This paper addresses key challenges in knowledge injection for large language models, including static representation, difficulty in updating, and limited domain adaptability. It proposes a dynamic fine-tuning framework designed for knowledge injection. The framework is based on parameter-efficient tuning strategies and introduces learnable adapter modules and gating mechanisms. These components enable selective integration and dynamic control of both structured and unstructured external knowledge. In the model design, a query encoder extracts semantic vectors from the input text. These vectors are matched with an external knowledge base to construct a dynamic knowledge subset. This subset guides task generation. Adapter modules and gating units are then applied across model layers to adjust knowledge enhancement. This ensures that external knowledge contributes to reasoning while preserving the model's original language capabilities. A unified joint loss function is also introduced. It coordinates the optimization between language modeling and knowledge alignment objectives. To evaluate the proposed method, the WikiHop dataset for multi-hop question answering is used. The model's behavior is analyzed under various experimental settings, including different parameter update ratios, knowledge densities, and cross-domain transfer scenarios. The results show that the method achieves strong performance on key metrics such as knowledge recall, F1 score, and inference efficiency, even with low parameter update ratios. This demonstrates the practicality and stability of the approach in dynamic knowledge integration tasks. The proposed work offers a flexible and efficient technical path for knowledge injection and domain adaptation in large language models.

Ad ranking systems must simultaneously optimize multiple objectives including click-through rate (CTR), conversion rate (CVR), revenue, and user experience metrics. However, production systems face critical challenges: score scale inconsistency across traffic segments undermines threshold transferability, and position bias in click logs causes offline-online metric discrepancies. We propose CaliCausalRank, a unified framework that integrates training-time scale calibration, constraint-based multi-objective optimization, and robust counterfactual utility estimation. Our approach treats score calibration as a first-class training objective rather than post-hoc processing, employs Lagrangian relaxation for constraint satisfaction, and utilizes variance-reduced counterfactual estimators for reliable offline evaluation. Experiments on the Criteo and Avazu datasets demonstrate that CaliCausalRank achieves 1.1% relative AUC improvement, 31.6% calibration error reduction, and 3.2% utility gain compared to the best baseline (PairRank) while maintaining consistent performance across different traffic segments.

This work addresses correlation bias and causal effect confounding in advertising recommendation systems and presents a causal learning–based recommendation framework. We first examine the limitations of conventional recommendation algorithms in complex advertising environments, where confounding variables and exposure bias often prevent models from capturing users’ true preferences. To tackle these issues, we design a unified embedding architecture that jointly represents user, advertisement, and contextual features, and incorporates a structural causal graph to explicitly model dependencies among variables. During model training, causal consistency regularization and inverse propensity weighting are integrated to mitigate the impact of biased exposure mechanisms and non-uniform sampling. A joint optimization objective is further formulated to couple click-through rate prediction with causal consistency estimation, enabling robust causal effect learning without sacrificing predictive accuracy. Extensive experiments on large-scale advertising datasets demonstrate that the proposed approach consistently outperforms several representative baselines in terms of Precision@10, Recall@10, NDCG@10, and MAP, while exhibiting strong robustness under multi-dimensional sensitivity analysis. Overall, this study highlights the practical value of causal modeling and consistency-aware learning in advertising recommendation and offers a computationally grounded approach for improving both interpretability and fairness in recommendation systems.

This study addresses the challenges of semantic mixing, limited interpretability, and complex feature structures in fine-grained sentiment and opinion classification by proposing an interpretable feature disentanglement framework built on the latent space of large language models. The framework constructs multi-component latent representations that separate emotional polarity, opinion direction, target attributes, and pragmatic cues during encoding, thus overcoming the limitations of traditional methods that merge diverse semantic factors into a single representation. During representation learning, the model first uses a large model encoder to generate basic semantic features and then builds multiple independent subspaces through learnable projections. A covariance constraint is introduced to reduce coupling across semantic components and to create clear boundaries in the latent space. To preserve the essential information of the original text, a reconstruction consistency mechanism integrates features from all subspaces to rebuild the global representation and enhance semantic completeness. The framework also incorporates semantic anchors to align latent components with interpretable semantic dimensions, giving each subspace a clear emotional or opinion-related meaning and improving transparency at the mechanism level. Experimental results show that the framework outperforms existing methods across multiple metrics and handles complex syntax, implicit semantics, and coexisting emotions with greater stability. It achieves high accuracy and interpretability in fine-grained sentiment and opinion analysis. Overall, the proposed disentanglement framework provides an effective approach for building structured, multidimensional, and interpretable representations of textual emotions and opinions and holds significant value for complex semantic understanding tasks.

This paper investigates the application of contrastive learning-based user and item representation learning in recommendation systems. A recommendation model combining contrastive loss with data augmentation strategies is proposed. Traditional recommendation systems typically rely on explicit or implicit feedback to model the relationships between users and items. However, traditional methods often show limited performance when dealing with issues such as data sparsity and cold start. To address this, the paper introduces a contrastive learning framework. By constructing positive and negative sample pairs, the model is guided to learn more discriminative representations. Various data augmentation methods are also applied to enhance the robustness and generalization capability of representation learning. Specifically, the paper compares different data augmentation strategies, including subsampling views, feature masking, and behavioral perturbation, and analyzes their performance under different temperature parameters and sparsity conditions. The experimental results show that the contrastive learning-based model effectively improves recommendation accuracy, particularly in addressing sparse data and cold start problems. Additionally, the paper explores the performance differences of the contrastive loss function under different training settings, validating the significant impact of appropriate hyperparameter tuning on recommendation system performance. By combining contrastive learning with data augmentation, this study provides a new perspective and significantly enhances the performance of recommendation systems in complex scenarios. The research offers both theoretical support and practical guidance for the future development of personalized recommendation technologies.