Vision-Language-Action (VLA) models have advanced in robotic manipulation, yet practical deployment remains hindered by two key limitations: **1) perceptual redundancy**, where irrelevant visual inputs are processed inefficiently, and **2) superficial instruction-vision alignment**, which hampers semantic grounding of actions. In this paper, we propose **SemanticVLA**, a novel VLA framework that performs Semantic-Aligned Sparsification and Enhancement for Efficient Robotic Manipulation. Specifically: **1)** To sparsify redundant perception while preserving semantic alignment, **Semantic-guided Dual Visual Pruner (SD-Pruner)** performs: Instruction-driven Pruner (ID-Pruner) extracts global action cues and local semantic anchors in SigLIP; Spatial-aggregation Pruner (SA-Pruner) compacts geometry-rich features into task-adaptive tokens in DINOv2. **2)** To exploit sparsified features and integrate semantics with spatial geometry, **Semantic-complementary Hierarchical Fuser (SH-Fuser)** fuses dense patches and sparse tokens across SigLIP and DINOv2 for coherent representation. **3)** To enhance the transformation from perception to action, **Semantic-conditioned Action Coupler (SA-Coupler)** replaces the conventional observation-to-DoF approach, yielding more efficient and interpretable behavior modeling for manipulation tasks. Extensive experiments on simulation and real-world tasks show that SemanticVLA sets a new SOTA in both performance and efficiency. SemanticVLA surpasses OpenVLA on LIBERO benchmark by **21.1%** in success rate, while reducing training cost and inference latency by **3.0×** and **2.7×**.

Computer-generated aesthetic patterns are widely used as design materials in various fields. The most common methods use fractals or dynamical systems as basic tools to create various patterns. To enhance aesthetics and controllability, some researchers have introduced symmetric layouts along with these tools. One popular strategy employs dynamical systems compatible with symmetries that construct functions with the desired symmetries. However, these are typically confined to simple planar symmetries. The other generates symmetrical patterns under the constraints of tilings. Although it is slightly more flexible, it is restricted to small ranges of tilings and lacks textural variations. Thus, we proposed a new approach for generating aesthetic patterns by symmetrizing quasi-regular patterns using general k-uniform tilings. We adopted a unified strategy to construct invariant mappings for k-uniform tilings that can eliminate texture seams across the tiling edges. Furthermore, we constructed three types of symmetries associated with the patterns: dihedral, rotational, and reflection symmetries. The proposed method can be easily implemented using GPU shaders and is highly efficient and suitable for complicated tiling with regular polygons. Experiments demonstrated the advantages of our method over state-of-the-art methods in terms of flexibility in controlling the generation of patterns with various parameters as well as the diversity of textures and styles.

In order to achieve electrical isolation, high-voltage transformers must use insulating materials. However, in some applications, such as medical treatment, transformers using insulating materials may cause problems of weight increase and flexible movement. As a result, the inductively coupled power transfer (ICPT) system are used to replace regular high transformers, with the air gap between the primary and secondary coils serving as electrical insulation. Under the SP resonance compensation structure, the ICPT system needs to achieve exceptionally high gain, which will result in the voltage gain being sensitive to frequency variations. When the frequency deviates slightly from the resonance point, more circulation and loss occurs, lowering the efficiency of the system. Therefore, this paper proposes a control strategy based on the combination of pulse-duration modulation (PDM) and maximum efficiency point tracking (MEPT) is proposed, at the first resonance point of the ICPT system based on the SP resonance compensation structure. This control method is capable of producing medium and high voltage constant voltage output, and the switch tube achieves zero current switching during operation. In the case of a wide load variation range, maximum efficiency point tracking may be obtained at the same time. Finally, a 100W high-gain ICPT experimental platform is built, and the constant voltage output with 25 times high gain is realized over the transmission distance of 11.2cm, which verifies the effectiveness of the control strategy.

With the rapid development of the Internet and global finance, more and more financial data is accumulated, while a majority of them are unstructured. Within the unstructured financial data, sentiment information is the most attractive and commonly desired to fetch for the financial participants, because sentiment can be an indicator to reflect the state of the financial market, which means the market can be predicted to some extent through sentiment analysis of the financial text on the Internet. Although sentiment analysis has been studied at the document level, sentence level, and aspect level, it may not provide enough information on what the financial participants want to know. Therefore, this paper aims at entity-level sentiment analysis of financial text based on the pre-trained language model and the improvement of granularity of the financial information. Bidirectional Encoder Representations from Transformers is adopted to illustrate the efficacy of pre-trained language model-based financial sentiment analysis. Experiments are conducted with a labeled test set and a train set with 8,000 and 10,000 pieces of financial network data, respectively. The experiment results show that at least 95% accuracy can be achieved for determining negative information, while at least 93% accuracy can be achieved for fetching financial entity that the negative information correlates to.