Beam Squint and Aperture--Bandwidth Limitations in Wideband RIS-Assisted THz Links
本文研究了在波束倾斜和有限分辨率相位控制下，宽带RIS辅助太赫兹链路的遍历速率问题，并通过精确推导信道矩及蒙特卡洛仿真验证了解析框架。
0 citationsRead paper
Selected work
Representative Papers
Design-Assisted Regression
本文提出一种设计辅助回归框架，通过利用协变量分布信息来稳定弱设计方向和修正潜在效应扭曲，从而改进估计性能。
0 citationsRead paper
InfluenceField: A Differentiable Field with Interventionally Identifiable Causal Structure for Multimodal World Modeling
为解决多模态大模型中视觉干预效果预测问题，提出InfluenceField方法，在视觉编码器和语言解码器间插入可微且具有因果结构的场，通过联合训练优化多个目标。
0 citationsRead paper
Rethinking Image Processing for the Age of AI: A Problem-First Framework for Scientific Progress
本文针对图像处理领域过度依赖模型优化而忽视实际问题的现象，提出了一种以问题为导向的研究框架，通过区分物理成像问题、解决方案原理等步骤，并强调现代AI的应用边界与未解决问题。
0 citationsRead paper
Diagnosing Dense Same-Class Attribute Misbinding in Large Vision-Language Models
This study addresses the challenge of attribute misbinding in large vision-language models within dense homogeneous scenes, where existing metrics prove inadequate. We formally define the DSCAM task and construct InstaBind-Lite, a controlled benchmark accompanied by a specialized evaluation framework. Through fine-grained instance annotation and multi-level question-answering design, this work enables quantitative assessment of attribute transfer while exposing critical blind spots in traditional evaluations. Experiments reveal misbinding rates of 19.84% for open-source models and 7.55% for API-based models, precisely localizing error sources. Ultimately, this research establishes a novel, traceable evaluation paradigm for assessing fine-grained attribute binding capabilities in large multimodal models.
0 citationsRead paper
Recent publications
Latest Papers
Beam Squint and Aperture--Bandwidth Limitations in Wideband RIS-Assisted THz Links
本文研究了在波束倾斜和有限分辨率相位控制下，宽带RIS辅助太赫兹链路的遍历速率问题，并通过精确推导信道矩及蒙特卡洛仿真验证了解析框架。
0 citationsRead paper
Design-Assisted Regression
本文提出一种设计辅助回归框架，通过利用协变量分布信息来稳定弱设计方向和修正潜在效应扭曲，从而改进估计性能。
0 citationsRead paper
InfluenceField: A Differentiable Field with Interventionally Identifiable Causal Structure for Multimodal World Modeling
为解决多模态大模型中视觉干预效果预测问题，提出InfluenceField方法，在视觉编码器和语言解码器间插入可微且具有因果结构的场，通过联合训练优化多个目标。
0 citationsRead paper
Rethinking Image Processing for the Age of AI: A Problem-First Framework for Scientific Progress
本文针对图像处理领域过度依赖模型优化而忽视实际问题的现象，提出了一种以问题为导向的研究框架，通过区分物理成像问题、解决方案原理等步骤，并强调现代AI的应用边界与未解决问题。
0 citationsRead paper
Diagnosing Dense Same-Class Attribute Misbinding in Large Vision-Language Models
This study addresses the challenge of attribute misbinding in large vision-language models within dense homogeneous scenes, where existing metrics prove inadequate. We formally define the DSCAM task and construct InstaBind-Lite, a controlled benchmark accompanied by a specialized evaluation framework. Through fine-grained instance annotation and multi-level question-answering design, this work enables quantitative assessment of attribute transfer while exposing critical blind spots in traditional evaluations. Experiments reveal misbinding rates of 19.84% for open-source models and 7.55% for API-based models, precisely localizing error sources. Ultimately, this research establishes a novel, traceable evaluation paradigm for assessing fine-grained attribute binding capabilities in large multimodal models.
0 citationsRead paper