Cognitive diagnosis (CD) is a central assessment approach in intelligent education that aims to uncover learners’ knowledge mastery and latent cognitive abilities under examination conditions. However, existing models primarily infer learners’ knowledge state based on their final performance, thus failing to capture the reasoning trajectories during problem-solving and struggling to address the dissociation between ability and performance caused by short-term affective fluctuations. To address these challenges, we propose a generative agent for CD (Agent4CD) in intelligent education that enhances interpretability and provides greater personalization and adaptability. Specifically, the framework constructs learner profiles and incorporates a memory module to integrate historical responses. It leverages the reasoning capabilities of large language models (LLMs) to dynamically model cognitive paths and knowledge evolution. The affect perception module utilizes LLMs’ contextual understanding to identify students’ transient affective state, helping to mitigate misjudgments of ability caused by affective fluctuations. Meanwhile, a dual-reflection mechanism supports both real-time intervention and long-term ability development. Experiments demonstrate that Agent4CD outperforms existing methods in diagnostic accuracy, interpretability, and personalization. To facilitate further research, data and code are publicly available at https://github.com/xinjiesun-ustc/Agent4CD

Intelligent learning (iLearning) represents a holistic approach to e-learning that dynamically aligns instructional materials with individual learner needs. Within the iLearning framework, the cognitive diagnosis model (CDM) serves as a cornerstone technology by inferring human cognitive states from diagnostic test response logs. Traditional CDMs typically treat cognitive states as learnable parameters, optimizing them through a score prediction-driven paradigm. However, this approach falls short in the nonidentifiability problem of diagnostic outputs and lacks efficiency in instant diagnosis for incoming learners. These limitations pose a significant challenge to the individual fairness and utility of cognitive assessments. In this study, we introduce a novel generative diagnosis paradigm designed to resolve these limitations. By enabling the inductive inference of cognitive states without the need for parameter reoptimization, our generative approach ensures strict identifiability and addresses the low efficiency of optimization-based diagnosis. We present two streamlined yet powerful instantiations of this paradigm: generative item response theory and the generative neural cognitive diagnosis model. Extensive experiments on real-world datasets confirm that our methodology significantly outperforms traditional models, offering superior scalability and reliability. Ultimately, this generative framework provides a robust foundation for the development of next-generation iLearning environments.

Time series forecasting is commonly formulated as a model-centric and single-pass prediction task. However, real-world forecasting often requires task understanding, data diagnosis, contextual feature acquisition, tool-assisted modeling, reflective verification, and human feedback. In this paper, we demonstrate CastClaw, an interactive agent system for context-aware time series forecasting. CastClaw organizes forecasting as a structured runtime workflow that includes intent understanding, data profiling, iterative prediction, reflective verification, and traceable report generation. Supported by a tool library, an execution environment, and state management, CastClaw can invoke forecasting tools, compare models across different families and forecasting skills, track intermediate states, and incorporate user feedback. Through demonstrations on real-world forecasting scenarios, CastClaw shows how forecasting systems can move beyond static predictors toward interactive, evidence-grounded, and verifiable forecasting. Our code is available at (https://github.com/ustc-time-series/CastClaw). The demonstration video could be found at the link (https://ustc-time-series.github.io/cast-claw/).

Computerized Adaptive Testing (CAT) offers an efficient and personalized method for assessing examinee proficiency by dynamically adjusting test questions based on individual performance. Compared to traditional, non-personalized testing methods, CAT requires fewer questions and provides more accurate assessments. As a result, CAT has been widely adopted across various fields, including education, healthcare, sports, sociology, and the evaluation of AI models. While traditional methods rely on psychometrics and statistics, the increasing complexity of large-scale testing has spurred the integration of machine learning techniques. This paper aims to provide a machine learning-focused survey on CAT, presenting a fresh perspective on this adaptive testing paradigm. We delve into measurement models, question selection algorithm, bank construction, and test control within CAT, exploring how machine learning can optimize these components. Through an analysis of current methods, strengths, limitations, and challenges, we strive to develop robust, fair, and efficient CAT systems. By bridging psychometric-driven CAT research with machine learning, this survey advocates for a more inclusive and interdisciplinary approach to the future of adaptive testing.

