Detecting AI-generated images remains a persistent challenge, as existing detectors often struggle to generalize to forgeries produced by previously unseen generative models. This generalization gap mainly stems from entanglement with semantic content and overfitting to model-specific artifacts. Moreover, many state-of-the-art methods rely on large pre-trained backbones or computationally intensive pipelines, which limit their applicability in real-world, resource-constrained environments. We propose RealNet, a lightweight and unsupervised framework that constructs a disentangled, forgery-aware representation space using only real images. RealNet first extracts semantic-agnostic representations through a dual adversarial denoising mechanism, producing compact features with low intra-class variance. These representations are then perturbed in feature space to generate pseudo-negative samples, which are combined with the original real features to train a lightweight discriminator, enabling robust detection without any dependence on synthetic images during training. Comprehensive evaluations across GAN, diffusion, and emerging VAR-based paradigms demonstrate that RealNet achieves superior cross-model generalization and robustness. RealNet surpasses previous state-of-the-art approaches by 4.51% in accuracy and 3.93% in average precision, while maintaining significantly lower computational cost. Furthermore, we introduce a medically relevant synthetic image dataset and show RealNet remains effective under severe distribution shifts, highlighting its potential for deployment in high-stakes real-world scenarios. Together, these advantages position RealNet as a practical, scalable and socially impactful solution for robust AI-generated image detection.

Accurately recognizing distracted driving activities in real-world scenarios is essential for improving road and pedestrian safety. However, existing approaches are prone to attending to irrelevant scene context and are susceptible to interference from redundant frames, compromising their robustness in complex driving environments. To overcome these limitations, we propose DualScope, a novel framework that captures behaviorally critical information from both spatial and temporal perspectives. In the spatial domain, we introduce a Synergistic Behavior-Centric Distillation mechanism that leverages two key information sources: (1) position-aware knowledge derived from the SAM model, which enhances the perception of critical regions and their semantic interaction structures; and (2) fine-grained visual details obtained from cropped key regions, which improve the model's ability to capture detailed patterns within behavior-relevant areas. In the temporal domain, we present the Saliency-Aware Fine-to-Coarse Temporal Modeling module, comprising three components: a Fine-Grained Motion Encoder for capturing local inter-frame dependencies; a Dynamic Difference Extractor for generating salient motion dynamics; and a Saliency-Aware Temporal Pyramid Mamba for integrating these representations to enable multi-scale temporal modeling. This design effectively captures both short-term motions and long-term behavioral patterns. Furthermore, incorporating salient dynamics enhances the model's focus on significant behavioral variations. Extensive experiments on seven publicly available DDAR datasets demonstrate that DualScope consistently outperforms state-of-the-art methods, validating its effectiveness in capturing behavioral cues across spatial and temporal dimensions.

Transformer-based large language models (LLMs) excel in natural language processing tasks by capturing long-range dependencies through self-attention mechanisms. However, long-context modeling faces significant computational inefficiencies due to redundant attention computations: while attention weights are often sparse, all tokens consume equal computational resources. In this paper, we reformulate traditional probabilistic sequence modeling as a supervised learning task, enabling the separation of relevant and irrelevant tokens and providing a clearer understanding of redundancy. Based on this reformulation, we theoretically analyze attention sparsity, revealing that only a few tokens significantly contribute to predictions. Building on this, we formulate attention optimization as a linear coding problem and propose a group coding strategy, theoretically showing its ability to improve robustness against random noise and enhance learning efficiency. Motivated by this, we propose Dynamic Group Attention (DGA), which leverages the group coding to explicitly reduce redundancy by aggregating less important tokens during attention computation. Empirical results show that our DGA significantly reduces computational costs while maintaining competitive performance.

With the explosive growth of data volumes, enterprises face unprecedented challenges in data governance. As a critical step in data modeling, multi-source data table merging enables the integration of multiple tables containing the same entities from different business system databases into a standardized entity table, thereby enhancing data quality and consistency. However, this process relies heavily on manual judgment, leading to inefficiencies. In recent years, the performance of large language models (LLMs) in natural language processing has provided new insights for addressing data governance issues. This paper proposes a "mutation-generation-tuning" approach that enhances the performance of small parameter LLMs (e.g., Baichuan2-7B) in multi-source data table merging tasks through secondary pre-training and fine-tuning. The specific methods include data augmentation and generating intermediate reasoning processes using large parameter LLMs. Experimental results indicate that the optimized MDTM-GPT model exhibits performance on par with high-parameter models in multi-source data table merging tasks while maintaining low cost and data security. The contributions of this paper lie in presenting a novel approach for small parameter LLMs and providing practical validation for applications in the field of data governance.

