Lei Wang

Cognitive diagnosis (CD) aims to infer students' latent proficiencies from their exercise responses. While most existing approaches are optimized primarily for response prediction, such prediction-oriented training can yield less reliable proficiency. Moreover, representing proficiency as independent concept-wise variables overlooks the relations among concepts, often resulting in fragmented and incoherent diagnoses. To address these issues, we represent each student as an explicit concept tree with node-wise proficiencies, enabling coarse-to-fine diagnosis with explicit concept relations. In this work, we propose LLM-Enhanced Hierarchical Cognitive Diagnosis (HiLLM-CD), a tree-structured encoder-decoder framework that uses the concept tree as an interpretable proficiency representation. Specifically, HiLLM-CD first encodes a student's response history into node-wise proficiencies on the concept tree, and then decodes the tree-structured proficiencies to predict responses via a GCN-based cognitive decoder. Moreover, HiLLM-CD employs a multi-agent LLM pipeline to generate exercise-concept links and automatically construct the concept tree from educational text, substantially reducing reliance on expert annotations and making the overall diagnosis pipeline automated; it further uses an LLM teacher to guide the model to learn more accurate proficiencies. We evaluate HiLLM-CD on MOOCCubeX across four subject domains (Computer Science, Mathematics, Physics, and Biology). In fully unannotated settings, HiLLM-CD improves response prediction by 11.1% AUC on average and improves proficiency reliability by 0.597 in Kendall's tau. We further validate its utility on computerized adaptive testing, where HiLLM-CD improves AUC by 11.4% on average. Our code is available at: https://github.com/xyq20011011/HiLLM.

A fundamental challenge in Continual Learning (CL) is catastrophic forgetting, where adapting to new tasks degrades the performance on previous ones. While the field has evolved with diverse methods, this rapid surge in diverse methodologies has culminated in a fragmented research landscape. The lack of a unified framework, including inconsistent implementations, conflicting dependencies, and varying evaluation protocols, makes fair comparison and reproducible research increasingly difficult. To address this challenge, we propose LibContinual, a comprehensive and reproducible library designed to serve as a foundational platform for realistic CL. Built upon a high-cohesion, low-coupling modular architecture, LibContinual integrates 19 representative algorithms across five major methodological categories, providing a standardized execution environment. Meanwhile, leveraging this unified framework, we systematically identify and investigate three implicit assumptions prevalent in mainstream evaluation: (1) offline data accessibility, (2) unregulated memory resources, and (3) intra-task semantic homogeneity. We argue that these assumptions often overestimate the real-world applicability of CL methods. Through our comprehensive analysis using strict online CL settings, a novel unified memory budget protocol, and a proposed category-randomized setting, we reveal significant performance drops in many representative CL methods when subjected to these real-world constraints. Our study underscores the necessity of resource-aware and semantically robust CL strategies, and offers LibContinual as a foundational toolkit for future research in realistic continual learning. The source code is available from \href{https://github.com/RL-VIG/LibContinual}{https://github.com/RL-VIG/LibContinual}.

Few-shot unsupervised domain adaptation (FS-UDA) leverages a limited amount of labeled data from a source domain to enable accurate classification in an unlabeled target domain. Despite recent advancements, current approaches of FS-UDA continue to confront a major challenge: models often demonstrate instability when adapted to new FS-UDA tasks and necessitate considerable time investment. To address these challenges, we put forward a novel framework called Enduring and Efficient Meta-Prompt Learning (E²MPL) for FS-UDA. Within this framework, we utilize the pre-trained CLIP model as the backbone of feature learning. Firstly, we design domain-shared prompts, consisting of virtual tokens, which primarily capture meta-knowledge from a wide range of meta-tasks to mitigate the domain gaps. Secondly, we develop a task prompt learning network that adaptively learns task-specific prompts with the goal of achieving fast and stable task generalization. Thirdly, we formulate the meta-prompt learning process as a bilevel optimization problem, consisting of (outer) meta-prompt learner and (inner) task-specific classifier and domain adapter. Also, the inner objective of each meta-task has the closed-form solution, which enables efficient prompt learning and adaptation to new tasks in a single step. Extensive experimental studies demonstrate the promising performance of our framework in a domain adaptation benchmark dataset DomainNet. Compared with state-of-the-art methods, our approach has improved the average accuracy by at least 15 percentage points and reduces the average time by 64.67% in the 5-way 1-shot task; in the 5-way 5-shot task, it achieves at least a 9-percentage-point improvement in average accuracy and reduces the average time by 63.18%. Moreover, our method exhibits more enduring and stable performance than the other methods, i.e., reducing the average IQR value by over 40.80% and 25.35% in the 5-way 1-shot and 5-shot task, respectively.