Text classification is a crucial and fundamental task in web content mining. Compared with the previous learning paradigm of pre-training and fine-tuning by cross entropy loss, the recently proposed supervised contrastive learning approach has received tremendous attention due to its powerful feature learning capability and robustness. Although several studies have incorporated this technique for text classification, some limitations remain. First, many text datasets are imbalanced, and the learning mechanism of supervised contrastive learning is sensitive to data imbalance, which may harm the model's performance. Moreover, these models leverage separate classification branches with cross entropy and supervised contrastive learning branches without explicit mutual guidance. To this end, we propose a novel model named SharpReCL for imbalanced text classification tasks. First, we obtain the prototype vector of each class in the balanced classification branch to act as a representation of each class. Then, by further explicitly leveraging the prototype vectors, we construct a proper and sufficient target sample set with the same size for each class to perform the supervised contrastive learning procedure. The empirical results show the effectiveness of our model, which even outperforms popular large language models across several datasets. Our code is available here.

The advancement of Internet technology has spurred a rise in the dissemination of misinformation, which has had profoundly negative impacts across a wide array of fields. To address this issue, the field of Misinformation Detection (MD), which focuses on the automated identification of online misinformation, has gained significant traction among researchers. In our study, we introduce an innovative plugand- play augmentation technique for MD, termed DEtecting Misinformation by Uncovering Commonsense Conflict (DEMUC). Our approach is grounded in previous psychological research that suggests that fake content often contains commonsense. Accordingly, we develop commonsense expressions for articles to highlight potential conflicts between the inferred commonsense triplets and the established ones derived from reliable commonsense reasoning tools. According to the used tools, we induce two variants DEMUC-KLM using the knowledge language model COMET and DEMUC-LLM using the large language models. These generated expressions are then applied as augmentations to each article, enabling any MD method to be trained on these augmented datasets. Additionally, we have manually compiled a new dataset CoMis, which consists exclusively of fake articles characterized by commonsense conflicts. By integrating DEMUC with various existing MD frameworks and evaluating them on four public benchmark datasets and CoMis, our empirical findings show that both DEMUC-KLM and DEMUC-LLM consistently and significantly outperform current MD baselines, while also generating precise commonsense expressions.

The proliferation of misinformation across diverse social media platforms has drawn significant attention from both academic and industrial communities due to its detrimental effects. Accordingly, automatically distinguishing misinformation, dubbed as Misinformation Detection (MD), has become an increasingly active research topic. The mainstream methods formulate MD as a static learning paradigm, which learns the mapping between the content, links, and propagation of news articles and the corresponding manual veracity labels. However, the static assumption is often violated, since in real-world scenarios, the veracity of news articles may vacillate within the dynamically evolving social environment. To tackle this problem, we propose a novel framework, namely Misinformation detection with Dynamic Environmental Representations (MISDER). The basic idea of MISDER lies in learning a social environmental representation for each period and employing a temporal model to predict the representation for future periods. In this work, we specify the temporal model as the LSTM model, continuous dynamics equation, and pre-trained dynamics system, suggesting three variants of MISDER, namely MISDER-LSTM, MISDER-ODE, and MISDER-PT, respectively. To evaluate the performance of MISDER, we compare it to various MD baselines across 2 prevalent datasets, and the experimental results can indicate the effectiveness of our proposed model.

Graph few-shot learning has garnered significant attention for its ability to rapidly adapt to downstream tasks with limited labeled data, sparking considerable interest among researchers. Recent advancements in graph few-shot learning models have exhibited superior performance across diverse applications. Despite their successes, several limitations still exist. First, existing models in the meta-training phase predominantly focus on instance-level features within tasks, neglecting crucial set-level features essential for distinguishing between different categories. Second, these models often utilize query sets directly on classifiers trained with support sets containing only a few labeled examples, overlooking potential distribution shifts between these sets and leading to suboptimal performance. Finally, previous models typically require necessitate abundant labeled data from base classes to extract transferable knowledge, which is typically infeasible in real-world scenarios. To address these issues, we propose a novel model named STAR, which leverages Set funcTions and optimAl tRansport for enhancing unsupervised graph few-shot learning. Specifically, STAR utilizes expressive set functions to obtain set-level features in an unsupervised manner and employs optimal transport principles to align the distributions of support and query sets, thereby mitigating distribution shift effects. Theoretical analysis demonstrates that STAR can capture more task-relevant information and enhance generalization capabilities. Empirically, extensive experiments across multiple datasets validate the effectiveness of STAR. Our code can be found here.