Arena-based evaluation has become a key method for assessing large language models (LLMs) through head-to-head model comparisons, closely reflecting human preferences. However, current arena rating systems (e.g., ELO rating system) often suffer from inefficiencies due to exhaustive or random model pair annotations, leading to redundant evaluations, longer evaluation times, and lower overall efficiency. To address these challenges, we propose a novel adaptive model-pair selection algorithm. By leveraging the asymptotic normality of LLM ability estimation under sparse conditions, our approach strategically selects high-value model pairs, focusing on confrontations with the lowest variance. Specifically, we introduce Fisher information as a metric to guide model pair selection, optimizing the evaluation process through A-optimality and D-optimality. A-optimality minimizes estimation variance, ensuring balanced reliability across models, while D-optimality reduces uncertainty by maximizing the determinant of the Fisher Information Matrix. Extensive experiments on both simulated and real-world datasets demonstrate that our method outperforms existing approaches in terms of information efficiency and result reliability. Notably, our method offers a flexible, general toolkit that can be easily integrated into existing arena-based platforms, greatly improving scalability and efficiency for large-scale LLM evaluations.

Large-scale learner-task interaction data are crucial for intelligent educational systems but are costly to collect and constrained by privacy and learner engagement. Learner simulators play a critical role in simulating scalable learner behavior without the need for continuous involvement of real learners. However, existing methods are predominantly \textbf{individual-centric}, pairing a simulator with each learner to iteratively infer latent knowledge states from dense interaction histories, which is both data- and computation-intensive, and fragile in cold-start scenarios. We propose a \textbf{cohort-aware roll-call simulation paradigm} that first constructs cohort-level proficiency priors and refines individual learner states through a small number of targeted diagnostic queries. Based on this paradigm, we introduce \textbf{Edu-Theater}, an LLM-powered agent system that performs cohort-aware learner simulation via a teacher agent and retrospective roll-call probing over learner logs. Edu-Theater enables scalable future behavior simulation without the need for dense per-learner histories. Experiments on two real-world datasets demonstrate that Edu-Theater achieves higher simulation accuracy with significantly fewer LLM calls, producing synthetic data that enhances downstream applications such as adaptive testing.

Large Language Models (LLMs) are increasingly used for question answering over scientific research papers. Existing retrieval augmentation methods often rely on isolated text chunks or concepts, but overlook deeper semantic connections between papers. This impairs the LLM's comprehension of scientific literature, hindering the comprehensiveness and specificity of its responses. To address this, we propose Central Entity-Guided Graph Optimization for Community Detection (CE-GOCD), a method that augments LLMs' scientific question answering by explicitly modeling and leveraging semantic substructures within academic knowledge graphs. Our approach operates by: (1) leveraging paper titles as central entities for targeted subgraph retrieval, (2) enhancing implicit semantic discovery via subgraph pruning and completion, and (3) applying community detection to distill coherent paper groups with shared themes. We evaluated the proposed method on three NLP literature-based question-answering datasets, and the results demonstrate its superiority over other retrieval-augmented baseline approaches, confirming the effectiveness of our framework.

Reinforcement learning (RL) has become a key technique for improving the agentic capabilities of large language models (LLMs). Although critic-free methods such as GRPO are increasingly popular, we argue that critic-based methods remain well suited to long-horizon agentic tasks because their critic models can assess each state and assign credit to different decisions. However, representative critic-based methods such as PPO still organize value estimation, credit assignment, and policy updates at token granularity, whereas the environment state changes only after the agent completes a full action. In this work, we propose CAPO, a critic-guided action-aligned policy optimization method for advancing LLM agent capabilities. CAPO introduces two action-aligned designs. First, it estimates the state value at the action boundary using a critic model and assigns the corresponding advantage to the complete action rather than individual tokens. Second, CAPO introduces a novel action-aware policy ratio that aggregates token ratios within each action with length calibration for consistent updates across actions of different lengths. These two designs effectively achieve fine-grained credit assignment and consistent policy updates while remaining compatible with token-level gradient computation. Experiments across multi-hop QA, academic paper search, and text-world action tasks show that CAPO consistently outperforms representative RL methods. Further analyses validate both designs and reveal how action-aligned optimization improves multi-turn RL training. We hope this work provides a practical path for training more capable LLM agents.

