual representations by a noun phrase-guided multi-modal interaction network, which augments the interaction between target-related phrases and the obtained query-aware visual representations to enhance the distinction of target regions, enhancing the subsequent referred target grounding and generalizing. In order to validate the proposed approach, we implement extensive experiments and ablation studies on public benchmark datasets, including RefCOCO, RefCOCO+, RefCOCOg, Flickr30K Entity, Flickr-Split-0 and Flickr-Split-1. Experimental results demonstrate that our approach substantially improves the grounding accuracy and achieves new state-of-the-art performance under single-stage training and testing.

Learning Unsupervised Cross-Domain Model for TIR Target Tracking

The limited availability of thermal infrared (TIR) training samples leads to suboptimal target representation by convolutional feature extraction networks, which adversely impacts the accuracy of TIR target tracking methods. To address this issue, we propose an unsupervised cross-domain model (UCDT) for TIR tracking. Our approach leverages labeled training samples from the RGB domain (source domain) to train a general feature extraction network. We then employ a cross-domain model to adapt this network for effective target feature extraction in the TIR domain (target domain). This cross-domain strategy addresses the challenge of limited TIR training samples effectively. Additionally, we utilize an unsupervised learning technique to generate pseudo-labels for unlabeled training samples in the source domain, which helps overcome the limitations imposed by the scarcity of annotated training data. Extensive experiments demonstrate that our UCDT tracking method outperforms existing tracking approaches on the PTB-TIR and LSOTB-TIR benchmarks.

A novel multi-layer discriminative dictionary learning approach for image classification

Discriminative dictionary learning (DDL) has been confirmed to be effective for image classification. However, existing DDL approaches often fail to extract deep hierarchical information due to the single-layer dictionary learning framework. Moreover, they overlook the atoms-label information in the dictionary, leading to reduced feature distinctiveness and lower classification accuracy. To overcome the above problem, a novel DDL method, called the Multi-layer local constraint and label embedding dictionary learning (M-LCLEDL), is proposed. Specifically, the novel multi-layer DDL framework, which is formed by stacking the DL process one by one, is designed to learn the deep hierarchical and nonlinear features. The layer-by-layer stacking of the DL process in the multi-layer DDL framework allows for the elimination of redundant and interfering features. This step-by-step elimination process enhances the stability and robustness of the framework. Additionally, to leverage the label information carried by the labeled training samples, atoms with label embedding and locality structure are introduced. The proposed approach includes a fast iteration strategy for efficient optimization. Experimental results demonstrate that the approach is relatively insensitive to dictionary size, achieving promising performance and greater stability compared to most DDL-based image classification approaches.

ADAC: an active domain adaptive network with progressive learning strategy for cloud detection of remote sensing imagery

ABSTRACT To enhance the adaptability and application capability of the cloud detection model in different remote sensing satellite domains, unsupervised domain adaptation methods are employed to improve the model’s robustness and transferability. However, conducting fully unlabeled training in the target domain to be adapted may impose difficulty on further improving the accuracy. It is more practical to obtain a few labeled target data for domain adaptation. Additionally, the detection of complex cloud scenes remains a persistent challenge. To address these issues, an Active Domain Adaptation network for Cloud detection (ADAC) is proposed in this paper. This network employs active learning to select a small number of representative samples for annotation, facilitating the transfer of a source domain-trained model to the target domain. Moreover, a novel Generator (G) designed for cloud segmentation is developed, where G utilizes a progressive learning strategy by dividing the network into two stages to learn gradient information, guiding cloud boundary segmentation and enabling the distinction between cloud-like objects and clouds. Furthermore, G integrates Transformer and Convolutional Neural Network to address the issues of misclassifying thin clouds. The proposed method is validated on Landsat 8 and GF-2 datasets, representing the source and target domains, respectively. The experimental results demonstrate ADAC achieved a 4% improvement in the overall accuracy with only 5% labeled training samples in the target domain, achieving the effective adaptation performance in the GF-2 domain.

