Chao Feng

Wangjie You, Xusheng Wang, Xing Wang et al.

While Large Language Models (LLMs) have demonstrated advanced reasoning capabilities, their comprehensive evaluation in general Chinese-language contexts remains understudied. To bridge this gap, we propose Chinese Commonsense Multi-hop Reasoning (CCMOR), a novel benchmark designed to evaluate LLMs' ability to integrate Chinese-specific factual knowledge with multi-step logical reasoning. Specifically, we first construct a domain-balanced seed set from existing QA datasets, then develop an LLM-powered pipeline to generate multi-hop questions anchored on factual unit chains. To ensure the quality of resulting dataset, we implement a human-in-the-loop verification system, where domain experts systematically validate and refine the generated questions. Using CCMOR, we evaluate state-of-the-art LLMs, demonstrating persistent limitations in LLMs' ability to process long-tail knowledge and execute knowledge-intensive reasoning. Notably, retrieval-augmented generation substantially mitigates these knowledge gaps, yielding significant performance gains.

Haiyang Yu, Yuchuan Wu, Fan Shi et al.

Chinese ancient documents, invaluable carriers of millennia of Chinese history and culture, hold rich knowledge across diverse fields but face challenges in digitization and understanding, i.e., traditional methods only scan images, while current Vision-Language Models (VLMs) struggle with their visual and linguistic complexity. Existing document benchmarks focus on English printed texts or simplified Chinese, leaving a gap for evaluating VLMs on ancient Chinese documents. To address this, we present AncientDoc, the first benchmark for Chinese ancient documents, designed to assess VLMs from OCR to knowledge reasoning. AncientDoc includes five tasks (page-level OCR, vernacular translation, reasoning-based QA, knowledge-based QA, linguistic variant QA) and covers 14 document types, over 100 books, and about 3,000 pages. Based on AncientDoc, we evaluate mainstream VLMs using multiple metrics, supplemented by a human-aligned large language model for scoring.

Xiaoxiao Yang, Chao Feng, Jiancheng Chen

Portable and wearable consumer-grade electroencephalography (EEG) devices, like Muse headbands, offer unprecedented mobility for daily brain-computer interface (BCI) applications, including cognitive load detection. However, the exacerbated non-stationarity in portable EEG signals constrains data fidelity and decoding accuracy, creating a fundamental trade-off between portability and performance. To mitigate such limitation, we propose MuseCogNet (Muse-based Cognitive Network), a unified joint learning framework integrating self-supervised and supervised training paradigms. In particular, we introduce an EEG-grounded self-supervised reconstruction loss based on average pooling to capture robust neurophysiological patterns, while cross-entropy loss refines task-specific cognitive discriminants. This joint learning framework resembles the bottom-up and top-down attention in humans, enabling MuseCogNet to significantly outperform state-of-the-art methods on a publicly available Muse dataset and establish an implementable pathway for neurocognitive monitoring in ecological settings.