Enabling machines to solve mathematical problems is a vital endeavor in developing intelligence that emulates human- like thinking and reasoning. However, most existing approaches focus on reconstructing human comprehension of problems, which are still far from enough since they neglect the fundamental human ability to learn knowledge from experiences. In this article, we focus on empowering models with the cognitive capacity to autonomously learn knowledge from mathematical problem-solving. We first propose a Cognitive Solver (CogSolver) that contains an intelligent BRAIN-ARM framework as the cognitive structure and operates the knowledge learning process in Store-Apply-Update steps inspired by two cognitive science theories. The BRAIN system stores three basic types of mathematical knowledge, and the ARM system applies them organically in answer reasoning process. After solving problems, the BRAIN updates its stored knowledge based on the ARM's feedback, with knowledge filters to eliminate redundancies and foster a more rational knowledge base. Our CogSolver carries out the above three steps iteratively, emulating a more human- like behavior. Furthermore, in order to overcome knowledge forgetting during the learning process, we extend CogSolver to CogSolver+ by incorporating an essential knowledge Recall mechanism, which is inspired by another prominent cognitive theory. We first discuss and fuse three crucial factors in simulating human memory replay. Then, we propose a influenced-based method with a theoretical guarantee of efficiency to consolidate the updated knowledge. Experiments on three math word problem benchmarks demonstrate the improvements of our CogSolver and CogSolver+ in answer reasoning and clearly illustrate how they acquire knowledge, leading to superior interpretability.

Multi-behavior sequential recommendation aims to capture users' dynamic interests by modeling diverse types of user interactions over time. Although several studies have explored this setting, the recommendation performance remains suboptimal, mainly due to two fundamental challenges: the heterogeneity of user behaviors and data sparsity. To address these challenges, we propose BLADE, a framework that enhances multi-behavior modeling while mitigating data sparsity. Specifically, to handle behavior heterogeneity, we introduce a dual item-behavior fusion architecture that incorporates behavior information at both the input and intermediate levels, enabling preference modeling from multiple perspectives. To mitigate data sparsity, we design three behavior-level data augmentation methods that operate directly on behavior sequences rather than core item sequences. These methods generate diverse augmented views while preserving the semantic consistency of item sequences. These augmented views further enhance representation learning and generalization via contrastive learning. Experiments on three real-world datasets demonstrate the effectiveness of our approach.

Specializing Large Language Models for educational domains is a key frontier in creating personalized learning tools. The central challenge is not data scarcity but its abundance: efficiently selecting a curated data subset from vast corpora to enhance specialized skills and foster generalization, without degrading existing abilities. Existing data selection paradigms, relying on superficial semantic similarity or model training dynamics, often lack a principled framework to identify data that promotes true cognitive growth. Our work proposes a paradigm shift from leveraging indirect proxies of learning value, such as semantic similarity and training dynamics, towards a framework that performs a direct, cognitive-level modeling of the learner's state. We introduce CASS, a novel framework that implements this cognitive approach through a clear pipeline, moving from an initial Diagnosis to the ultimate goal of expanding the model's cognitive frontier. First, CASS diagnoses the LLM's cognitive frontier using Multidimensional Item Response Theory. Leveraging this diagnosis, it then employs Fisher Information to select a data subset situated at LLM's cognitive frontier that offers maximum informational gain. Finally, the model is fine-tuned on this curated data using a structured, easy-to-hard curriculum to ensure effective learning. Experiments on our new multi-subject dataset show that models trained with CASS not only achieve superior accuracy in the target domain but also exhibit enhanced generalization. CASS provides a more efficient, effective, and theoretically-grounded paradigm for building expert educational LLMs.

Knowledge tagging for questions is a fundamental task in intelligent educational systems, aiming to identify the most relevant knowledge concepts for a given question. Traditionally, this task is performed by domain experts, as it requires not only a deep understanding of both the question content and knowledge definitions, but also strong logical reasoning ability. Although recent studies have attempted to automate this process using various approaches, their performance remains limited. To address this challenge, we propose PRKT, a Process-supervised multi-step Reasoning framework for Knowledge Tagging that leverages the reasoning capabilities of large language models (LLMs). Specifically, we first construct a multi-step reasoning dataset with step-by-step supervision and fine-tune an LLM to align it with structured reasoning processes, resulting in a base model RKT. We then design a Step-level Score Preference Optimization (SSPO) method to improve the quality of intermediate reasoning steps of RKT, yielding the final model PRKT. Moreover, we introduce a score-guided decoding strategy to better guide the inference process. Extensive experiments demonstrate that PRKT outperforms existing methods on knowledge tagging tasks.