Output Regularization With Cluster-Based Soft Targets.

While supervised learning of over-parameterized neural networks achieved state-of-the-art performance in image classification, it tends to over-fit the labeled training samples to give inferior generalization ability. Output regularization deals with over-fitting by using soft targets as additional training signals. Although clustering is one of the most fundamental data analysis tools for discovering general-purpose and data-driven structures, it has been ignored in existing output regularization approaches. In this article, we leverage this underlying structural information by proposing Cluster-based soft targets for Output Regularization (CluOReg). This approach provides a unified way for simultaneous clustering in embedding space and neural classifier training with cluster-based soft targets via output regularization. By explicitly calculating a class relationship matrix in the cluster space, we obtain classwise soft targets shared by all samples in each class. Results of image classification experiments under various settings on a number of benchmark datasets are provided. Without resorting to external models or designed data augmentation, we get consistent and significant reductions in classification error compared with other approaches, demonstrating that cluster-based soft targets effectively complement the ground-truth label.

Multi-scale task-aware structure graph modeling for few-shot image recognition

The Few-shot image recognition attempts to recognize images from a novel class with only a limited number of labeled training images, which is a few-shot learning (FSL) task. FSL is very challenging. Limited labeled training samples cannot adequately represent the distribution of classes, and the base and novel classes in the training and testing stages do not intersect and have different distributions, leading to a domain shift problem in generalizing the learned model to the novel class dataset. In this paper, we propose multi-scale task-aware structure graph modeling for few-shot image recognition. We train a meta-filter learner to generate task-aware local structure filters for each scale and adaptively capture the local structures at each scale. Moreover, we introduce a novel multi-scale graph attention network (MGAT) module to model the multi-scale local structures of an image, fully exploring the correlations between different local structures at all scales of the image. Finally, we leverage the attention mechanism of graph attention network to achieve information aggregation and propagation, aiming to obtain more representative and discriminative local structure features that integrate both local and global information. We conducted comprehensive experiments on four benchmark datasets widely adopted in FSL tasks. The experimental results demonstrate that the MTSGM obtains state-of-the-art performance, which validates the effectiveness of MTSGM.

Element-wise parallel deep learning for structural distributed damage diagnosis by leveraging physical properties of long-gauge static strain transmissibility under moving loads

The Transmissibility Function (TF) has gained considerable interest in structural damage detection because of its relatively high sensitivity to damage and robustness to excitation. This study proposed a distributed damage diagnosis method for beam-like structures based on a long-gauge static strain TF defined as the ratio of the Fourier transform of static strain response under moving loads from a target long-gauge sensor to that from the reference sensor. It has been discovered that each long-gauge static strain TF is independent of moving loads, making it an ideal damage indicator. It exhibits direct proportionality to the ratio of bending stiffness between the two measured zones covered by the target and reference long-gauge strain sensors while being unaffected by the stiffness of any other zones not covered by these two sensors. Moreover, the TF at zero frequency was proved to be equivalent to the ratio of static strain time history areas, relying solely on the stiffness of the covered zones. Leveraged by the physical properties of long-gauge static strain TF, an element-wise parallel deep neural network architecture was developed to decompose damage diagnosis into independent sub-tasks on the element level, utilizing a variational autoencoder (VAE) in the Bayesian inference framework for extracting features from the long-gauge strain TF and a regressor for mapping the features to elemental stiffness reduction. The divide-and-conquer strategy and element-wise parallel learning architecture allow for a significant reduction in the number of labeled training samples generated based on the updated numerical baseline model as it only requires simulated scenarios involving a single damage element. The efficiency and robustness of the proposed method were demonstrated through a numerical simply supported beam and a laboratory experiment on a two-span bridge.

