Large language models (LLMs) have demonstrated strong capabilities in structured query generation, making them a natural choice for Text-to-SPARQL, which translates natural language questions into executable SPARQL queries over knowledge graphs. However, their initial outputs remain unreliable: generated queries may be executable yet semantically misaligned with input questions, leading to incorrect retrieval. To address this issue, we propose Generator-Gate-Corrector (GGC), a framework for reliable LLM-based Text-to-SPARQL generation. GGC first uses a Generator to produce an initial query, then applies a Gate to predict whether correction is needed, and finally invokes a Corrector only for selected high-risk queries. This selective correction mechanism avoids unnecessary modifications and reduces the risk of degrading originally correct queries. Experiments on MCQA show that GGC improves query-level accuracy from 90.23% to 98.33% while reducing inference overhead by 45% compared with correcting all generated queries. Ablation studies show that the Gate is robust across thresholds and that Corrector training data composition affects correction effectiveness and stability. Overall, the results demonstrate that selective correction enhances the accuracy, reliability, and efficiency of LLM-based text-to-SPARQL generation.

Understanding how people argue across ideological divides online is important for studying political polarization, misinformation, and content moderation. Existing datasets capture only part of this problem: some preserve text but ignore interaction structure, some model structure without rich semantics, and others represent conversations without stable user-level ideological identity. We introduce ControBench, a benchmark for controversial discourse analysis that combines heterogeneous social interaction graphs with rich textual semantics. Built from Reddit discussions on three topics, Trump, abortion, and religion, ControBench contains 7,370 users, 1,783 posts, and 26,525 interactions. The graph contains user and post nodes connected by semantically enriched edges; in particular, user-comment-user edges encode both a reply and the parent comment that it responds to, preserving local argumentative context. User labels are derived from self-declared Reddit flairs, providing a scalable proxy for ideological identity without manual annotation. The resulting datasets exhibit low or negative adjusted homophily (Trump: -0.77, Abortion: 0.06, Religion: 0.04), reflecting the cross-cutting structure of real-world debate. We evaluate graph neural networks, pretrained language models, and large language models on ControBench and observe distinct performance patterns across topics and model families, especially when ideological boundaries are ambiguous. These results position ControBench as a challenging and realistic benchmark for controversial discourse analysis.

Reasoning models enhance problem-solving by scaling test-time compute, yet they face a critical paradox: excessive thinking tokens often degrade performance rather than improve it. We attribute this to a fundamental architectural flaw: standard LLMs operate as "malloc-only" engines, continuously accumulating valid and redundant steps alike without a mechanism to prune obsolete information. To break this cycle, we propose Free()LM, a model that introduces an intrinsic self-forgetting capability via the Free-Module, a plug-and-play LoRA adapter. By iteratively switching between reasoning and cleaning modes, Free()LM dynamically identifies and prunes useless context chunks, maintaining a compact and noise-free state. Extensive experiments show that Free()LM provides consistent improvements across all model scales (8B to 685B). It achieves a 3.3

Parody is an emerging phenomenon on social media, where individuals imitate a role or position opposite to their own, often for humor, provocation, or controversy. Detecting and analyzing parody can be challenging and is often reliant on context, yet it plays a crucial role in understanding cultural values, promoting subcultures, and enhancing self-expression. However, the study of parody is hindered by limited available data and deficient diversity in current datasets. To bridge this gap, we built seven parody datasets from both English and Chinese corpora, with 14,755 annotated users and 21,210 annotated comments in total. To provide sufficient context information, we also collect replies and construct user-interaction graphs to provide richer contextual information, which is lacking in existing datasets. With these datasets, we test traditional methods and Large Language Models (LLMs) on three key tasks: (1) parody detection, (2) comment sentiment analysis with parody, and (3) user sentiment analysis with parody. Our extensive experiments reveal that parody-related tasks still remain challenging for all models, and contextual information plays a critical role. Interestingly, we find that, in certain scenarios, traditional sentence embedding methods combined with simple classifiers can outperform advanced LLMs, i.e. DeepSeek-R1 and GPT-o3, highlighting parody as a significant challenge for LLMs.

