This study addresses the generalization limitations commonly observed in large language models under multi-task and cross-domain settings. Unlike prior methods such as SPoT, which depends on fixed prompt templates, our study introduces a unified multi-task learning framework with dynamic prompt scheduling mechanism. By introducing a prompt pool and a task-aware scheduling strategy, the method dynamically combines and aligns prompts for different tasks. This enhances the model’s ability to capture semantic differences across tasks. During prompt fusion, the model uses task embeddings and a gating mechanism to finely control the prompt signals. This ensures alignment between prompt content and task-specific demands. At the same time, it builds flexible sharing pathways across tasks. In addition, the proposed optimization objective centers on joint multi-task learning. It incorporates an automatic learning strategy for scheduling weights, which effectively mitigates task interference and negative transfer. To evaluate the effectiveness of the method, a series of sensitivity experiments were conducted. These experiments examined the impact of prompt temperature parameters and task number variation. The results confirm the advantages of the proposed mechanism in maintaining model stability and enhancing transferability. Experimental findings show that the prompt scheduling method significantly improves performance on a range of language understanding and knowledge reasoning tasks. These results fully demonstrate its applicability and effectiveness in unified multi-task modeling and cross-domain adaptation.

This paper proposes an enhanced feature pyramid-based segmentation algorithm designed for accurate skin disease lesion segmentation. The method utilizes a multi-level convolutional encoder to extract hierarchical features and introduces a cross-scale enhancement module to strengthen the fusion of multi-scale contextual information. Additionally, a space attention mechanism is applied to refine spatial localization and highlight critical lesion regions. The network architecture is systematically constructed to balance the extraction of fine-grained details and global semantic representations. Extensive experiments on the ISIC 2018 dataset demonstrate that the proposed method achieves superior performance compared to several baseline models, showing notable improvements in mIOU, mDice, and F1-Score metrics. Intuitive visualization of segmentation masks and feature maps further validates the model's ability to accurately capture lesion boundaries and suppress background noise. The effectiveness of the enhanced feature pyramid structure confirms its potential to advance the precision and robustness of skin disease image analysis. Through comprehensive evaluations, the method is shown to offer a reliable and scalable solution for skin lesion segmentation tasks.

This paper proposes a heterogeneous federated learning method for collaborative modeling, addressing data heterogeneity and privacy protection in multi-center scenarios. Based on the traditional federated learning framework, the method introduces a regularization term to mitigate performance instability caused by inconsistent data distributions across clients. It also designs a weighted aggregation strategy based on dynamic client features to enhance model convergence and robustness in multi-source heterogeneous environments. To ensure data privacy, the method applies differential privacy during local training. Noise is added to the model updates to prevent leakage of sensitive information. During development, the method is evaluated using a real federated learning dataset. It is compared with several mainstream approaches to validate its advantages in key performance metrics such as Accuracy, AUC, and F1-Score. Experimental results show that the proposed method effectively balances the personalized needs of local models with the goal of global consistency. It adapts to differences across clients and achieves better federated modeling performance. The overall architecture demonstrates strong practicality, security, and generalization ability. It is well-suited for real-world applications where centralized data storage is not feasible.

This paper addresses the challenges of high computational cost and slow inference in deploying large language models. It proposes a distillation strategy guided by multiple teacher models. The method constructs several teacher models and integrates their output probability distributions and intermediate semantic features. This guides the student model to learn from multiple sources of knowledge. As a result, the student model gains stronger language understanding and generation ability while maintaining a small parameter size. To achieve this, the paper introduces a weighted output fusion mechanism, a feature alignment loss function, and an entropy-driven dynamic teacher weighting strategy. These components improve the quality and stability of knowledge transfer during distillation. Under multi-teacher guidance, the student model captures semantic information more effectively and demonstrates strong performance across multiple evaluation metrics. In particular, the method shows high consistency in expression, generalization ability, and task adaptability in tasks such as language modeling, text generation, and multi-task learning. The experiments compare the proposed method with several widely adopted distillation approaches. The results further confirm its overall advantages in perplexity, distillation loss, and generation quality. This study provides a feasible technical path for the efficient compression of large-scale language models. It also demonstrates the effectiveness of multi-teacher collaborative mechanisms in complex language modeling tasks.

This paper proposes a federated learning-based method for multi-party collaborative mining and heterogeneous data source fusion. The goal is to address the shortcomings of traditional centralized learning in data privacy protection and cross-domain collaboration. To tackle the heterogeneity problem among multiple distributed data sources, the proposed method uses the federated learning framework to achieve collaborative training across different data sources. This approach avoids direct data transmission, effectively protecting data privacy. By transmitting model parameters to a central server for aggregation, the method enables efficient model fusion and collaborative learning across multiple data sources, significantly improving learning performance in non-i.i.d. data environments. Experimental results show that the proposed model demonstrates high accuracy and stability when handling heterogeneous data sources in multi-party collaborative learning. It outperforms other existing federated learning methods, such as FedAvg and SCAFFOLD. The model is able to ensure data privacy while improving data fusion efficiency and model generalization. This approach holds promising application potential. This study provides an innovative solution to the challenges of multi-party collaborative mining and data privacy protection, particularly in fields such as healthcare, financial risk control, and intelligent manufacturing, where data security is paramount. It offers technological support for real-world data sharing and cross-domain collaboration.