DAAL-WS: A weakly-supervised method integrated with data augmentation and active learning strategies for MLS point cloud semantic segmentation

Mobile laser scanning (MLS) point clouds have increasingly been a significant data source for acquiring accurate three-dimensional (3D) semantic information from complex scenes. However, most current point cloud semantic segmentation methods heavily depend on a huge number of manually labeled training samples, which is labor-intensive and time-consuming. To address the above challenge, a weakly-supervised MLS point cloud semantic segmentation method integrated with data augmentation and active learning strategies is proposed (termed DAAL-WS). By taking advantage of our previously presented multi-branch weakly-supervised network (WSPointNet) in weakly supervisory signals and ensemble predictions, the DAAL-WS integrates WSPointNet with two essential components, i.e., an elevation-calibrated Mix3D (EC-Mix3D) data augmentation strategy and a point-level ensemble prediction-based (PLEP) active learning strategy. Specifically, the EC-Mix3D data augmentation strategy leverages elevation information to calibrate sub-point clouds and generates new contextual scenes by mixing the elevation-calibrated sub-point clouds, thereby augmenting the training point cloud distribution. Designed to reduce labeling redundancy, the PLEP active learning strategy selects the most important labeled points for model training. This strategy first measures the uncertainty for each unlabeled point by ensemble predictions and then employs a feature-distance suppression module to select the significant and discriminating unlabeled points for manual labeling. The proposed DAAL-WS method was evaluated on three public MLS datasets, including Toronto3D, Paris-Lille-3D, and WHU-MLS datasets, on which DAAL-WS obtained a competitive performance over fully-supervised baselines using only 0.015 %, 0.03 %, and 0.03 % labeled points, with mean Intersection over Union (mIoU) scores of 81.91 %, 82.59 %, and 60.36 %, respectively.

Self-supervised motor imagery EEG recognition model based on 1-D MTCNN-LSTM network

Objective. Aiming for the research on the brain–computer interface (BCI), it is crucial to design a MI-EEG recognition model, possessing a high classification accuracy and strong generalization ability, and not relying on a large number of labeled training samples. Approach. In this paper, we propose a self-supervised MI-EEG recognition method based on self-supervised learning with one-dimensional multi-task convolutional neural networks and long short-term memory (1-D MTCNN-LSTM). The model is divided into two stages: signal transform identification stage and pattern recognition stage. In the signal transform recognition phase, the signal transform dataset is recognized by the upstream 1-D MTCNN-LSTM network model. Subsequently, the backbone network from the signal transform identification phase is transferred to the pattern recognition phase. Then, it is fine-tuned using a trace amount of labeled data to finally obtain the motion recognition model. Main results. The upstream stage of this study achieves more than 95% recognition accuracy for EEG signal transforms, up to 100%. For MI-EEG pattern recognition, the model obtained recognition accuracies of 82.04% and 87.14% with F1 scores of 0.7856 and 0.839 on the datasets of BCIC-IV-2b and BCIC-IV-2a. Significance. The improved accuracy proves the superiority of the proposed method. It is prospected to be a method for accurate classification of MI-EEG in the BCI system.

DMAF-NET: Deep Multi-Scale Attention Fusion Network for Hyperspectral Image Classification with Limited Samples

In recent years, deep learning methods have achieved remarkable success in hyperspectral image classification (HSIC), and the utilization of convolutional neural networks (CNNs) has proven to be highly effective. However, there are still several critical issues that need to be addressed in the HSIC task, such as the lack of labeled training samples, which constrains the classification accuracy and generalization ability of CNNs. To address this problem, a deep multi-scale attention fusion network (DMAF-NET) is proposed in this paper. This network is based on multi-scale features and fully exploits the deep features of samples from multiple levels and different perspectives with an aim to enhance HSIC results using limited samples. The innovation of this article is mainly reflected in three aspects: Firstly, a novel baseline network for multi-scale feature extraction is designed with a pyramid structure and densely connected 3D octave convolutional network enabling the extraction of deep-level information from features at different granularities. Secondly, a multi-scale spatial–spectral attention module and a pyramidal multi-scale channel attention module are designed, respectively. This allows modeling of the comprehensive dependencies of coordinates and directions, local and global, in four dimensions. Finally, a multi-attention fusion module is designed to effectively combine feature mappings extracted from multiple branches. Extensive experiments on four popular datasets demonstrate that the proposed method can achieve high classification accuracy even with fewer labeled samples.

