Multi-instance multi-label classification (MIML) is a fundamental task in machine learning, where each data sample comprises a bag containing several instances and multiple binary labels. Despite its wide applications, the data collection process involves matching multiple instances and labels, typically resulting in high annotation costs. In this paper, we study a novel yet practical crowdsourced multi-instance multi-label classification (CMIML) setup, where labels are collected from multiple crowd sources. To address this problem, we first propose a novel data generation process for CMIML, i.e., cross-label transition, where cross-label annotation error is more likely to appear rather than previous single-label transition assumption, due to the inherent similarity of localized instances from different classes. Then, we formally define the cross-label transition by cross-label transition matrices which are dependent across classes. Subsequently, we establish the first unbiased risk estimator for CMIML and further improve it through aggregation techniques, along with a rigorous generalization error bound. We also provide a practical implementation of cross-label transition matrix estimation. Comprehensive experiments on six benchmark datasets under various scenarios demonstrate that our algorithm outperforms the baselines by a large margin, validating its effectiveness in handling the CMIML problem.

Mingxuan Xia, Haobo Wang, Yixuan Li, Zewei Yu, Jindong Wang, Junbo Zhao, Runze Wu. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

Hierarchical text classification aims to classify documents into multiple labels within a hierarchical taxonomy, making it an essential yet challenging task in natural language processing.Recently, using Large Language Models (LLM) to tackle hierarchical text classification in a zero-shot manner has attracted increasing attention due to their cost-efficiency and flexibility.Given the challenges of understanding the hierarchy, various HTC prompting strategies have been explored to elicit the best performance from LLMs.However, our empirical study reveals that LLMs are highly sensitive to these prompting strategies-(i) within a task, different strategies yield substantially different results, and (ii) across various tasks, the relative effectiveness of a given strategy varies significantly.To address this, we propose a novel ensemble method, HiEPS, which integrates the results of diverse prompting strategies to promote LLMs' reliability.We also introduce a path-valid voting mechanism for ensembling, which selects a valid result with the highest path frequency score.Extensive experiments on three benchmark datasets show that HiEPS boosts the performance of single prompting strategies and achieves SOTA results.The source code is available at https: //github.com/MingxuanXia/HiEPS.

Traditional Incremental Learning (IL) targets to handle sequential fully-supervised learning problems where novel classes emerge from time to time. However, due to inherent annotation uncertainty and ambiguity, collecting high-quality annotated data in a dynamic learning system can be extremely expensive. To mitigate this problem, we propose a novel weakly-supervised learning paradigm called Incremental Partial Label Learning (IPLL), where the sequentially arrived data relate to a set of candidate labels rather than the ground truth. Technically, we develop the Prototype-Guided Disambiguation and Replay Algorithm (PGDR) which leverages the class prototypes as a proxy to mitigate two intertwined challenges in IPLL, i.e., label ambiguity and catastrophic forgetting. To handle the former, PGDR encapsulates a momentum-based pseudo-labeling algorithm along with prototype-guided initialization, resulting in a balanced perception of classes. To alleviate forgetting, we develop a memory replay technique that collects well-disambiguated samples while maintaining representativeness and diversity. By jointly distilling knowledge from curated memory data, our framework exhibits a great disambiguation ability for samples of new tasks and achieves less forgetting of knowledge. Extensive experiments demonstrate that PGDR achieves superior performance over the baselines in the IPLL task.

Black-box domain adaptation (BDA) targets to learn a classifier on an unsupervised target domain while assuming only access to black-box predictors trained from unseen source data. Although a few BDA approaches have demonstrated promise by manipulating the transferred labels, they largely overlook the rich underlying structure in the target domain. To address this problem, we introduce a novel separation and alignment framework for BDA. Firstly, we locate those well-adapted samples via loss ranking and a flexible confidence-thresholding procedure. Then, we introduce a novel graph contrastive learning objective that aligns under-adapted samples to their local neighbors and well-adapted samples. Lastly, the adaptation is finally achieved by a nearest-centroid-augmented objective that exploits the clustering effect in the feature space. Extensive experiments demonstrate that our proposed method outperforms best baselines on benchmark datasets, e.g. improving the averaged per-class accuracy by 4.1% on the VisDA dataset. The source code is available at: https://github.com/MingxuanXia/SEAL.