Graph neural networks have been demonstrated as a powerful paradigm for effectively learning graph-structured data on the web and mining content from it. %the wide web. for downstream task analysis. Current leading graph models require a large number of labeled samples for training, which unavoidably leads to overfitting in few-shot scenarios. Recent research has sought to alleviate this issue by simultaneously leveraging graph learning and meta-learning paradigms. However, these graph meta-learning models assume the availability of numerous meta-training tasks to learn transferable meta-knowledge. Such assumption may not be feasible in the real world due to the difficulty of constructing tasks and the substantial costs involved. Therefore, we propose a SiMple yet effectIve approach for graph few-shot Learning with fEwer tasks, named SMILE. We introduce a dual-level mixup strategy, encompassing both within-task and across-task mixup, to simultaneously enrich the available nodes and tasks in meta-learning. Moreover, we explicitly leverage the prior information provided by the node degrees in the graph to encode expressive node representations. Theoretically, we demonstrate that SMILE can enhance the model generalization ability. Empirically, SMILE consistently outperforms other competitive models by a large margin across all evaluated datasets with in-domain and cross-domain settings. Our anonymous code can be found https://github.com/KEAML-JLU/SMILE.

Graphs, as a fundamental data structure, have proven efficacy in modeling complex relationships between objects and are therefore found in wide web applications. Graph classification is an essential task in graph data analysis, which can effectively assist in extracting information and mining content from the web. Recently, few-shot graph classification, a more realistic and challenging task, has garnered great research interest. Existing few-shot graph classification models are all supervised, assuming abundant labeled data in base classes for meta-training. However, sufficient annotation is often challenging to obtain in practice due to high costs or demand for expertise. Moreover, they commonly adopt complicated meta-learning algorithms via episodic training to transfer prior knowledge from base classes. To break free from these constraints, in this paper, we propose a simple yet effective approach named SMART for unsupervised few-shot graph classification without using any labeled data. SMART employs transfer learning philosophy instead of the previously prevailing meta-learning paradigm, avoiding the need for sophisticated meta-learning algorithms. Additionally, we adopt a novel mixup strategy to augment the original graph data and leverage unsupervised pretraining on these data to obtain the expressive graph encoder. We also utilize the prompt tuning technique to alleviate the overfitting and low fine-tuning efficiency caused by the limited support samples of novel classes. Extensive experimental results demonstrate the superiority of our proposed approach, significantly surpassing even leading supervised few-shot graph classification models. Our code is available here.

Node classification is an essential problem in graph learning. However, many models typically obtain unsatisfactory performance when applied to few-shot scenarios. Some studies have attempted to combine meta-learning with graph neural networks to solve few-shot node classification on graphs. Despite their promising performance, some limitations remain. First, they employ the node encoding mechanism of homophilic graphs to learn node embeddings, even in heterophilic graphs. Second, existing models based on meta-learning ignore the interference of randomness in the learning process. Third, they are trained using only limited labeled nodes within the specific task, without explicitly utilizing numerous unlabeled nodes. Finally, they treat almost all sampled tasks equally without customizing them for their uniqueness. To address these issues, we propose a novel framework for few-shot node classification called Meta-GPS \(++\) . Specifically, we first adopt an efficient method to learn discriminative node representations on homophilic and heterophilic graphs. Then, we leverage a prototype-based approach to initialize parameters and contrastive learning for regularizing the distribution of node embeddings. Moreover, we apply self-training to extract valuable information from unlabeled nodes. Additionally, we adopt S \({}^{2}\) (scaling and shifting) transformation to learn transferable knowledge from diverse tasks. The results on real-world datasets show the superiority of Meta-GPS \(++\) . Our code is available here .