Under circumstances of heterophily, where nodes with different labels tend to be connected based on semantic meanings, Graph Neural Networks (GNNs) often exhibit suboptimal performance. Current studies on graph heterophily mainly focus on aggregation calibration or neighbor extension and address the heterophily issue by utilizing node features or structural information to improve GNN representations. In this paper, we propose and demonstrate that the valuable semantic information inherent in heterophily can be utilized effectively in graph learning by investigating the distribution of neighbors for each individual node within the graph. The theoretical analysis is carried out to demonstrate the efficacy of the idea in enhancing graph learning. Based on this analysis, we propose HiGNN, an innovative approach that constructs an additional new graph structure, that integrates heterophilous information by leveraging node distribution to enhance connectivity between nodes that share similar semantic characteristics. We conduct empirical assessments on node classification tasks using both homophilous and heterophilous benchmark datasets and compare HiGNN to popular GNN baselines and SoTA methods, confirming the effectiveness in improving graph representations. In addition, by incorporating heterophilous information, we demonstrate a notable enhancement in existing GNN-based approaches, and the homophily degree across real-world datasets, thus affirming the efficacy of our approach.

Retrieval-Augmented Generation (RAG) systems enable large language models (LLMs) instant access to relevant information for the generative process, demonstrating their superior performance in addressing common LLM challenges such as hallucination, factual inaccuracy, and the knowledge cutoff. Graph-based RAG further extends this paradigm by incorporating knowledge graphs (KGs) to leverage rich, structured connections for more precise and inferential responses. A critical challenge, however, is that most Graph-based RAG systems rely on LLMs for automated KG construction, often yielding noisy KGs with redundant entities and unreliable relationships. This noise degrades retrieval and generation performance while also increasing computational cost. Crucially, current research does not comprehensively address the denoising problem for LLM-generated KGs. In this paper, we introduce DEnoised knowledge Graphs for Retrieval Augmented Generation (DEG-RAG), a framework that addresses these challenges through: (1) entity resolution, which eliminates redundant entities, and (2) triple reflection, which removes erroneous relations. Together, these techniques yield more compact, higher-quality KGs that significantly outperform their unprocessed counterparts. Beyond the methods, we conduct a systematic evaluation of entity resolution for LLM-generated KGs, examining different blocking strategies, embedding choices, similarity metrics, and entity merging techniques. To the best of our knowledge, this is the first comprehensive exploration of entity resolution in LLM-generated KGs. Our experiments demonstrate that this straightforward approach not only drastically reduces graph size but also consistently improves question answering performance across diverse popular Graph-based RAG variants.

Graph Neural Networks (GNNs) have demonstrated strong capabilities in processing structured data. While traditional GNNs typically treat each feature dimension equally important during graph convolution, we raise an important question: **Is the graph convolution operation equally beneficial for each feature?** If not, the convolution operation on certain feature dimensions can possibly lead to harmful effects, even worse than convolution-free models. Therefore, it is required to distinguish convolution-favored and convolution-disfavored features. Traditional feature selection methods mainly focus on identifying informative features or reducing redundancy, but they are not suitable for structured data as they overlook graph structures. In graph community, some studies have investigated the performance of GNN with respect to node features using feature homophily metrics, which assess feature consistency across graph topology. Unfortunately, these metrics do not effectively align with GNN performance and cannot be reliably used for feature selection in GNNs. To address these limitations, we introduce a novel metric, Topological Feature Informativeness (TFI), to distinguish GNN-favored and GNN-disfavored features, where its effectiveness is validated through both theoretical analysis and empirical observations. Based on TFI, we propose a simple yet effective Graph Feature Selection (GFS) method, which processes GNN-favored and GNN-disfavored features with GNNs and non-GNN models separately. Compared to original GNNs, GFS significantly improves the extraction of useful topological information from each feature with comparable computational costs. Extensive experiments show that after applying GFS to $\textbf{8}$ baseline and state-of-the-art (SOTA) GNN architectures across $\textbf{10}$ datasets, $\textbf{90\%}$ of the GFS-augmented cases show significant performance boosts. Furthermore, our proposed TFI metric outperforms other feature selection methods for GFS. These results verify the effectiveness of both GFS and TFI. Additionally, we demonstrate that GFS's improvements are robust to hyperparameter tuning, highlighting its potential as a universally valid method for enhancing various GNN architectures.

