arXiv subjects

Ningning Xu

Publications and source records attributed to Ningning Xu.

8 recordsLinked to original sources

arXiv2025

Learning How to Use Tools, Not Just When: Pattern-Aware Tool-Integrated Reasoning

Tool-integrated reasoning (TIR) has become a key approach for improving large reasoning models (LRMs) on complex problems. Prior work has mainly studied when to invoke tools, while overlooking how tools are applied. We identify two common patterns: a calculator pattern that uses code for direct computation, and an algorithmic pattern that encodes problems as programs. Misaligned choices often cause failures even when reasoning is sound. We propose a two-stage framework that first builds code competence from both patterns and then aligns pattern selection with teacher preferences. Across challenging math datasets, our pattern-aware method substantially improves both code usage and accuracy, for instance raising Code@1 on MATH500 from 64.0% to 70.5% and on AIME24 from 26.7% to 50.0%. These gains highlight the effectiveness of a pattern-aware approach for tool-integrated reasoning.

cs.AI↗

arXiv2025

Multi-Agent Visual-Language Reasoning for Comprehensive Highway Scene Understanding

This paper introduces a multi-agent framework for comprehensive highway scene understanding, designed around a mixture-of-experts strategy. In this framework, a large generic vision-language model (VLM), such as GPT-4o, is contextualized with domain knowledge to generates task-specific chain-of-thought (CoT) prompts. These fine-grained prompts are then used to guide a smaller, efficient VLM (e.g., Qwen2.5-VL-7B) in reasoning over short videos, along with complementary modalities as applicable. The framework simultaneously addresses multiple critical perception tasks, including weather classification, pavement wetness assessment, and traffic congestion detection, achieving robust multi-task reasoning while balancing accuracy and computational efficiency. To support empirical validation, we curated three specialized datasets aligned with these tasks. Notably, the pavement wetness dataset is multimodal, combining video streams with road weather sensor data, highlighting the benefits of multimodal reasoning. Experimental results demonstrate consistently strong performance across diverse traffic and environmental conditions. From a deployment perspective, the framework can be readily integrated with existing traffic camera systems and strategically applied to high-risk rural locations, such as sharp curves, flood-prone lowlands, or icy bridges. By continuously monitoring the targeted sites, the system enhances situational awareness and delivers timely alerts, even in resource-constrained environments.

cs.CV↗

arXiv2025

Structured Prompting and Multi-Agent Knowledge Distillation for Traffic Video Interpretation and Risk Inference

Comprehensive highway scene understanding and robust traffic risk inference are vital for advancing Intelligent Transportation Systems (ITS) and autonomous driving. Traditional approaches often struggle with scalability and generalization, particularly under the complex and dynamic conditions of real-world environments. To address these challenges, we introduce a novel structured prompting and knowledge distillation framework that enables automatic generation of high-quality traffic scene annotations and contextual risk assessments. Our framework orchestrates two large Vision-Language Models (VLMs): GPT-4o and o3-mini, using a structured Chain-of-Thought (CoT) strategy to produce rich, multi-perspective outputs. These outputs serve as knowledge-enriched pseudo-annotations for supervised fine-tuning of a much smaller student VLM. The resulting compact 3B-scale model, named VISTA (Vision for Intelligent Scene and Traffic Analysis), is capable of understanding low-resolution traffic videos and generating semantically faithful, risk-aware captions. Despite its significantly reduced parameter count, VISTA achieves strong performance across established captioning metrics (BLEU-4, METEOR, ROUGE-L, and CIDEr) when benchmarked against its teacher models. This demonstrates that effective knowledge distillation and structured multi-agent supervision can empower lightweight VLMs to capture complex reasoning capabilities. The compact architecture of VISTA facilitates efficient deployment on edge devices, enabling real-time risk monitoring without requiring extensive infrastructure upgrades.

cs.CV↗

arXiv2024

Leveraging Scene Geometry and Depth Information for Robust Image Deraining

Image deraining holds great potential for enhancing the vision of autonomous vehicles in rainy conditions, contributing to safer driving. Previous works have primarily focused on employing a single network architecture to generate derained images. However, they often fail to fully exploit the rich prior knowledge embedded in the scenes. Particularly, most methods overlook the depth information that can provide valuable context about scene geometry and guide more robust deraining. In this work, we introduce a novel learning framework that integrates multiple networks: an AutoEncoder for deraining, an auxiliary network to incorporate depth information, and two supervision networks to enforce feature consistency between rainy and clear scenes. This multi-network design enables our model to effectively capture the underlying scene structure, producing clearer and more accurately derained images, leading to improved object detection for autonomous vehicles. Extensive experiments on three widely-used datasets demonstrated the effectiveness of our proposed method.