Text clustering is a critical step in text data analysis and has been extensively studied by the text mining community. Most existing text clustering algorithms are based on the bag-of-words model, which faces the high-dimensional and sparsity problems and ignores text structural and sequence information. Deep learning-based models such as convolutional neural networks and recurrent neural networks regard texts as sequences but lack supervised signals and explainable results. In this paper, we propose a deep feature-based text clustering (DFTC) framework that incorporates pretrained text encoders into text clustering tasks. This model, which is based on sequence representations, breaks the dependency on supervision. The experimental results show that our model outperforms classic text clustering algorithms on almost all the considered datasets. In addition, the explanation of the clustering results is significant for understanding the principles of the deep learning approach. Our proposed clustering framework includes an explanation module that can help users understand the meaning and quality of the clustering results. Our code is available at https://github.com/KEAML-JLU/DeepTextClustering.

The cold-start problem is a major factor that limits the effectiveness of recommendation systems. Having too few available interaction records brings a series of challenges when predicting user preferences. At present, there are two main kinds of strategies for solving this problem from different perspectives. One is cross-domain recommendation (CDR), which introduces additional information by domain knowledge propagation with transfer learning. However, CDR methods follow traditional training processes in machine learning and cannot solve this typical few-shot problem from the perspective of optimization. The other type of methods that has recently emerged is based on meta-learning. Most of these approaches focus only on generating a meta-model to perform better on new tasks and ignore improvements based on cross-domain information. Therefore, it is necessary to design a novel approach to solve this problem with both domain knowledge and meta-optimization. To achieve this goal, a novel cross-domain meta-learner for cold-start recommendation (MetaCDR) is proposed. In MetaCDR, we design a domain knowledge meta-transfer module to connect different domain networks. In addition, we introduce a pretraining strategy to ensure its efficiency. The experimental results show that MetaCDR performs significantly better than state-of-the-art models in a variety of scenarios.

User cold-start is a major challenge in building personalized recommender systems. Due to the lack of sufficient interactions, it is difficult to effectively model new users. One of the main solutions is to obtain an initial model through meta-learning (mainly gradient-based methods) and adapt it to new users with a few steps of gradient descent. Although these methods have achieved remarkable performance, they are still far from being usable in real-world applications due to their high-demand data processing, heavy computational burden, and inability to perform effective user-incremental update. In this paper, we propose a d eployable and c ontinuable m eta-learning-based r ecommendation (DCMR) approach, which can achieve fast user-incremental updating with task replay and first-order gradient descent. Specifically, we introduce a dual-constrained task sampler, distillation-based loss functions, and an adaptive controller in this framework to balance the trade-off between stability and plasticity in updating. In summary, DCMR can be updated while serving new users; in other words, it learns continuously and rapidly from a sequential user stream and is able to make recommendations at any time. The extensive experiments conducted on three benchmark datasets illustrate the superiority of our model.

Attributed networks, as a manifestation of data in non-Euclidean domains, have a wide range of applications in the real world, such as molecular property prediction, social network analysis and anomaly detection. Node classification, as a fundamental research problem in attributed networks, has attracted increasing attention among research communities. However, most existing models cannot be directly applied to the data with limited labeled instances (\textiti.e., the few-shot scenario). Few-shot node classification on attributed networks is gradually becoming a research hotspot. Although several methods aim to integrate meta-learning with graph neural networks to address this problem, some limitations remain. First, they all assume node representation learning using graph neural networks in homophilic graphs. %Hence, suboptimal performance is obtained when these models are applied to heterophilic graphs. Second, existing models based on meta-learning entirely depend on instance-based statistics. %which in few-shot settings are unavoidably degraded by data noise or outliers. Third, most previous models treat all sampled tasks equally and fail to adapt their uniqueness. %which has a significant impact on the overall performance of the model. To solve the above three limitations, we propose a novel graph Meta -learning framework called G raph learning based on P rototype and S caling & shifting transformation (Meta-GPS ). More specifically, we introduce an efficient method for learning expressive node representations even on heterophilic graphs and propose utilizing a prototype-based approach to initialize parameters in meta-learning. Moreover, we also leverage S$^2$ (scaling & shifting) transformation to learn effective transferable knowledge from diverse tasks. Extensive experimental results on six real-world datasets demonstrate the superiority of our proposed framework, which outperforms other state-of-the-art baselines by up to 13% absolute improvement in terms of related metrics.