Graph homophily refers to the phenomenon that connected nodes tend to share similar characteristics. Understanding this concept and its related metrics is crucial for designing effective Graph Neural Networks (GNNs). The most widely used homophily metrics, such as edge or node homophily, quantify such "similarity" as label consistency across the graph topology. These metrics are believed to be able to reflect the performance of GNNs, especially on node-level tasks. However, many recent studies have empirically demonstrated that the performance of GNNs does not always align with homophily metrics, and how homophily influences GNNs still remains unclear and controversial. Then, a crucial question arises: What is missing in our current understanding of homophily? To figure out the missing part, in this paper, we disentangle the graph homophily into three aspects: label, structural, and feature homophily, which are derived from the three basic elements of graph data. We argue that the synergy of the three homophily can provide a more comprehensive understanding of GNN performance. Our new proposed structural and feature homophily consider the neighborhood consistency and feature dependencies among nodes, addressing the previously overlooked structural and feature aspects in graph homophily. To investigate their synergy, we propose a Contextual Stochastic Block Model with three types of Homophily (CSBM-3H), where the topology and feature generation are controlled by the three metrics. Based on the theoretical analysis of CSBM-3H, we derive a new composite metric, named Tri-Hom, that considers all three aspects and overcomes the limitations of conventional homophily metrics. The theoretical conclusions and the effectiveness of Tri-Hom have been verified through synthetic experiments on CSBM-3H. In addition, we conduct experiments on $31$ real-world benchmark datasets and calculate the correlations between homophily metrics and model performance. Tri-Hom has significantly higher correlation values than $17$ existing metrics that only focus on a single homophily aspect, demonstrating its superiority and the importance of homophily synergy. Our code is available at https://github.com/zylMozart/Disentangle_GraphHom.

To improve the performance of Graph Neural Networks (GNNs), Graph Structure Learning (GSL) has been extensively applied to reconstruct or refine original graph structures, effectively addressing issues like heterophily, over-squashing, and noisy structures. While GSL is generally thought to improve GNN performance, it often leads to longer training times and more hyperparameter tuning. Besides, the distinctions among current GSL methods remain ambiguous from the perspective of GNN training, and there is a lack of theoretical analysis to quantify their effectiveness. Recent studies further suggest that, under fair comparisons with the same hyperparameter tuning, GSL does not consistently outperform baseline GNNs. This motivates us to ask a critical question: is GSL really useful for GNNs? To address this question, this paper makes two key contributions. First, we propose a new GSL framework, which includes three steps: GSL base (the representation used for GSL) construction, new structure construction, and view fusion, to better understand the effectiveness of GSL in GNNs. Second, after graph convolution, we analyze the differences in mutual information (MI) between node representations derived from the original topology and those from the newly constructed topology. Surprisingly, our empirical observations and theoretical analysis show that no matter which type of graph structure construction methods are used, after feeding the same GSL bases to the newly constructed graph, there is no MI gain compared to the original GSL bases. To fairly reassess the effectiveness of GSL, we conduct ablation experiments and find that it is the pretrained GSL bases that enhance GNN performance, and in most cases, GSL cannot improve GNN performance. This finding encourages us to rethink the essential components in GNNs, such as self-training and structural encoding, in GNN design rather than GSL.

