Selected work

Representative Papers

FoCal: Frequency-Oriented Cross-Modal Interaction and Spectral Calibration for Aerial Visible-Infrared Object Detection

Sep 24, 2026

This study addresses the insufficient cross-modal frequency interaction and improper utilization of spectral discrepancies in aerial visible-infrared object detection by proposing the FoCal framework. FoCal introduces a pioneering frequency-aware dual-domain calibration mechanism alongside discrepancy-guided spectral modulation. By integrating cross-modal attention with adaptive gating, the framework achieves low-frequency consensus fusion while preserving high-frequency specificity, thereby accurately extracting complementary information across frequency bands and avoiding indiscriminate fusion. Extensive experiments demonstrate that FoCal attains an mAP50 of 83.5% on benchmarks such as DroneVehicle. Notably, with only 3.0M parameters, it achieves an inference speed of 113.6 FPS, effectively balancing detection accuracy with computational efficiency for real-time deployment.

0 citationsRead paper

GeoRay: Gauge-Aware Feed-Forward Satellite 3D Reconstruction in the Geodetic Frame

Aug 30, 2026

本文提出GeoRay方法，解决卫星摄影测量中绝对地理框架下的密集表面高度重建问题，通过轻量级射线一致适配器和显式基准机制实现。

0 citationsRead paper

DroneFINE: Domain-Aware Parameter-Efficient Fine-Tuning of Vision-Language Detectors for Drone Images

Jun 30, 2026

Existing vision-language models exhibit limited performance in drone image object detection due to significant domain discrepancies, and conventional parameter-efficient fine-tuning (PEFT) approaches struggle to address the unique challenges of aerial imagery—namely, the bird’s-eye view, background-dominated scenes, and extremely small objects. To overcome these limitations, this work proposes DroneFINE, a novel domain-aware dual-module framework. It introduces a dynamic multi-path HyperAdapter for flexible feature adaptation and a text-guided SemanticGate to effectively suppress irrelevant background clutter. By transcending the static architectural constraints of traditional PEFT methods, DroneFINE achieves substantial performance gains on the VisDrone and UAVDT benchmarks, closely approaching the accuracy of full fine-tuning while requiring only a minimal number of trainable parameters.

0 citationsRead paper

Holistic Data Scheduler for LLM Pre-training via Multi-Objective Reinforcement Learning

Jun 23, 2026

Existing online data mixing methods are limited to a single optimization objective, making them ill-suited for the multidimensional and dynamic data composition requirements of large language model pretraining. This work formulates data scheduling as a reinforcement learning problem in a continuous control space and introduces the first multi-objective reward function that integrates data-driven, loss-driven, and model-driven perspectives. Leveraging the Soft Actor-Critic algorithm, the approach enables efficient exploration and policy optimization. Experimental results demonstrate that the proposed method achieves the validation perplexity of the best baseline on The Pile using only 56% of the training steps, while also delivering a 7.2% improvement on zero-shot MMLU performance and consistent gains across multiple benchmark evaluations.

0 citationsRead paper

ReMMD: Realistic Multilingual Multi-Image Agentic Verification for Multimodal Misinformation Detection

Jun 22, 2026

Existing approaches struggle to effectively detect multimodal misinformation in real-world scenarios characterized by multilingual long-form text, multiple images, heterogeneous sources, and fine-grained textual-visual inconsistencies. To address this challenge, this work introduces ReMMDBench, the first real-world benchmark supporting multilingualism, multi-image inputs, fine-grained labels, and evidence provenance. Furthermore, it proposes an agent-based verification framework endowed with persistent memory, which decomposes claims into atomic assertions and caches structured evidence for reuse. By integrating large vision-language models with cross-modal fusion strategies, the framework achieves efficient and accurate misinformation detection. On a five-class veracity assessment task, the method attains 41.80% accuracy and 39.12% macro F1-score, reducing verification costs by up to 79.9% compared to baseline approaches.