User cold-start recommendation is a serious problem that limits the performance of recommender systems (RSs). Recent studies have focused on treating this issue as a few-shot problem and seeking solutions with model-agnostic meta-learning (MAML). Such methods regard making recommendations for one user as a task and adapt to new users with a few steps of gradient updates on the meta-model. However, none of those methods consider the limitation of user representation learning imposed by the special task setting of MAML-based RSs. And they learn a common meta-model for all users while ignoring the implicit grouping distribution induced by the correlation differences among users. In response to the above problems, we propose a pretrained network modulation and task adaptation approach (PNMTA) for user cold-start recommendation. In the pretraining stage, a pretrained model is obtained with non-meta-learning methods to achieve better user representation and generalization, which can also transfer the learned knowledge to the meta-learning stage for modulation. During the meta-learning stage, an encoder modulator is utilized to realize the memorization and correction of prior parameters for the meta-learning task, and a predictor modulator is introduced to condition the model initialization on the task identity for adaptation steps. In addition, PNMTA can also make use of the existing non-cold-start users for pretraining. Comprehensive experiments on two benchmark datasets demonstrate that our model can achieve significant and consistent improvements against other state-of-the-art methods.

Text clustering is a critical step in text data analysis and has been extensively studied by the text mining community. Most existing text clustering algorithms are based on the bag-of-words model, which faces the high-dimensional and sparsity problems and ignores text structural and sequence information. Deep learning-based models such as convolutional neural networks and recurrent neural networks regard texts as sequences but lack supervised signals and explainable results. In this paper, we propose adeepfeature-basedtextclustering (DFTC) framework that incorporates pretrained text encoders into text clustering tasks. This model, which is based on sequence representations, breaks the dependency on supervision. The experimental results show that our model outperforms classic text clustering algorithms and the state-of-the-art pretrained language model, i.e., BERT, on almost all the considered datasets. In addition, the explanation of the clustering results is significant for understanding the principles of the deep learning approach. Our proposed clustering framework includes an explanation module that can help users understand the meaning and quality of the clustering results.

Paper-publication venue prediction aims to predict candidate publication venues that effectively suit given submissions. This technology is developing rapidly with the popularity of machine learning models. However, most previous methods ignore the structure information of papers, while modeling them with graphs can naturally solve this drawback. Meanwhile, they either use hand-crafted or bag-of-word features to represent the papers, ignoring the ones that involve high-level semantics. Moreover, existing methods assume that the venue where a paper is published as a correct venue for the data annotation, which is unrealistic. One paper can be relevant to many venues. In this paper, we attempt to address these problems above and develop a novel prediction model, namelyVenue Prediction with Abstract-Level Graph (Vpalg xspace), which can serve as an effective decision-making tool for venue selections. Specifically, to achieve more discriminative paper abstract representations, we construct each abstract as a semantic graph and perform a dual attention message passing neural network for representation learning. Then, the proposed model can be trained over the learned abstract representations with their labels and generalized via self-training. Empirically, we employ the PubMed dataset and further collect two new datasets from the top journals and conferences in computer science. Experimental results indicate the superior performance of Vpalg xspace, consistently outperforming the existing baseline methods.

Based on an effective clustering algorithm-Affinity Propagation (AP)-we present in this paper a novel semisupervised text clustering algorithm, called Seeds Affinity Propagation (SAP). There are two main contributions in our approach: 1) a new similarity metric that captures the structural information of texts, and 2) a novel seed construction method to improve the semisupervised clustering process. To study the performance of the new algorithm, we applied it to the benchmark data set Reuters-21578 and compared it to two state-of-the-art clustering algorithms, namely, k-means algorithm and the original AP algorithm. Furthermore, we have analyzed the individual impact of the two proposed contributions. Results show that the proposed similarity metric is more effective in text clustering (F-measures ca. 21 percent higher than in the AP algorithm) and the proposed semisupervised strategy achieves both better clustering results and faster convergence (using only 76 percent iterations of the original AP). The complete SAP algorithm obtains higher F-measure (ca. 40 percent improvement over k-means and AP) and lower entropy (ca. 28 percent decrease over k-means and AP), improves significantly clustering execution time (20 times faster) in respect that k-means, and provides enhanced robustness compared with all other methods.