To address the problem of poor entity recognition performance caused by the lack of Chinese annotation in clinical electronic medical records, this paper proposes a multi-medical entity recognition method F-MNER using a fusion technique combining BART, Bi-LSTM, and CRF. First, after cleaning, encoding, and segmenting the electronic medical records, the obtained semantic representations are dynamically fused using a bidirectional autoregressive transformer (BART) model. Then, sequential information is captured using a bidirectional long short-term memory (Bi-LSTM) network. Finally, the conditional random field (CRF) is used to decode and output multi-task entity recognition. Experiments are performed on the CCKS2019 dataset, with micro avg Precision , macro avg Recall , weighted avg Precision reaching 0.880, 0.887, and 0.883, and micro avg F1-score , macro avg F1-score , weighted avg F1-score reaching 0.875, 0.876, and 0.876 respectively. Compared with existing models, our method outperforms the existing literature in three evaluation metrics ( micro average , macro average , weighted average ) under the same dataset conditions. In the case of weighted average, the Precision , Recall , and F1-score are 19.64%, 15.67%, and 17.58% higher than the existing BERT-BiLSTM-CRF model respectively. Experiments are performed on the actual clinical dataset with our MF-MNER, the Precision , Recall , and F1-score are 0.638, 0.825, and 0.719 under the micro-avg evaluation mechanism. The Precision , Recall , and F1-score are 0.685, 0.800, and 0.733 under the macro-avg evaluation mechanism. The Precision , Recall , and F1-score are 0.647, 0.825, and 0.722 under the weighted avg evaluation mechanism. The above results show that our method MF-MNER can integrate the advantages of BART, Bi-LSTM, and CRF layers, significantly improving the performance of downstream named entity recognition tasks with a small amount of annotation, and achieving excellent performance in terms of recall score, which has certain practical significance. Source code and datasets to reproduce the results in this paper are available at https://github.com/xfwang1969/MF-MNER . Graphical Abstract Illustration of the proposed MF-MNER. The method mainly includes four steps: (1) medical electronic medical records need to be cleared, coded, and segmented. (2) The semantic representation obtained by dynamic fusion of the bidirectional autoregressive converter (BART) model. (3) The sequence information is captured by a bi-directional short-term memory (Bi-LSTM) network. (4) the multi-task entity recognition is decoded and output by conditional random field (CRF).

Graph Neural Networks (GNNs) have demonstrated strong capabilities in processing structured data. While traditional GNNs typically treat each feature dimension equally during graph convolution, we raise an important question: Is the graph convolution operation equally beneficial for each feature? If not, the convolution operation on certain feature dimensions can possibly lead to harmful effects, even worse than the convolution-free models. In prior studies, to assess the impacts of graph convolution on features, people proposed metrics based on feature homophily to measure feature consistency with the graph topology. However, these metrics have shown unsatisfactory alignment with GNN performance and have not been effectively employed to guide feature selection in GNNs. To address these limitations, we introduce a novel metric, Topological Feature Informativeness (TFI), to distinguish between GNN-favored and GNN-disfavored features, where its effectiveness is validated through both theoretical analysis and empirical observations. Based on TFI, we propose a simple yet effective Graph Feature Selection (GFS) method, which processes GNN-favored and GNN-disfavored features separately, using GNNs and non-GNN models. Compared to original GNNs, GFS significantly improves the extraction of useful topological information from each feature with comparable computational costs. Extensive experiments show that after applying GFS to 8 baseline and state-of-the-art (SOTA) GNN architectures across 10 datasets, 83.75 GFS-augmented cases show significant performance boosts. Furthermore, our proposed TFI metric outperforms other feature selection methods. These results validate the effectiveness of both GFS and TFI. Additionally, we demonstrate that GFS's improvements are robust to hyperparameter tuning, highlighting its potential as a universal method for enhancing various GNN architectures.

Sentence representation is one of the most fundamental research topics in natural language processing (NLP), as its quality directly affects various downstream task performances. Recent studies for sentence representations have established state-of-the-art (SOTA) performance on semantic representation tasks. However, embeddings by those approaches share unsatisfying transferability when applied to various specific applications. Seldom work studies the transferability of semantic sentence embeddings. In this paper, we first explore the transferability characteristic of the sentence embeddings, and present BlendCSE, a new sentence embedding model targeting rich semantics and transferability. BlendCSE blends three recent advanced NLP learning methodologies, namely, continue learning on masked language modeling (MLM), contrastive learning (CL) with data augmentations (DA), and semantic supervised learning. The main objectives of BlendCSE are to capture token/word level information, diversified linguistic properties, and sentence semantics, respectively. Empirical studies demonstrate that BlendCSE captures semantics comparably well on STS tasks, yet surpasses existing methods on various transfer tasks, yielding even stronger transferability on document-level applications. Ablation studies verified that the three learning objectives synergy well to capture semantics and transferability effectively.