This work studies the novel Crowdsourced Multi-Label Learning (CMLL) problem, where each instance is related to multiple true labels but the model only receives unreliable labels from different annotators. Although a few Crowdsourced Multi-Label Inference (CMLI) methods have been developed, they require both the training and testing sets to be assigned crowdsourced labels and focus on true label inferring rather than prediction, making them less practical. In this paper, by excavating the generation process of crowdsourced labels, we establish the first unbiased risk estimator for CMLL based on the crowdsourced transition matrices. To facilitate transition matrix estimation, we upgrade our unbiased risk estimator by aggregating crowdsourced labels and transition matrices from all annotators while guaranteeing its theoretical characteristics. Integrating with the unbiased risk estimator, we further propose a decoupled autoencoder framework to exploit label correlations and boost performance. We also provide a generalization error bound to ensure the convergence of the empirical risk estimator. Experiments on various CMLL scenarios demonstrate the effectiveness of our proposed method. The source code is available at https://github.com/MingxuanXia/CLEAR.

Couriers’ package pick-up route prediction is a fundamental task in the emerging intelligent logistics systems. It is beneficial for order dispatching and arrival-time estimation by leveraging the predicted routes to improve those downstream tasks. However, the package pick-up route prediction problem is challenging since couriers’ behaviors are affected by both strict Spatial–Temporal Constraints (STC) and personalized spatial-transfer patterns (STP). Specifically, couriers have to consider explicitly spatial-temporal requirements such as the locations of packages and the promised pick-up time when selecting future routes. In addition, couriers have personalized mobility patterns between different locations, which are implicit patterns hidden behind couriers’ historical trajectories and cannot be ignored to precisely depict their behaviors. This paper proposes a novel framework, named CP-Route, to predict a specific courier’s future package pick-up route under strict spatial-temporal constraints, and enhance the prediction performance by learning the spatial-transfer patterns (i.e., the routing patterns) of couriers. A sophisticated encoder is designed to capture the STC and STP, and a mixed-distribution-based decoder is designed to simultaneously consider the influence of spatial-temporal constraints and routing patterns on couriers’ final decisions. Extensive experiments conducted on an industry-scale logistics dataset demonstrate the superiority of our proposed framework against the existing baseline methods. The online A/B test shows our contribution to the improvement of arrival time prediction.

Partial-label learning (PLL) is a peculiar weakly-supervised learning task where the training samples are generally associated with a set of candidate labels instead of single ground truth. While a variety of label disambiguation methods have been proposed in this domain, they normally assume a class-balanced scenario that may not hold in many real-world applications. Empirically, we observe degenerated performance of the prior methods when facing the combinatorial challenge from the long-tailed distribution and partial-labeling. In this work, we first identify the major reasons that the prior work failed. We subsequently propose SoLar, a novel Optimal Transport-based framework that allows to refine the disambiguated labels towards matching the marginal class prior distribution. SoLar additionally incorporates a new and systematic mechanism for estimating the long-tailed class prior distribution under the PLL setup. Through extensive experiments, SoLar exhibits substantially superior results on standardized benchmarks compared to the previous state-of-the-art PLL methods. Code and data are available at: https://github.com/hbzju/SoLar.

Diagnosing the fault as early as possible is significant to guarantee the safety and reliability of the high-speed train. Incipient fault always makes the monitored signals deviate from their normal values, which may lead to serious consequences gradually. Due to the obscure early stage symptoms, incipient faults are difficult to detect. This article develops a stacked generalization (stacking)-based incipient fault diagnosis scheme for the traction system of high-speed trains. To extract the fault feature from the faulty data signals, which are similar to the normal ones, the extreme gradient boosting (XGBoost), random forest (RF), extra trees (ET), and light gradient boosting machine (LightGBM) are chosen as the base estimators in the first layer of the stacking. Then, the logistic regression (LR) is taken as the meta estimator in the second layer to integrate the results from the base estimators for fault classification. Thanks to the generalization ability of stacking, the incipient fault diagnosis performance of the proposed stacking-based method is better than that of the single model (XGBoost, RF, ET, and LightGBM), although they can be used to detect the incipient faults, separately. Moreover, to find out the optimal hyperparameters of the base estimators, a swarm intelligent optimization algorithm, pigeon-inspired optimization (PIO), is employed. The proposed method is tested on a semiphysical platform of the CRH2 traction system in CRRC Zhuzhou Locomotive Company Ltd. The results show that the fault diagnosis rate of the proposed scheme is over 96%.