arXiv subjects

Qing Miao

Publications and source records attributed to Qing Miao.

8 recordsLinked to original sources

arXiv2026

ConSteer-RL: Steering Reasoning Capabilities in Large Language Models via Confidence-Aware Reinforcement Learning

Reinforcement Learning from Verifiable Rewards (RLVR) has recently become a key paradigm for improving the reasoning abilities of Large Language Models (LLMs), yet it remains limited by sparse binary rewards and its ignorance of model-internal uncertainty. In this paper, we propose ConSteer-RL, a simple yet effective framework that integrates token-level confidence signals derived from model log-probabilities into RLVR training. Specifically, building upon the Group Relative Policy Optimization (GRPO) framework, we construct a confidence-aware reward by aggregating per-token probabilities into a scalar confidence score and incorporating it into an awareness-based reward shaping mechanism that penalizes overconfident errors while reinforcing correct and confident reasoning. Experimental results demonstrate that ConSteer-RL consistently outperforms strong GRPO baselines, achieving average improvements of 2.3%-4.0% across different model scales.

cs.LG↗

arXiv2026

VideoSeeker: Incentivizing Instance-level Video Understanding via Native Agentic Tool Invocation

Large Vision-Language Models (LVLMs) have shown significant progress in video understanding, yet they face substantial challenges in tasks requiring precise spatiotemporal localization at the instance level. Existing methods primarily rely on text prompts for human-model interaction, but these prompts struggle to provide precise spatial and temporal references, resulting in poor user experience. Furthermore, current approaches typically decouple visual perception from language reasoning, centering reasoning around language rather than visual content, which limits the model's ability to proactively perceive fine-grained visual evidence. To address these challenges, we propose VideoSeeker, a novel paradigm for instance-level video understanding through visual prompts. VideoSeeker seamlessly integrates agentic reasoning with instance-level video understanding tasks, enabling the model to proactively perceive and retrieve relevant video segments on demand. We construct a four-stage fully automated data synthesis pipeline to efficiently generate large-scale, high-quality instance-level video data. We internalize tool-calling and proactive perception capabilities into the model via cold-start supervision and RL training, building a powerful video understanding model. Experiments demonstrate that our model achieves an average improvement of +13.7% over baselines on instance-level video understanding tasks, surpassing powerful closed-source models such as GPT-4o and Gemini-2.5-Pro, while also showing effective transferability on general video understanding benchmarks. The relevant datasets and code will be released publicly.

cs.CV↗

arXiv2026

Nemotron 3 Nano Omni: Efficient and Open Multimodal Intelligence

We introduce Nemotron 3 Nano Omni, the latest model in the Nemotron multimodal series and the first to natively support audio inputs alongside text, images, and video. Nemotron 3 Nano Omni delivers consistent accuracy improvements over its predecessor, Nemotron Nano V2 VL, across all modalities, enabled by advances in architecture, training data and recipes. In particular, Nemotron 3 delivers leading results in real-world document understanding, long audio-video comprehension, and agentic computer use. Built on the highly efficient Nemotron 3 Nano 30B-A3B backbone, Nemotron 3 Nano Omni further incorporates innovative multimodal token-reduction techniques to deliver substantially lower inference latency and higher throughput than other models of similar size. We are releasing model checkpoints in BF16, FP8, and FP4 formats, along with portions of the training data and codebase to facilitate further research and development.

cs.LG↗

arXiv2025

Automotive Radar Multi-Frame Track-Before-Detect Algorithm Considering Self-Positioning Errors

This paper presents a method for the joint detection and tracking of weak targets in automotive radars using the multi-frame track-before-detect (MF-TBD) procedure. Generally, target tracking in automotive radars is challenging due to radar field of view (FOV) misalignment, nonlinear coordinate conversion, and self-positioning errors of the ego-vehicle, which are caused by platform motion. These issues significantly hinder the implementation of MF-TBD in automotive radars. To address these challenges, a new MF-TBD detection architecture is first proposed. It can adaptively adjust the detection threshold value based on the existence of moving targets within the radar FOV. Since the implementation of MF-TBD necessitates the inclusion of position, velocity, and yaw angle information of the ego-vehicle, each with varying degrees of measurement error, we further propose a multi-frame energy integration strategy for moving-platform radar and accurately derive the target energy integration path functions. The self-positioning errors of the ego-vehicle, which are usually not considered in some previous target tracking approaches, are well addressed. Numerical simulations and experimental results with real radar data demonstrate large detection and tracking gains over standard automotive radar processing in weak target environments.

eess.SP↗

arXiv2025

V2P-Bench: Evaluating Video-Language Understanding with Visual Prompts for Better Human-Model Interaction