A sea–land clutter classification framework for over-the-horizon radar based on weighted loss semi-supervised generative adversarial network

Deep convolutional neural network has made great achievements in sea–land clutter classification for over-the-horizon radar (OTHR). The premise is that a large number of labeled training samples must be provided for a sea–land clutter classifier. In practical engineering applications, it is relatively easy to obtain label-free sea–land clutter samples. However, the labeling process is extremely cumbersome and requires expertise in the field of OTHR. To solve this problem, we propose an improved generative adversarial network, namely weighted loss semi-supervised generative adversarial network (WL-SSGAN). Specifically, we propose a joint feature matching loss by weighting the middle layer features of the discriminator of semi-supervised generative adversarial network (SSGAN). Furthermore, we propose the weighted loss of WL-SSGAN by linearly weighting the standard adversarial loss of SSGAN and the joint feature matching loss. The classification performance of WL-SSGAN is evaluated on sea–land clutter datasets. The experimental results show that WL-SSGAN can improve the performance of the fully supervised classifier with only a small number of labeled samples by utilizing a large number of unlabeled sea–land clutter samples. Further, the proposed weighted loss is superior to both the adversarial loss and the feature matching loss. Additionally, we compare WL-SSGAN with conventional semi-supervised classification methods and demonstrate that WL-SSGAN achieves the highest classification accuracy.

Matching Compound Prototypes for Few-Shot Action Recognition

The task of few-shot action recognition aims to recognize novel action classes using only a small number of labeled training samples. How to better describe the action in each video and how to compare the similarity between videos are two of the most critical factors in this task. Directly describing the video globally or by its individual frames cannot well represent the spatiotemporal dependencies within an action. On the other hand, naively matching the global representations of two videos is also not optimal since action can happen at different locations in a video with different speeds. In this work, we propose a novel approach that describes each video using multiple types of prototypes and then computes the video similarity with a particular matching strategy for each type of prototypes. To better model the spatiotemporal dependency, we describe the video by generating prototypes that model the multi-level spatiotemporal relations via transformers. There are a total of three types of prototypes. The first type of prototypes are trained to describe specific aspects of the action in the video e.g., the start of the action, regardless of its timestamp. These prototypes are directly matched one-to-one between two videos to compare their similarity. The second type of prototypes are the timestamp-centered prototypes that are trained to focus on specific timestamps of the video. To deal with the temporal variation of actions in a video, we apply bipartite matching to allow the matching of prototypes of different timestamps. The third type of prototypes are generated from the timestamp-centered prototypes, which regularize their temporal consistency while serving as an auxiliary summarization of the whole video. Experiments demonstrate that our proposed method achieves state-of-the-art results on multiple benchmarks.

Label noise correction for crowdsourcing using dynamic resampling

Crowdsourcing provides a cost-effective labeling solution for the acquisition of labeled training samples for machine learning by employing workers on the Internet. A common approach to improving the label quality is to employ a truth inference method to infer integrated labels for samples from their multiple noisy labels obtained from different crowd workers. Although the quality of integrated labels is significantly improved compared with that of the original noisy ones, it still cannot completely eliminate the noises inevitably existing in the integrated labels. To further improve the label quality, this paper proposes a novel label noise correction method for crowdsourcing based on dynamic resampling (DRNC). DRNC first divides the dataset with inferred labels into a clean set and a noisy set through a filter. According to a certain proportion, the clean set and the noisy set are resampled to train multiple heterogeneous classifiers, which form an ensemble classifier. Then, the dataset is divided by the ensemble classifier into a new sub-noisy set and a sub-clean set. The whole process repeats multiple rounds, generating multiple sub-clean sets. Finally, these sub-clean sets are used to train classifiers, which jointly correct the wrong labels in the dataset by voting. Experimental results on 25 simulated and 4 real-world datasets consistently show that the proposed DRNC averagely improves the quality of labels as well as the quality of learned models in the range of 1 to 10 percentage points, compared with four state-of-the-art crowdsourcing noise correction methods.