This paper improves contrastive learning for sentence embeddings from two perspectives: handling dropout noise and addressing feature corruption. Specifically, for the first perspective, we identify that the dropout noise from negative pairs affects the model's performance. Therefore, we propose a simple yet effective method to deal with such type of noise. Secondly, we pinpoint the rank bottleneck of current solutions to feature corruption and propose a dimension-wise contrastive learning objective to address this issue. Both proposed methods are generic and can be applied to any contrastive learning based models for sentence embeddings. Experimental results on standard benchmarks demonstrate that combining both proposed methods leads to a gain of 1.8 points compared to the strong baseline SimCSE configured with BERT base. Furthermore, applying the proposed method to DiffCSE, another strong contrastive learning based baseline, results in a gain of 1.4 points.

Despite being studied for decades, current clustering approaches still face difficulties in handling high-dimensional datasets. For high-dimensional data, it is essential to get a good feature representation for the clustering algorithm to conduct on. Most of the clustering algorithms do not explicitly encourage the preservation of pairwise distance within the input data when learning feature representation. We proposed an improvement over the Deep Embedded Clustering (DEC) by including a penalty term in the loss function for the differences between the input data and the random projection of its corresponding feature embedding. The idea behind this penalty term is one of the properties of random projection, that is the pairwise distance is preserved between the low-dimensional manifold and the data space. In this way, the network encourages the learning towards preserving data similarities in the feature space. We named the proposed method as DEC-RPP. The experiments show significant improvements on clustering metrics over four datasets compared with the baseline DEC and another work IDEC that improves DEC by preserving local structure.

Neural network pruning is a practical way for reducing the size of trained models and the number of floating-point operations. One way of pruning is to use the relative Hessian trace to calculate sensitivity of each channel, as compared to the more common magnitude pruning approach. However, the stochastic approach used to estimate the Hessian trace needs to iterate over many times before it can converge. This can be time-consuming when used for larger models with many millions of parameters. To address this problem, we modify the existing approach by estimating the Hessian trace using FP16 precision instead of FP32. We test the modified approach (EHAP) on ResNet-32/ResNet-56/WideResNet-28-8 trained on CIFAR10/CIFAR100 image classification tasks and achieve faster computation of the Hessian trace. Specifically, our modified approach can achieve speed ups ranging from 17% to as much as 44% during our experiments on different combinations of model architectures and GPU devices. Our modified approach also takes up around 40% less GPU memory when pruning ResNet-32 and ResNet-56 models, which allows for a larger Hessian batch size to be used for estimating the Hessian trace. Meanwhile, we also present the results of pruning using both FP16 and FP32 Hessian trace calculation and show that there are no noticeable accuracy differences between the two. Overall, it is a simple and effective way to compute the relative Hessian trace faster without sacrificing on pruned model performance. We also present a full pipeline using EHAP and quantization aware training (QAT), using INT8 QAT to compress the network further after pruning. In particular, we use symmetric quantization for the weights and asymmetric quantization for the activations.

This paper proposes the DistillCSE framework, which performs contrastive learning under the self-training paradigm with knowledge distillation. The potential advantage of DistillCSE is its self-enhancing feature: using a base model to provide additional supervision signals, a stronger model may be learned through knowledge distillation. However, the vanilla DistillCSE through the standard implementation of knowledge distillation only achieves marginal improvements due to severe overfitting. The further quantitative analyses demonstrate the reason that the standard knowledge distillation exhibits a relatively large variance of the teacher model's logits due to the essence of contrastive learning. To mitigate the issue induced by high variance, this paper accordingly proposed two simple yet effective solutions for knowledge distillation: a Group-P shuffling strategy as an implicit regularization and the averaging logits from multiple teacher components. Experiments on standard benchmarks demonstrate that the proposed DistillCSE outperforms many strong baseline methods and yields a new state-of-the-art performance.