With the rapid growth of data and the increasing awareness of privacy protection, data privacy issues have become particularly important in the field of machine learning. Federated learning, as a distributed learning method, achieves collaborative training of models while preserving data privacy by keeping the data stationary and allowing the model to move. However, during the federated learning process, there is still a risk of privacy leakage when aggregating the intermediate parameters of models trained by different data providers. Researchers have found that adding noise to the intermediate parameters of the model using differential privacy can effectively prevent privacy inference on the data contributors. Nevertheless, there exists an inherent trade-off between the accuracy and privacy in federated learning models under differential privacy. Strengthening privacy protection often leads to a decrease in model performance. This trade-off becomes more pronounced in complex deep learning models that require multiple iterations to converge. To address the issues of data privacy, data silos, and the trade-off between data privacy leakage and model availability in deep learning within federated learning, this paper proposes a relaxed differential privacy federated learning approach. It reduces the impact of noise on the final results by selectively perturbing gradients when data providers return intermediate model parameters. Experiments demonstrate that this approach achieves a high level of accuracy while preserving data privacy. Additionally, it exhibits superior performance in terms of computational efficiency, striking a well-balanced compromise between accuracy and privacy.

Previous works on font generation mainly focus on the standard print fonts where character's shape is stable and strokes are clearly separated. There is rare research on brush hand-writing font generation, which involves holistic structure changes and complex strokes transfer. To address this issue, we propose a novel GAN-based image translation model by integrating the skeleton information. We first extract the skeleton from training images, then design an image encoder and a skeleton encoder to extract corresponding features. A self-attentive refined attention module is devised to guide the model to learn distinctive features between different domains. A skeleton discriminator is involved to first synthesize the skeleton image from the generated image with a pre-trained generator, then to judge its realness to the target one. We also contribute a large-scale brush handwriting font image dataset with six styles and 15,000 high-resolution images. Both quantitative and qualitative experimental results demonstrate the competitiveness of our proposed model.

In this paper, we present a novel system (denoted as Polaca) to generate poetic Chinese landscape painting with calligraphy. Unlike previous single image-to-image painting generation, Polaca takes the classic poetry as input and outputs the artistic landscape painting image with the corresponding calligraphy. It is equipped with three different modules to complete the whole piece of landscape painting artwork: the first one is a text-to-image module to generate landscape painting image, the second one is an image-to-image module to generate stylistic calligraphy image, and the third one is an image fusion module to fuse the two images into a whole piece of aesthetic artwork.

As an essential means of privacy protection technology, zero-knowledge proof has gradually been applied into various fields with the development of blockchain technology, such as the Internet of Vehicles and Bitcoin. Personal comprehensive credit score as a measure to promote social governance is closely related to personal privacy. Although there are currently credit score calculation systems for various application scenarios, these systems almost ignore user privacy protection, which leads to user information leakage or abuse. The combination of zero-knowledge proof and personal credit score calculation has been studied by a number of researchers at the present stage. However, all data are provided by users directly in the current schemes, which did not consider the data authenticity under the situation that users provided false data. In this paper, we utilize zero-knowledge proof to design a novel privacy protection scheme for personal credit score calculation, taking into account the authenticity verification of multi-dimensional user data. In addition, our scheme also proposes the concept of a universal verification platform based on blockchain for personal credit scores. This platform has more substantial applicability and versatility for any qualified institution that requires querying and verifying user’s credit scores. At the end of the paper, we conducted a security analysis and performance evaluation for the overall scheme.

Emotion plays a critical role in calligraphy composition, which makes the calligraphy artwork impressive and have a soul. However, previous research on calligraphy generation all neglected the emotion as a major contributor to the artistry of calligraphy. Such defects prevent them from generating aesthetic, stylistic, and diverse calligraphy artworks, but only static handwriting font library instead. To address this problem, we propose a novel cross-modal approach to generate stylistic and diverse Chinese calligraphy artwork driven by different emotions automatically. We firstly detect the emotions in the text by a classifier, then generate the emotional Chinese character images via a novel modified Generative Adversarial Network (GAN) structure, finally we predict the layout for all character images with a recurrent neural network. We also collect a large-scale stylistic Chinese calligraphy image dataset with rich emotions. Experimental results demonstrate that our model outperforms all baseline image translation models significantly for different emotional styles in terms of content accuracy and style discrepancy. Besides, our layout algorithm can also learn the patterns and habits of calligrapher, and makes the generated calligraphy more artistic. To the best of our knowledge, we are the first to work on emotion-driven discourse-level Chinese calligraphy artwork composition.