0 citationsRead paper

Recent publications

Latest Papers

FoCal: Frequency-Oriented Cross-Modal Interaction and Spectral Calibration for Aerial Visible-Infrared Object Detection

Sep 24, 2026

This study addresses the insufficient cross-modal frequency interaction and improper utilization of spectral discrepancies in aerial visible-infrared object detection by proposing the FoCal framework. FoCal introduces a pioneering frequency-aware dual-domain calibration mechanism alongside discrepancy-guided spectral modulation. By integrating cross-modal attention with adaptive gating, the framework achieves low-frequency consensus fusion while preserving high-frequency specificity, thereby accurately extracting complementary information across frequency bands and avoiding indiscriminate fusion. Extensive experiments demonstrate that FoCal attains an mAP50 of 83.5% on benchmarks such as DroneVehicle. Notably, with only 3.0M parameters, it achieves an inference speed of 113.6 FPS, effectively balancing detection accuracy with computational efficiency for real-time deployment.

0 citationsRead paper

GeoRay: Gauge-Aware Feed-Forward Satellite 3D Reconstruction in the Geodetic Frame

Aug 30, 2026

本文提出GeoRay方法，解决卫星摄影测量中绝对地理框架下的密集表面高度重建问题，通过轻量级射线一致适配器和显式基准机制实现。

0 citationsRead paper

DroneFINE: Domain-Aware Parameter-Efficient Fine-Tuning of Vision-Language Detectors for Drone Images

Jun 30, 2026

Existing vision-language models exhibit limited performance in drone image object detection due to significant domain discrepancies, and conventional parameter-efficient fine-tuning (PEFT) approaches struggle to address the unique challenges of aerial imagery—namely, the bird’s-eye view, background-dominated scenes, and extremely small objects. To overcome these limitations, this work proposes DroneFINE, a novel domain-aware dual-module framework. It introduces a dynamic multi-path HyperAdapter for flexible feature adaptation and a text-guided SemanticGate to effectively suppress irrelevant background clutter. By transcending the static architectural constraints of traditional PEFT methods, DroneFINE achieves substantial performance gains on the VisDrone and UAVDT benchmarks, closely approaching the accuracy of full fine-tuning while requiring only a minimal number of trainable parameters.

0 citationsRead paper

Holistic Data Scheduler for LLM Pre-training via Multi-Objective Reinforcement Learning

Jun 23, 2026

Existing online data mixing methods are limited to a single optimization objective, making them ill-suited for the multidimensional and dynamic data composition requirements of large language model pretraining. This work formulates data scheduling as a reinforcement learning problem in a continuous control space and introduces the first multi-objective reward function that integrates data-driven, loss-driven, and model-driven perspectives. Leveraging the Soft Actor-Critic algorithm, the approach enables efficient exploration and policy optimization. Experimental results demonstrate that the proposed method achieves the validation perplexity of the best baseline on The Pile using only 56% of the training steps, while also delivering a 7.2% improvement on zero-shot MMLU performance and consistent gains across multiple benchmark evaluations.

0 citationsRead paper

ReMMD: Realistic Multilingual Multi-Image Agentic Verification for Multimodal Misinformation Detection

Jun 22, 2026

Existing approaches struggle to effectively detect multimodal misinformation in real-world scenarios characterized by multilingual long-form text, multiple images, heterogeneous sources, and fine-grained textual-visual inconsistencies. To address this challenge, this work introduces ReMMDBench, the first real-world benchmark supporting multilingualism, multi-image inputs, fine-grained labels, and evidence provenance. Furthermore, it proposes an agent-based verification framework endowed with persistent memory, which decomposes claims into atomic assertions and caches structured evidence for reuse. By integrating large vision-language models with cross-modal fusion strategies, the framework achieves efficient and accurate misinformation detection. On a five-class veracity assessment task, the method attains 41.80% accuracy and 39.12% macro F1-score, reducing verification costs by up to 79.9% compared to baseline approaches.

0 citationsRead paper