A hybrid convolution transformer for hyperspectral image classification

ABSTRACT Hyperspectral images play a crucial role in remote sensing applications surveillance, environment and precision agriculture, containing abundant object information. However, they often face challenges such as limited labelled data and imbalanced classes. In recent years, convolutional neural networks (CNNs) have shown impressive performance in computer vision tasks, including hyperspectral image classification. The emergence of transformers has also attracted attention for hyperspectral image analysis due to their promising capabilities. Nevertheless, transformers typically demand a substantial amount of training data, making their application challenging in scenarios with limited labelled samples. To overcome this limitation, we propose a hybrid convolution transformer framework. Our method uses a vision transformer and a residual 3D convolutional neural network model. It also uses a sequence aggregation layer to avoid overfitting issues that come up when there isn’t enough training data. Our proposed residual channel attention module captures richer spatial-spectral complementary information and maintains spectral details during the feature extraction process. We conducted experiments on three benchmark datasets. The proposed model achieved state of the art performance 99.75 % , 99.46 % and 99.95 % in terms of overall accuracy (OA) using only 5 % , 5 % and 10 % labelled training samples respectively. Which is better than other state of the art methods.

Research on Entity and Relationship Extraction with Small Training Samples for Cotton Pests and Diseases

The extraction of entities and relationships is a crucial task in the field of natural language processing (NLP). However, existing models for this task often rely heavily on a substantial amount of labeled data, which not only consumes time and labor but also hinders the development of downstream tasks. Therefore, with a focus on enhancing the model’s ability to learn from small samples, this paper proposes an entity and relationship extraction method based on the Universal Information Extraction (UIE) model. The core of the approach is the design of a specialized prompt template and schema on cotton pests and diseases as one of the main inputs to the UIE, which, under its guided fine-tuning, enables the model to subdivide the entity and relationship in the corpus. As a result, the UIE-base model achieves an accuracy of 86.5% with only 40 labeled training samples, which really solves the problem of the existing models that require a large amount of manually labeled training data for knowledge extraction. To verify the generalization ability of the model in this paper, experiments are designed to compare the model with four classical models, such as the Bert-BiLSTM-CRF. The experimental results show that the F1 value on the self-built cotton data set is 1.4% higher than that of the Bert-BiLSTM-CRF model, and the F1 value on the public data set is 2.5% higher than that of the Bert-BiLSTM-CRF model. Furthermore, experiments are designed to verify that the UIE-base model has the best small-sample learning performance when the number of samples is 40. This paper provides an effective method for small-sample knowledge extraction.

Multi-Modal Attentive Prompt Learning for Few-shot Emotion Recognition in Conversations

Emotion recognition in conversations (ERC) has emerged as an important research area in Natural Language Processing and Affective Computing, focusing on accurately identifying emotions within the conversational utterance. Conventional approaches typically rely on labeled training samples for fine-tuning pre-trained language models (PLMs) to enhance classification performance. However, the limited availability of labeled data in real-world scenarios poses a significant challenge, potentially resulting in diminished model performance. In response to this challenge, we present the Multi-modal Attentive Prompt (MAP) learning framework, tailored specifically for few-shot emotion recognition in conversations. The MAP framework consists of four integral modules: multi-modal feature extraction for the sequential embedding of text, visual, and acoustic inputs; a multi-modal prompt generation module that creates six manually-designed multi-modal prompts; an attention mechanism for prompt aggregation; and an emotion inference module for emotion prediction. To evaluate our proposed model’s efficacy, we conducted extensive experiments on two widely recognized benchmark datasets, MELD and IEMOCAP. Our results demonstrate that the MAP framework outperforms state-of-the-art ERC models, yielding notable improvements of 3.5% and 0.4% in micro F1 scores. These findings highlight the MAP learning framework’s ability to effectively address the challenge of limited labeled data in emotion recognition, offering a promising strategy for improving ERC model performance.