Large Vision-Language Models (LVLMs) have made significant strides in the field of video understanding in recent times. Nevertheless, existing video benchmarks predominantly rely on text prompts for evaluation, which often require complex referential language and diminish both the accuracy and efficiency of human model interaction in turn. To address this limitation, we propose V2P-Bench, a robust and comprehensive benchmark for evaluating the ability of LVLMs to understand Video Visual Prompts in human model interaction scenarios. V2P-Bench consists of 980 videos and 1172 well-structured high-quality QA pairs, each paired with manually annotated visual prompt frames. The benchmark spans three main tasks and twelve categories, thereby enabling fine-grained, instance-level evaluation. Through an in-depth analysis of current LVLMs, we identify several key findings: 1) Visual prompts are both more model-friendly and user-friendly in interactive scenarios than text prompts, leading to significantly improved model performance and enhanced user experience. 2) Models are reasonably capable of zero-shot understanding of visual prompts, but struggle with spatiotemporal understanding. Even o1 achieves only 71.8%, far below the human expert score of 88.3%, while most open-source models perform below 60%. 3) LVLMs exhibit pervasive Hack Phenomena in video question answering tasks, which become more pronounced as video length increases and frame sampling density decreases, thereby inflating performance scores artificially. We anticipate that V2P-Bench will not only shed light on these challenges but also serve as a foundational tool for advancing human model interaction and improving the evaluation of video understanding.

cs.CV↗

arXiv2023

Learning with Noisy Labels Using Collaborative Sample Selection and Contrastive Semi-Supervised Learning

Learning with noisy labels (LNL) has been extensively studied, with existing approaches typically following a framework that alternates between clean sample selection and semi-supervised learning (SSL). However, this approach has a limitation: the clean set selected by the Deep Neural Network (DNN) classifier, trained through self-training, inevitably contains noisy samples. This mixture of clean and noisy samples leads to misguidance in DNN training during SSL, resulting in impaired generalization performance due to confirmation bias caused by error accumulation in sample selection. To address this issue, we propose a method called Collaborative Sample Selection (CSS), which leverages the large-scale pre-trained model CLIP. CSS aims to remove the mixed noisy samples from the identified clean set. We achieve this by training a 2-Dimensional Gaussian Mixture Model (2D-GMM) that combines the probabilities from CLIP with the predictions from the DNN classifier. To further enhance the adaptation of CLIP to LNL, we introduce a co-training mechanism with a contrastive loss in semi-supervised learning. This allows us to jointly train the prompt of CLIP and the DNN classifier, resulting in improved feature representation, boosted classification performance of DNNs, and reciprocal benefits to our Collaborative Sample Selection. By incorporating auxiliary information from CLIP and utilizing prompt fine-tuning, we effectively eliminate noisy samples from the clean set and mitigate confirmation bias during training. Experimental results on multiple benchmark datasets demonstrate the effectiveness of our proposed method in comparison with the state-of-the-art approaches.

cs.CV↗

arXiv2018

Estimating Distances via Received Signal Strength and Connectivity in Wireless Sensor Networks

Distance estimation is vital for localization and many other applications in wireless sensor networks (WSNs). Particularly, it is desirable to implement distance estimation as well as localization without using specific hardware in low-cost WSNs. As such, both the received signal strength (RSS) based approach and the connectivity based approach have gained much attention. The RSS based approach is suitable for estimating short distances, whereas the connectivity based approach obtains relatively good performance for estimating long distances. Considering the complementary features of these two approaches, we propose a fusion method based on the maximum-likelihood estimator (MLE) to estimate the distance between any pair of neighboring nodes in a WSN through efficiently fusing the information from the RSS and local connectivity. Additionally, the method is reported under the practical log-normal shadowing model, and the associated Cramer-Rao lower bound (CRLB) is also derived for performance analysis. Both simulations and experiments based on practical measurements are carried out, and demonstrate that the proposed method outperforms any single approach and approaches to the CRLB as well.

eess.SP↗

arXiv2015

Dual condensates at finite isospin chemical potential

The dual observables as order parameters for center symmetry are tested at finite isospin chemical potential $μ_I$ in a Polyakov-loop enhanced chiral model of QCD with physical quark masses. As a counterpart of the dressed Polyakov-loop, the first Fourier moment of pion condensate is introduced for $μ_I>{m_π}/{2}$ under the temporal twisted boundary conditions for quarks. We demonstrate that this dual condensate exhibits the similar temperature dependence as the conventional Polyakov-loop. We confirm that its rapid increase with $T$ is driven by the evaporating of pion condensation. On the other hand, the dressed Polyakov-loop shows abnormal thermal behavior, which even decreases with $T$ at low temperatures due to the influence of pion condensate. We thus argue that in QCD the critical temperature extracting from a dual observable may have nothing to do with the quark confinement-deconfinement transition if the quark mass is very small.

hep-ph↗