The construction of S-scheme heterostructure is a prospective strategy to remove polycyclic aromatic hydrocarbons (PAHs) from simulated seawater. In this study, a novel Bi2S3/CeVO4 S-scheme heterojunction photocatalyst was prepared by solvent thermal method. The successful construction of S-scheme heterostructure is beneficial to the separation and migration of photogenerated carriers, thus achieving excellent redox capacity and efficient production of free radicals, which is conducive to promoting the photocatalyst to transform pollutants into H2O, CO2. In pure water, the optimized Bi2S3/CeVO4 heterojunction (Bi/Ce-30) achieved a high degradation efficiency (97.39%) for naphthalene (NAP). In simulated seawater, the degradation efficiency of NAP still reaches 87.20% (k = 0.01653 min-1), which is 7.58 folds higher than pure CeVO4. Furthermore, the effect (such as solution pH, pollutant concentration, etc.) on the photocatalytic performance of Bi2S3/CeVO4 heterojunction was investigated. The possible photocatalytic degradation processes of NAP were qualitatively analyzed using 3D fluorescence spectroscopy and HPLC-MS to reveal the major degradation pathways. In addition, toxicity assessment software was used to demonstrate the gradual decrease in toxicity of the intermediate products. Furthermore, we elucidated the possible photocatalytic mechanism of Bi2S3/CeVO4 S-scheme heterojunction for the degradation of NAP. This work provides new insights into the realization of S-scheme heterojunction for removing polycyclic aromatic hydrocarbons from natural seawater.

Hierarchical multi-label text classification (HMTC) has been gaining popularity in recent years thanks to its applicability to a plethora of real-world applications. The existing HMTC algorithms largely focus on the design of classifiers, such as the local, global, or a combination of them. However, very few studies have focused on hierarchical feature extraction and explore the association between the hierarchical labels and the text. In this paper, we propose a Label-based Attention for Hierarchical Multi-label Text Classification Neural Network (LA-HCN), where the novel label-based attention module is designed to hierarchically extract important information from the text based on the labels from different hierarchy levels. Besides, hierarchical information is shared across levels while preserving the hierarchical label-based information. Separate local and global document embeddings are obtained and used to facilitate the respective local and global classifications. In our experiments, LA-HCN outperforms other state-of-the-art neural network-based HMTC algorithms on four public HMTC datasets. The ablation study also demonstrates the effectiveness of the proposed label-based attention module as well as the novel local and global embeddings and classifications. By visualizing the learned attention (words), we find that LA-HCN is able to extract meaningful information corresponding to the different labels which provides explainability that may be helpful for the human analyst.

Back translation (BT) is one of the most significant technologies in NMT research fields. Existing attempts on BT share a common characteristic: they employ either beam search or random sampling to generate synthetic data with a backward model but seldom work studies the role of synthetic data in the performance of BT. This motivates us to ask a fundamental question: what kind of synthetic data contributes to BT performance? Through both theoretical and empirical studies, we identify two key factors on synthetic data controlling the back-translation NMT performance, which are quality and importance. Furthermore, based on our findings, we propose a simple yet effective method to generate synthetic data to better trade off both factors so as to yield a better performance for BT. We run extensive experiments on WMT14 DE-EN, EN-DE, and RU-EN benchmark tasks. By employing our proposed method to generate synthetic data, our BT model significantly outperforms the standard BT baselines (i.e., beam and sampling based methods for data generation), which proves the effectiveness of our proposed methods.

Heterogeneous information networks(HINs) become popular in recent years for its strong capability of modelling objects with abundant information using explicit network structure. Network embedding has been proved as an effective method to convert information networks into lower-dimensional space, whereas the core information can be well preserved. However, traditional network embedding algorithms are sub-optimal in capturing rich while potentially incompatible semantics provided by HINs. To address this issue, a novel meta-path-based HIN representation learning framework named mSHINE is designed to simultaneously learn multiple node representations for different meta-paths. More specifically, one representation learning module inspired by the RNN structure is developed and multiple node representations can be learned simultaneously, where each representation is associated with one respective meta-path. By measuring the relevance between nodes with the designed objective function, the learned module can be applied in downstream link prediction tasks. A set of criteria for selecting initial meta-paths is proposed as the other module in mSHINE which is important to reduce the optimal meta-path selection cost when no prior knowledge of suitable meta-paths is available. To corroborate the effectiveness of mSHINE, extensive experimental studies including node classification and link prediction are conducted on five real-world datasets. The results demonstrate that mSHINE outperforms other state-of-the-art HIN embedding methods.