Long Chain-of-Thought (LCoT), achieved by Reinforcement Learning with Verifiable Rewards (RLVR), has proven effective in enhancing the reasoning capabilities of Large Language Models (LLMs). However, reasoning in current LLMs is primarily generated as plain text, where performing semantic evaluation on such unstructured data creates a computational bottleneck during training. Despite RLVR-based optimization, existing methods still suffer from coarse-grained supervision, reward hacking, high training costs, and poor generalization. To address these issues, we propose the Graph Reasoning Paradigm (GRP), which realizes structured and symbolic reasoning, implemented via graph-structured representations with step-level cognitive labels. Building upon GRP, we further design Process-Aware Stratified Clipping Group Relative Policy Optimization (PASC-GRPO), which leverages structured evaluation to replace semantic evaluation, achieves process-aware verification through graph-structured outcome rewards, and mitigates reward hacking via stratified clipping advantage estimation. Experiments demonstrate significant improvements across mathematical reasoning and code generation tasks. Data, models, and code will be released later.

With the advancement of digital education, low student engagement in online environments has become a critical challenge, highlighting the need for reliable, robust student engagement prediction models. While deep learning models have been developed for this task, conventional single-stream visual approaches are susceptible to real-world confounding factors like occlusions and fail to capture complex spatio-temporal dynamics, while multimodal fusion strategies often neglect the hierarchical and complementary relationships between diverse data streams. To address these limitations, we propose a novel hierarchical multimodal framework termed GLP-fusion: specifically, it first extracts three complementary feature modalities-global facial features, local key point features, and physiological features-to comprehensively characterize student engagement, and subsequently adopts an innovative hierarchical fusion mechanism integrating mid-level and decision-level fusion, which effectively models intra-stream temporal dependencies and inter-level relational interactions while mitigating interference and redundancy. Comprehensive experimental evaluations and visualization analyses on the EmotiW 2023 and DAISEE benchmark datasets demonstrate that GLP-fusion outperforms state-of-the-art methods, validating its superior effectiveness, robustness, and temporal feasibility-particularly excelling in scenarios involving complex interference.

Large Language Models (LLMs) show strong cross-domain performance, further enhanced via fine-tuning on domain-specific data. While they have key breakthroughs, LLMs still struggle with complex tasks like accelerating chemical research–this needs integrating intensive environmental interactions (high-throughput experiments), sophisticated theoretical simulations (wavefunction- and density functional theory-based calculations), and reasoning (data analysis, optimization direction identification). To date, the most effective method is LLMs collaborating with existing chemical toolkits. Yet augmenting LLMs with these toolkits is challenging: they need diverse inputs, generate heterogeneous outputs, and raise LLM adaptation training costs. To tackle this, we propose SiMiao, a chemistry research assistant platform integrating LLMs (fine-tuned on chemical datasets) with chemistry-specific toolkits. SiMiao outperforms mainstream general-purpose and domain-specific models in chemical tasks, including advanced knowledge QA, literature understanding, molecular understanding, and scientific knowledge deduction. Consequently, it supports professional chemists, lowers barriers for non-experts, bridges theoretical-experimental chemistry gaps, and drives scientific progress. The official platform access link is accessible at https://ai4s.iflytek.com/chat.

Reinforcement learning (RL) for large language models usually supervises reasoning with scalar outcome rewards, such as binary correctness. Such rewards provide an optimization direction but rarely explain how a model should revise its mistaken reasoning, which can encourage shortcut learning and brittle policies. We propose SocraticPO (Socratic Policy Optimization), a policy-optimization framework that augments RL rollouts with Socratic-style natural-language guidance. During rollout, the student first answers independently; if the answer is incorrect, a teacher diagnoses the attempt and provides concise corrective guidance, after which the student continues under the expanded context. Crucially, this guidance is paired with reward decay: correct answers obtained after teacher intervention only receive decayed rewards, preventing the policy from treating teacher help as a free path to reward. Since SocraticPO only modifies the rollout process while leaving the standard expected-reward objective intact, it can be plugged into existing policy-gradient backends such as Reinforce++. Moreover, because the teacher provides only text-level guidance, SocraticPO can leverage stronger black-box teacher models without requiring access to logits or distribution matching. On undergraduate-level scientific reasoning benchmarks from SciKnowEval, SocraticPO improves over strong RL and self-distillation baselines. Ablations show that both targeted guidance and reward decay are necessary, with reward decay mitigating reliance on assisted correction.

