https://repositori.mypolycc.edu.my/jspui/handle/123456789/6661 2026-06-01T18:08:52Z https://repositori.mypolycc.edu.my/jspui/handle/123456789/9932 Title: MULTIMODAL ALIGNMENT AND HIERARCHICAL FUSION NETWORK FOR MULTIMODAL SENTIMENT ANALYSIS Authors: Huang, Jiasheng; Li, Huan; Mo, Xinyue Abstract: The widespread emergence of multimodal data on social platforms has presented new opportunities for sentiment analysis. However, previous studies have often overlooked the issue of detail loss during modal interaction fusion. They also exhibit limitations in addressing semantic alignment challenges and the sensitivity of modalities to noise. To enhance analytical accuracy, a novel model named MAHFNet is proposed. The proposed architecture is composed of three main components. Firstly, an attention-guided gated interaction alignment module is developed for modeling the semantic interaction between text and image using a gated network and a cross-modal attention mechanism. Next, a contrastive learning mechanism is introduced to encourage the aggregation of semantically aligned image-text pairs. Subsequently, an intra-modality emotion extraction module is designed to extract local emotional features within each modality. This module serves to compensate for detail loss during interaction fusion. The intra-modal local emotion features and cross-modal interaction features are then fed into a hierarchical gated fusion module, where the local features are fused through a cross-gated mechanism to dynamically adjust the contribution of each modality while suppressing modality-specific noise. Then, the fusion results and cross-modal interaction features are further fused using a multi-scale attention gating module to capture hierarchical dependencies between local and global emotional information, thereby enhancing the model’s ability to perceive and integrate emotional cues across multiple semantic levels. Finally, extensive experiments have been conducted on three public multimodal sentiment datasets, with results demonstrating that the proposed model outperforms existing methods across multiple evaluation metrics. Specifically, on the TumEmo dataset, our model achieves improvements of 2.55% in ACC and 2.63% in F1 score compared to the second-best method. On the HFM dataset, these gains reach 0.56% in ACC and 0.9% in F1 score, respectively. On the MVSA-S dataset, these gains reach 0.03% in ACC and 1.26% in F1 score. These findings collectively validate the overall effectiveness of the proposed model. 2025-09-26T00:00:00Z https://repositori.mypolycc.edu.my/jspui/handle/123456789/9931 Title: STRENGTHENING SMALL OBJECT DETECTION IN ADAPTED RT-DETR THROUGH ROBUST ENHANCEMENTS Authors: Madan, Manav; Reich, Christoph Abstract: RT-DETR (Real-Time DEtection TRansformer) has recently emerged as a promising model for object detection in images, yet its performance on small objects remains limited, particularly in terms of robustness. While various approaches have been explored, developing effective solutions for reliable small object detection remains a significant challenge. This paper introduces an adapted variant of RT-DETR, specifically designed to enhance robustness in small object detection. The model was first designed on one dataset and subsequently transferred to others to validate generalization. Key contributions include replacing components of the feed-forward neural network (FFNN) within a hybrid encoder with Hebbian, randomized, and Oja-inspired layers; introducing a modified loss function; and applying multi-scale feature fusion with fuzzy attention to refine encoder representations. The proposed model is evaluated on the Al-Cast Detection X-ray dataset, which contains small components from high-pressure die-casting machines, and the PCB quality inspection dataset, which features tiny hole anomalies. The results show that the optimized model achieves an mAP of 0.513 for small objects—an improvement from the 0.389 of the baseline RT-DETR model on the Al-Cast dataset—confirming its effectiveness. In addition, this paper contributes a mini-literature review of recent RT-DETR enhancements, situating our work within current research trends and providing context for future development. 2025-09-27T00:00:00Z https://repositori.mypolycc.edu.my/jspui/handle/123456789/9930 Title: DOCUMENT-LEVEL FUTURE EVENT PREDICTION INTEGRATING EVENT KNOWLEDGE GRAPH AND LLM TEMPORAL REASONING Authors: Huang, Shaonian; Wang, Huanran; Li, Peilin; Chen, Zhixin Abstract: Predicting future events is crucial for temporal reasoning, providing valuable insights for decision-making across diverse domains. However, the intricate global interactions and temporal–causal relationships at the document level event present significant challenges. This study introduces a novel document-level future event prediction method that integrates an event knowledge graph and a large language model (LLM) reasoning frame work based on metacognitive theory. Initially, an event knowledge graph is constructed by extracting event chains from the original document-level event texts. An LLM-based approach is then used to generate diverse and rational positive and negative training samples. Subsequently, a future event reasoning framework based on metacognitive theory is introduced. This framework enhances the model’s reasoning capabilities through a cyclic process of task understanding, reasoning strategy planning, strategy execution, and strategy reflection. Experimental results demonstrate that the proposed approach outperforms baseline models. Notably, the incorporation of the event knowledge graph significantly enhances the performance of different reasoning methods, while the proposed reasoning framework achieves superior performance in document-level future event pre diction tasks. Furthermore, the interpretability analysis of the prediction results validates the effectiveness of the proposed method. This study advances research on document-level future event prediction, highlighting the critical role of event knowledge graphs and large language models in temporal reasoning. It offers a more sophisticated future event prediction framework for government management departments, facilitating the enhancement of government safety management strategies. 2025-09-26T00:00:00Z https://repositori.mypolycc.edu.my/jspui/handle/123456789/9929 Title: CHANNEL TEMPERATURE MEASUREMENT OF GAN HEMT USED IN KILOWATT-LEVEL POWER AMPLIFIER Authors: Zhong, Sheng; Fang, Wenrao; Zhao, Juan; Huang, Wenhua; Fu, Chao; Wang, Lulu; He, Tianwei Abstract: This paper presents an electrical thermometry method designed for kilowatt(kW)-level Gallium Nitride (GaN) High Electron Mobility Transistors (HEMTs). The dependence of the drain current on the channel temperature in GaN HEMTs is utilized as a means to measure the transient channel temperature. However, in kW-class GaN HEMTs, the gate current can reach tens of milliamperes, and trap-induced capture resulting from high doping concentrations can both influence the drain current. Through modifications to the gate bias circuit, the gate voltage self-biasing phenomenon caused by the gate current is mitigated. A theoretical model is derived to express the relationship between the drain current and the channel temperature. Experimentally, amplifier modules equipped with kW-level HEMTs were placed on thermal stages set at 45 ◦C, 60 ◦C, and 80 ◦C. The transient drain current curves and the corresponding channel temperature profiles were measured. The measured drain current versus channel temperature curves at different ambient temperatures were fitted and compared with the theoretically derived formula. The relative error between the measured and calculated drain current values at the same channel temperature was found to be within 1% 2025-09-29T00:00:00Z