The transaction efficiency of blockchains has been a painful point haunting researchers all the time. The metrics of Transaction per Second (TPS) present a low value, especially as the number of network nodes increases. In order to improve the blockchain's TPS, numerous articles explored the solutions from multiple aspects, such as: using a new block system based on the Directed Acyclic Graph (DAG), adjusting the block structure, optimizing the consensus algorithm or adopting a sharding protocol. The sharding protocols play an extremely important role among these schemes, while the existing ones suffer from secure risks due to their static sharding. In this paper, we proposed a dynamic sharding protocol for consortium blockchains, employing random values to randomly partition the network. Our protocol aims for improving the throughput of the blockchain system as well as ensuring its security. The OMNET++ simulation results showed that, in the case of a large-scale consortium blockchain network of 400 nodes which is divided into 16 shards, the performance of our scheme is able to attain up to 50 times promotion, compared to the system without sharding.

Real-world integrated personalized recommendation systems usually deal with millions of heterogeneous items. It is extremely challenging to conduct full corpus retrieval with complicated models due to the tremendous computation costs. Hence, most large-scale recommendation systems consist of two modules: a multi-channel matching module to efficiently retrieve a small subset of candidates, and a ranking module for precise personalized recommendation. However, multi-channel matching usually suffers from cold-start problems when adding new channels or new data sources. To solve this issue, we propose a novel Internal and contextual attention network (ICAN), which highlights channel-specific contextual information and feature field interactions between multiple channels. In experiments, we conduct both offline and online evaluations with case studies on a real-world integrated recommendation system. The significant improvements confirm the effectiveness and robustness of ICAN, especially for cold-start channels. Currently, ICAN has been deployed on WeChat Top Stories used by millions of users. The source code can be obtained from https://github.com/zhijieqiu/ICAN.

We present a novel Chinese calligraphy artwork composition system (MaLiang) which can generate aesthetic, stylistic and diverse calligraphy images based on the emotion status from the input text. Different from previous research, it's the first work to endow the calligraphy synthesis with the ability to express fickle emotions and composite a whole piece of discourse-level calligraphy artwork instead of single character images. The system consists of three modules: emotion detection, character image generation, and layout prediction. As a creative form of interactive art, MaLiang has been exhibited in several famous international art festivals.

Sequential recommendation and group recommendation are two important branches in the field of recommender system. While considerable efforts have been devoted to these two branches in an independent way, we combine them by proposing the novel sequential group recommendation problem which enables modeling group dynamic representations and is crucial for achieving better group recommendation performance. The major challenge of the problem is how to effectively learn dynamic group representations based on the sequential user-item interactions of group members in the past time frames. To address this, we devise a Group-aware Long- and Short-term Graph Representation Learning approach, namely GLS-GRL, for sequential group recommendation. Specifically, for a target group, we construct a group-aware long-term graph to capture user-item interactions and item-item co-occurrence in the whole history, and a group-aware short-term graph to contain the same information regarding only the current time frame. Based on the graphs, GLS-GRL performs graph representation learning to obtain long-term and short-term user representations, and further adaptively fuse them to gain integrated user representations. Finally, group representations are obtained by a constrained user-interacted attention mechanism which encodes the correlations between group members. Comprehensive experiments demonstrate that GLS-GRL achieves better performance than several strong alternatives coming from sequential recommendation and group recommendation methods, validating the effectiveness of the core components in GLS-GRL.

We present a novel real-time, collaborative, and interactive AI painting system, Mappa Mundi, for artistic Mind Map creation. The system consists of a voice-based input interface, an automatic topic expansion module, and an image projection module. The key innovation is to inject Artificial Imagination into painting creation by considering lexical and phonological similarities of language, learning and inheriting artist’s original painting style, and applying the principles of Dadaism and impossibility of improvisation. Our system indicates that AI and artist can collaborate seamlessly to create imaginative artistic painting and Mappa Mundi has been applied in art exhibition in UCCA, Beijing.

DC-link capacitors are usually utilized in voltage source converters to stabilize the DC-link voltage and suppress the transmission of current harmonics into the power grid. To guarantee the reliability of voltage source converters, in addition to select the DC-link capacitors with optimum capacitance, appropriate rated voltage and enough ripple current capacity, thermal design is also of great importance. In this paper, the DC-link capacitors heating in voltage source inverters is evaluated. A novel cooling structure of DC-link capacitor is designed. The heating dissipation capability of the novel structure capacitor increases more than 60% compared to the traditional one, which is a key step to increase the power density of voltage source inverter.

Silicon carbide (SiC) devices have its unique advantages of high operating temperature, high voltage capability with low switching losses when compared to its silicon counterparts. This paper presents a systematic method for long-term reliability analysis for SiC devices. The reliability block diagrams (RBDs) have been established for a SiC power module with standard wire-bonded package. The reliability function and mean time to failure (MTTF) for this module has been estimated by using the lifetime models for single bond wire and SiC chip. The analysis shows that the bond wires are the weakest line of the whole device's reliability. Therefore, it is obvious that utilizing to the full potential of SiC devices is critical for increasing the reliability of package. Design for novel package structure (e.g. without bond wires) may be a possible solution.