Yanxin Chen

Ling He, Yanxin Chen, Xiaoqiang Hu

This study delves into the application potential of the large language models (LLMs) ChatGLM in the automatic generation of structured questions for National Teacher Certification Exams (NTCE). Through meticulously designed prompt engineering, we guided ChatGLM to generate a series of simulated questions and conducted a comprehensive comparison with questions recollected from past examinees. To ensure the objectivity and professionalism of the evaluation, we invited experts in the field of education to assess these questions and their scoring criteria. The research results indicate that the questions generated by ChatGLM exhibit a high level of rationality, scientificity, and practicality similar to those of the real exam questions across most evaluation criteria, demonstrating the model's accuracy and reliability in question generation. Nevertheless, the study also reveals limitations in the model's consideration of various rating criteria when generating questions, suggesting the need for further optimization and adjustment. This research not only validates the application potential of ChatGLM in the field of educational assessment but also provides crucial empirical support for the development of more efficient and intelligent educational automated generation systems in the future.

Pengyu Wang, Chenkun Tan, Shaojun Zhou et al.

Video understanding is shifting from the offline paradigm -- taking a fully recorded video as input and producing a single answer after it ends -- toward real-time interaction, in which the model perceives new frames while still replying, revises its answer as new evidence appears, and remains silent when there is nothing to say. We present MOSS-Video-Preview to validate this paradigm. Our central claim is that perception must not be blocked by generation; its natural realization is a two-channel architecture. We argue that a cross-attention backbone is better suited to real-time vision-language fusion than the prevailing decoder-only design: visual features enter through a side channel rather than joining the autoregressive sequence, so perception and generation run on separate, non-blocking pathways -- reducing the frequency of visual processing and exposing a clean channel-wise interface for independent compression. We complement this with a data synthesis pipeline that converts dense captions into real-time understanding QA whose answers are revised to match what the model has perceived so far, and we specialize an offline model on these data to elicit real-time behavior. Our model trails the strong Qwen2.5-VL-7B baseline overall -- a gap we attribute primarily to data and scale rather than the architecture -- yet attains competitive offline video and multimodal understanding, remains robust on the spatial and fine-grained temporal reasoning central to real-time use, and acquires behaviors that offline models lack: continuous perception, answer revision, and timely silence. On a single H200 with 256 frames per video, it achieves about a 5x speedup in time to first token and 2.7x higher decoding throughput, with negligible degradation in offline ability. Our study of paradigm, architecture, and data outlines a viable path toward real-time video understanding.

Hui Li, Pengfei Yang, Juanyang Chen et al.

Knowledge distillation as an efficient knowledge transfer technique, has achieved remarkable success in unimodal scenarios. However, in cross-modal settings, conventional distillation methods encounter significant challenges due to data and statistical heterogeneities, failing to leverage the complementary prior knowledge embedded in cross-modal teacher models. This paper empirically reveals two critical issues in existing approaches: distillation path selection and knowledge drift. To address these limitations, we propose MST-Distill, a novel cross-modal knowledge distillation framework featuring a mixture of specialized teachers. Our approach employs a diverse ensemble of teacher models across both cross-modal and multimodal configurations, integrated with an instance-level routing network that facilitates adaptive and dynamic distillation. This architecture effectively transcends the constraints of traditional methods that rely on monotonous and static teacher models. Additionally, we introduce a plug-in masking module, independently trained to suppress modality-specific discrepancies and reconstruct teacher representations, thereby mitigating knowledge drift and enhancing transfer effectiveness. Extensive experiments across five diverse multimodal datasets, spanning visual, audio, and text, demonstrate that our method significantly outperforms existing state-of-the-art knowledge distillation methods in cross-modal distillation tasks. The source code is available at https://github.com/Gray-OREO/MST-Distill.