cs.CV↗

arXiv2024

OCEAN: Flexible Feature Set Aggregation for Analysis of Multi-omics Data

Integrated analysis of multi-omics datasets holds great promise for uncovering complex biological processes. However, the large dimension of omics data poses significant interpretability and multiple testing challenges. Simultaneous Enrichment Analysis (SEA) was introduced to address these issues in single-omics analysis, providing an in-built multiple testing correction and enabling simultaneous feature set testing. In this paper, we introduce OCEAN, an extension of SEA to multi-omics data. OCEAN is a flexible approach to analyze potentially all possible two-way feature sets from any pair of genomics datasets. We also propose two new error rates which are in line with the two-way structure of the data and facilitate interpretation of the results. The power and utility of OCEAN is demonstrated by analyzing copy number and gene expression data for breast and colon cancer.

stat.ME↗

arXiv2024

Depth Estimation Based on 3D Gaussian Splatting Siamese Defocus

Depth estimation is a fundamental task in 3D geometry. While stereo depth estimation can be achieved through triangulation methods, it is not as straightforward for monocular methods, which require the integration of global and local information. The Depth from Defocus (DFD) method utilizes camera lens models and parameters to recover depth information from blurred images and has been proven to perform well. However, these methods rely on All-In-Focus (AIF) images for depth estimation, which is nearly impossible to obtain in real-world applications. To address this issue, we propose a self-supervised framework based on 3D Gaussian splatting and Siamese networks. By learning the blur levels at different focal distances of the same scene in the focal stack, the framework predicts the defocus map and Circle of Confusion (CoC) from a single defocused image, using the defocus map as input to DepthNet for monocular depth estimation. The 3D Gaussian splatting model renders defocused images using the predicted CoC, and the differences between these and the real defocused images provide additional supervision signals for the Siamese Defocus self-supervised network. This framework has been validated on both artificially synthesized and real blurred datasets. Subsequent quantitative and visualization experiments demonstrate that our proposed framework is highly effective as a DFD method.

cs.CV↗

arXiv2022

Terahertz sensing of highly absorptive water-methanol mixtures with multiple resonances in metamaterials

Ultrasensitive terahertz sensing of highly absorptive aqueous solutions remains challenging due to strong absorption of water in the terahertz regime. Here, we experimentally demonstrate a cost-effective metamaterial-based sensor integrated with terahertz time-domain spectroscopy for highly absorptive water-methanol mixture sensing. This metamaterial has simple asymmetric wire structures that support multiple resonances including a fundamental Fano resonance and higher order dipolar resonance in the terahertz regime. Both the resonance modes have strong intensity in the transmission spectra which we exploit for detection of the highly absorptive water-methanol mixtures. The experimentally characterized sensitivities of the Fano and dipole resonances for the water-methanol mixtures are found to be 160 and 305 GHz/RIU, respectively. This method provides a route for readily available metamaterial-assisted terahertz spectroscopy for ultrasensitive sensing of highly absorptive chemical and biochemical materials with multiple resonances and high accuracy.

physics.optics↗

arXiv2021

Pathway Testing in Metabolomics with Globaltest, Allowing Post Hoc Choice of Pathways

The Globaltest is a powerful test for the global null hypothesis that there is no association between a group of features and a response of interest, which is popular in pathway testing in metabolomics. Evaluating multiple pathways, however, requires multiple testing correction. In this paper, we propose a multiple testing method, based on closed testing, specifically designed for the Globaltest. The proposed method controls the family-wise error rate simultaneously over all possible feature sets, and therefore allows post hoc inference, i.e. the researcher may choose the pathway database after seeing the data without jeopardizing error control. To circumvent the exponential computation time of closed testing, we derive a novel shortcut that allows exact closed testing to be performed on the scale of metabolomics data. An R package ctgt is available on CRAN. We illustrate the shortcut on several metabolomics data examples.

stat.ME↗