Geological interpretation infers subsurface properties and structures from indirect geophysical observations. Well-log classification provides a measurable setting by assigning geological classes to depth-indexed petrophysical records. The task is difficult because different subsurface units may exhibit similar logging responses, whereas accurate interpretation often depends on local measurements, depth-wise context, domain knowledge, and reasonable transitions between neighboring layers. Existing automated methods mainly follow fixed prediction pipelines, leaving little room to gather additional evidence for difficult samples. In this work, we propose GeoDecider, an evidence-grounded agent for deliberative geological interpretation. GeoDecider retains efficient numerical prediction as the first stage, then selectively invokes tool-assisted reasoning for difficult intervals. A lightweight classifier produces point-wise predictions and estimates sample difficulty from its prediction scores. High-difficulty points act as routing anchors, triggering interval-level analysis so that nearby observations can be examined together. For each activated interval, specialized tools build an Evidence Profile that summarizes geological knowledge, depth-wise trends, previous predictions from the same well, and similar cases retrieved from training wells. Three complementary scientific views generate candidate interpretations. GeoDecider compares their supporting evidence, resolves disagreements, then applies geology-informed checks on continuity, boundary cues, and petrophysical consistency. Experiments on four public well-log benchmarks show that GeoDecider consistently outperforms representative baselines, demonstrating the value of selective evidence gathering and deliberative reasoning for geological interpretation. [Our code is available at https://github.com/Xiaoyu-Tao/GeoDecider]

Knowledge Tracing (KT) aims to assess students' dynamic knowledge states from their learning histories. While most existing KT methods focus on single-domain learning with notable success, real-world learning scenarios often involve multiple domains simultaneously, introducing two critical factors: 1) Cognitive load, arising from managing learning across domains in both temporal and knowledge dimensions. 2) Knowledge transfer, where knowledge states in one domain influence related states both within and across domains. In this paper, we focus on exploring these factors to improve students' knowledge state assessment in multi-domain learning scenarios and propose a novel method incorporating cognitive Load and knowledge Transfer for Multi-domain Knowledge Tracing (LT-MKT). Specifically, to bridge isolated domains, LT-MKT first integrates textual information from questions and their associated concepts to construct a Multi-domain Hierarchical Graph, leveraging the advanced representational capabilities of large language models (LLMs). Then, cross-domain features in both the temporal and knowledge dimensions are explicitly modeled to capture the effects of cognitive load. Additionally, a knowledge transfer module is designed to model the propagation of knowledge states within and across domains. By jointly modeling these factors, LT-MKT enables more accurate prediction of students' future performance. Finally, extensive experiments on real-world datasets demonstrate that our method achieves state-of-the-art performance.

While Large Language Models (LLMs) excel in natural language tasks, single-agent architectures face an inherent optimization conflict between maximizing generative response quality and mitigating factual hallucinations. To address this, we propose the Multi-Agent Collaborative Framework (MA-CF), a structured prompt-driven pipeline that decomposes generation into distinct phases of drafting, parallelized evaluation, and final synthesis. By architecturally decoupling factuality verification from linguistic assessment into separate agent roles, MA-CF empirically mitigates the suboptimal trade-offs and consensus biases observed in both monolithic single-agent systems and traditional multi-agent debate frameworks. We evaluated MA-CF using small-scale backbones (Llama3.1-8B, Qwen3-8B) against single-agent baselines, iterative debate architectures, and large-scale models across three datasets: PreciseWiki (N = 2000), LongWiki (N = 250), and HaluEval 2.0 (N = 800). Experimental results demonstrate substantial quantitative improvements. On the LongWiki dataset, MA-CF increased Llama3.1-8B's Recall@32 from 15.38% to 35.22% and F1@32 from 22.63% to 44.22%. On the cross-domain HaluEval 2.0 benchmark, the framework improved the Average Factual Rate from 65.24% to 73.28%. Furthermore, on the PreciseWiki dataset, MA-CF significantly improved the model's ability to reject unanswerable queries, effectively reducing the probability of model error. Significantly, our findings confirm that a strategic ensemble of 8B-parameter models can match or surpass the hallucination mitigation capabilities of 671B-parameter models (e.g., DeepSeek-V3), offering a computationally efficient architecture for deploying reliable LLM systems in resource-constrained information environments.