1
阿里推出国内首个AI语音平台CosyVoice Studio，将语义理解融入语音能力
阿里发布CosyVoice Studio，AI语音平台融入语义理解，听感更懂人话，专业数字转写也精准。
AI语音“听、说、创”一站式满足
2
ArtAnno: Annotating Implicit Semantics in Artworks through LLM Agent-Driven Bidirectional Human-AI Augmentation
用AI智能体给艺术品标注隐性语义，开启人机双向增强新范式。
arXiv:2608.05026v1 Announce Type: cross Abstract: High-quality annotation of artworks is essential for computational art research, yet extracting impl…
3
Unsupervised Pixel-Level Semantic Left-Right Understanding of In-the-Wild Images
不用人工标注，模型也能学会自然图像的左右语义，像素级理解新思路。
arXiv:2607.05006v1 Announce Type: new Abstract: While various works address reflective symmetry understanding in 3D data and images, pixel-level seman…
4
A Hybrid Framework for Song Lyric Annotation Based on Human-LLM Alignment
人类智慧与AI大模型协同，突破歌词标注的语义理解瓶颈
arXiv:2606.29273v1 Announce Type: new Abstract: Emotion recognition of song lyrics is a challenging task since lyrics may not necessarily align with t…
5
Metaphors are a Source of Cross-Domain Misalignment of Large Reasoning Models
大型推理模型在跨域隐喻理解上存在对齐偏差，揭示隐喻是模型跨域误对齐的关键来源，为提升AI语义鲁棒性提供新视角。
arXiv:2601.03388v3 Announce Type: replace-cross Abstract: Earlier research has shown that metaphors influence human decision-making, raising the quest…
6
When Similar Means Different: Evaluating LLMs on Arabic--Hebrew Cognates
测试大模型能否辨识阿拉伯语与希伯来语中形似义异的同源词，揭示多语言语义理解短板。
arXiv:2606.13218v1 Announce Type: new Abstract: Arabic and Hebrew, as closely related Semitic languages, share a substantial lexicon of true cognates,…
7
System Report for CCL25-Eval Task 5: New Dataset and LoRA-Fine-Tuned Qwen2.5
新数据集与LoRA微调Qwen2.5突破古典诗歌翻译和情感语义理解瓶颈，技术细节值得深挖。
arXiv:2606.12392v1 Announce Type: cross Abstract: Recently, large language models (LLMs) have achieved promising progress in the fields of classical C…
8
Beyond Semantic Understanding: Preserving Collaborative Frequency Components in LLM-based Recommendation
揭露LLM推荐系统过度依赖语义的缺陷，提出保留协同频率成分的新思路
arXiv:2508.10312v2 Announce Type: replace Abstract: Recommender systems in concert with Large Language Models (LLMs) present promising avenues for gen…
9
LLMs Lean on Priors, Not Programming Language Semantics
ICML 2026论文揭示：LLM代码理解靠先验知识，而非编程语言语义
arXiv:2510.03415v3 Announce Type: replace-cross Abstract: Recent work asks whether large language models (LLMs) condition their reasoning on explicit …
10
Does Slightly Mean Somewhat? Measuring Vague Intensity Words in LLM Numeric Actions
语言模型如何理解“稍微”与“大幅”的细微差别？这项研究用一个精心构建的程度词量表测试了LLM在数值行动中的语义一致性。
arXiv:2605.21827v1 Announce Type: new Abstract: Do language models preserve the ordinal meaning of intensity words when those words must produce numer…
11
Presupposition and Reasoning in Conditionals: A Theory-Based Study of Humans and LLMs
对比人类与LLM在条件句预设投射上的推理能力，揭示模型语义理解的差异与局限。
arXiv:2605.18352v1 Announce Type: new Abstract: Presupposition projection in conditionals is central to theories of meaning and pragmatics, yet it rem…