Knowledge Tracing (KT) is a critical technique for modeling student knowledge to support personalized learning. However, most KT systems focus on binary correctness prediction and cannot diagnose the underlying conceptual misunderstandings that lead to errors. Such fine-grained diagnostic feedback is essential for designing targeted instruction and effective remediation. In this work, we introduce the task of concept-level deficiency prediction, which extends traditional KT by identifying the specific concepts a student is likely to struggle with on future problems. We present ConceptKT, a dataset annotated with labels that capture both the concepts required to solve each question and the missing concepts underlying incorrect responses. We investigate in-context learning approaches to KT and evaluate the diagnostic capabilities of various Large Language Models (LLMs) and Large Reasoning Models (LRMs). Different strategies for selecting informative historical records are explored. Experimental results demonstrate that selecting response histories based on conceptual alignment and semantic similarity leads to improved performance on both correctness prediction and concept-level deficiency identification.

Aspect-based summarization has attracted significant attention for its ability to generate more fine-grained and user-aligned summaries. While most existing approaches assume a set of predefined aspects as input, real-world scenarios often present challenges where these given aspects may be incomplete, irrelevant, or entirely missing from the document. Users frequently expect systems to adaptively refine or filter the provided aspects based on the actual content. In this paper, we initiate this novel task setting, termed Content-Aware Refinement of Provided Aspects for Summarization (CARPAS), with the aim of dynamically adjusting the provided aspects based on the document context before summarizing. We construct three new datasets and conduct pilot experiments with four representative prompting strategies. Our analysis reveals that LLMs often over-generate aspects, resulting in excessively long and misaligned summaries. Building on this observation, we introduce a two-stage framework that derives lightweight scope guidance before aspect refinement and summarization, improving focus and reducing over-generation. Our extensive experiments show that the proposed approach significantly improves performance across all datasets. Moreover, our deeper analyses uncover LLMs’ compliance when the requested number of aspects differs from their own estimations, establishing a crucial insight for the deployment of LLMs in similar real-world applications.

Large language models are increasingly used in practical systems, making efficient model selection important for reducing deployment cost. LLM routing has emerged as a practical solution for allocating each input query to an appropriate model under a desired cost-performance trade-off. Existing routing methods often estimate model suitability from the surface semantics or embedding similarity of the input query. However, such methods may ignore the underlying difficulty of a query, leading to suboptimal routing decisions. To address the challenge, we propose VDAR-Router, a difficulty-aware retrieval-based routing framework. For each input query, VDAR-Router first generates an explicit difficulty analysis. It then retrieves historical examples with similar difficulty profiles. Based on the retrieved records, it estimates candidate model suitability and selects the model using a reward function that considers both performance and cost. Experiments on three datasets show that VDAR-Router consistently achieves better cost-performance trade-offs than existing baselines. These results demonstrate the effectiveness of difficulty-aware retrieval for training-free LLM routing. Case studies further show that explicit query analysis helps retrieve more relevant examples and supports more reliable routing decisions.

VLMs are increasingly positioned as daily assistants that perceive first-person environments, follow user dialogue, and decide how to help. Existing egocentric benchmarks mainly evaluate visual understanding in isolation, leaving open whether models can arbitrate between visual evidence and user-provided language when the two are helpful, irrelevant, or conflicting. We introduce EgoArgus, a human-annotated dataset for evaluating egocentric assistants on understanding and decision tasks in five dialogue-video daily scenarios. Our results demonstrate that it is still challenging for current VLMs as reliable egocentric assistants, which requires identifying which modality is trustworthy and deciding when intervention is warranted. Deeper analysis also shows that existing modality bias mitigation methods are quite restricted to enhance performance, providing insights to aid practioners into the deployment of current VLMs as daily assistants.

First-person assistance situations are difficult to collect with controlled timing and evidence: rare hazards should not be staged, while ordinary needs often depend on subtle visual and conversational context. We present VISTA, an interactive platform that treats scenario creation as an inspectable authoring workflow rather than a one-shot video prompt. A six-stage compiler turns a seed into typed intent, scene, event, interaction, rendering, and media artifacts; human reviewers can inspect revisions and provider requests before approving execution. We evaluate the rendered outputs, not the authoring interface. In a blinded study of 60 scenarios, 11 evaluators supplied 165 judgments. They selected the human-reviewed VISTA output as the best match to the seed in 52.1% of judgments, versus 22.4% for a one-pass seed-plus-assets baseline and 25.5% for seed-only generation. Case-balanced seed-match scores were 3.65, 3.27, and 3.22, respectively. These results support the complete workflow under the tested renderers; they do not establish interface usability, authoring efficiency, or a novel planning algorithm. The public artifact is a credential-free prepared replay with a narrated screencast.

The increasing reliance on Large Language Models (LLMs) across various domains extends to education, where students progressively use generative AI as a tool for learning. While prior work has examined LLMs’ mathematical ability, their reliability in grading authentic student problem-solving processes and delivering effective feedback remains underexplored. This study introduces MathEDU, a dataset consisting of student problem-solving processes in mathematics and corresponding teacher-written feedback. We systematically evaluate the reliability of various models across three hierarchical tasks: answer correctness classification, error identification, and feedback generation. Experimental results show that fine-tuning strategies effectively improve performance in classifying correctness and locating erroneous steps. However, the generated feedback across models shows a considerable gap from teacher-written feedback. Critically, the generated feedback is often verbose and fails to provide targeted explanations for the student’s underlying misconceptions. This emphasizes the urgent need for trustworthy and pedagogy-aware AI feedback in education.

The dialogue systems in customer services have been developed with neural models to provide users with precise answers and round-the-clock support in task-oriented conversations by detecting customer intents based on their utterances. Existing intent detection approaches have highly relied on adaptively pre-training language models with large-scale datasets, yet the predominant cost of data collection may hinder their superiority. In addition, they neglect the information within the conversational responses of the agents, which have a lower collection cost, but are significant to customer intent as agents must tailor their replies based on the customers’ intent. In this paper, we propose RSVP, a self-supervised framework dedicated to task-oriented dialogues, which utilizes agent responses for pre-training in a two-stage manner. Specifically, we introduce two pre-training tasks to incorporate the relations of utterance-response pairs: 1) Response Retrieval by selecting a correct response from a batch of candidates, and 2) Response Generation by mimicking agents to generate the response to a given utterance. Our benchmark results for two real-world customer service datasets show that RSVP significantly outperforms the state-of-the-art baselines by 4.95% for accuracy, 3.4% for MRR@3, and 2.75% for MRR@5 on average. Extensive case studies are investigated to show the validity of incorporating agent responses into the pre-training stage.

This paper presents a state-of-the-art solution to the LT-EDI-ACL 2022 Task 4: Detecting Signs of Depression from Social Media Text. The goal of this task is to detect the severity levels of depression of people from social media posts, where people often share their feelings on a daily basis. To detect the signs of depression, we propose a framework with pre-trained language models using rich information instead of training from scratch, gradient boosting and deep learning models for modeling various aspects, and supervised contrastive learning for the generalization ability. Moreover, ensemble techniques are also employed in consideration of the different advantages of each method. Experiments show that our framework achieves a 2nd prize ranking with a macro F1-score of 0.552, showing the effectiveness and robustness of our approach.