Diagnosing aesthetic perception plays a crucial role in deepening our understanding of student creativity, emotional expression, and the pursuit of lifelong learning within art education. This task encompasses the evaluation and analysis of students' sensitivity, preference, and capacity to perceive and appreciate beauty across different sensory domains. Currently, this assessment frequently relies on subjective evaluations of student artworks. There are two limitations: 1) the diagnosis is possibly limited by instructors' bias; and 2) the heavy workload of instructors for conducting comprehensive assessments. These limitations motivate us to ask: Can we automatically and objectively conduct aesthetic perception diagnosis? To this end, we propose an innovative deep hybrid framework, AestheNet, to automatically evaluate aesthetic perception by analyzing numerous collected student paitings. More specially, we first utilize Convolutional Neural Networks (CNNs) to extract the significant features from the student artworks. Then, we employ the Transformer model to capture the intricate relationships among multiple aesthetic perception dimensions for objective diagnosis. Finally, we validate the effectiveness of the framework by creating a new dataset consisting of 2,153 paintings drawn by 675 students. These paintings are annotated by human experts from 77 dimensions based on domain expertise. Extensive experiments have shown the effectiveness of AestheNet in aesthetic perception diagnosis. AestheNet is dedicated to to overcoming the subjectivity inherent in traditional assessment methods, providing a new, quantifiable, and standardized way to evaluate aesthetic perception. This research not only opens up new perspectives in understanding students' aesthetic development during the art education process but also explores the innovation of using artificial intelligence (AI) technologies in the assessment of art education.

Aesthetic perception, the cognitive process through which individuals interpret and evaluate the expressive and emotional qualities of visual art, is fundamental to students' creative and emotional development. Recent progress in artificial intelligence has enabled computational models to assist in aesthetic analysis by identifying patterns in visual composition and affective expression. However, such models often struggle to recognize abstract or context-dependent aesthetic dimensions, and improving these aspects through comprehensive annotation remains costly and time-consuming. This study presents a human-AI teaming framework designed to identify the aesthetic perception dimensions that AI models find most difficult to interpret and to allocate these dimensions to human experts for annotation. The framework employs a multi-agent reinforcement learning (MARL) mechanism, where each agent is assigned to a specific aesthetic dimension and learns a policy for determining whether expert annotation is required. Two complementary state representation strategies are introduced: a statistical representation that captures the model's predictive distribution across dimensions, and a graph-based attention module that models interdependencies among aesthetic attributes. A reward mechanism further guides agents to balance the improvement of model perception with the minimization of human annotation effort. Experiments conducted on two real-world datasets demonstrate that the proposed framework effectively identifies the challenging dimensions for AI models and strategically delegates them for human evaluation. This targeted collaboration significantly enhances annotation efficiency and model interpretability, providing a scalable approach for improving human-AI synergy in aesthetic perception analysis.

Personalized exercise recommendation is essential for enhancing learning efficiency by adapting educational content to individual student needs. However, current approaches often fail to simultaneously minimize the number of exercises while ensuring students achieve specific learning objectives, and they overlook the nuanced knowledge structures revealed during problem-solving. To address these limitations, this paper introduces a novel framework for dynamic requirement-driven exercise recommendation (DRER), which combines confusion-aware knowledge tracing with nonlinear combinatorial optimization. DRER consists of two key stages. First, in the knowledge tracing stage, a hybrid knowledge concept matrix is constructed to model both the inherent relationships among knowledge concepts and the latent structures revealed through student interactions. Second, in the exercise recommendation stage, the problem is formulated as a Minimum nonlinear weighted set cover problem (MNLWSCP), aiming to identify the smallest set of exercises that enables students to reach a predefined proficiency threshold (e.g., 0.6). To solve this efficiently, a heuristic-based local search algorithm is proposed. Extensive experiments on real-world and academic datasets validate the effectiveness of the framework, demonstrating its ability to significantly reduce the number of recommended exercises while ensuring high accuracy in achieving learning objectives. This work represents a significant integration of data mining and combinatorial optimization, offering a scalable and practical solution for personalized education.

The prediction of drug-target interactions (DTIs) has emerged as a vital step in drug discovery. Recently, biomedical knowledge graph enables the utilization of multi-omics resources for modelling complex biological systems and further improves overall performance of specific predictive task. However, due to the scale and generalization of biomedical knowledge graph, it is necessary to capture task-specific knowledge from biomedical knowledge graph for DTI prediction. Moreover, although biomedical knowledge graph has rich interactions between biological entities, there still needs to contain unignorable structural information of drugs or targets in the multi-modal fusion manner. To this end, we develop a novel DTI identification framework, CBKG-DTI, which aims to distill task-specific knowledge from the complex knowledge graph to the lightweight DTI prediction model. Specifically, CBKG-DTI first introduces a hierarchy-aware knowledge graph embedding as teacher model to capture semantic hierarchy information of biomedical knowledge graph. Then, to further improve model performance, CBKG-DTI integrates information from multiple aspects such as relational information and structural information by constructing a heterogeneous network and then employs a heterogeneous graph attention network framework as the lightweight student model. Moreover, we design a multi-level distillation mechanism to improve the representation and prediction ability of the lightweight student model via capturing the representation and logit distribution of the teacher model. Finally, we conduct the extensive comparison experiments and can reach the AUC of 0.9751 and the AUPR of 0.6310 under 5-fold cross validation. This not only demonstrates the superiority of CBKG-DTI in DTI prediction, but also, more importantly, validate the effectiveness of the framework capturing task-specific knowledge from biomedical knowledge graph.