arXiv subjects

Qiang Lu

Publications and source records attributed to Qiang Lu.

At least 19 recordsLinked to original sources

arXiv2026

PredErase: Training-Free Object-and-Effect Removal with Predictive Latent Guidance

Removing an object is not the same as filling its mask. Cast shadows and contact shading usually lie outside the user-provided instance mask M_obj, so a frozen Fill model that edits only that mask leaves the object's photometric footprint on nearby surfaces. Supervised removers learn this joint erasure from paired clean plates. Training-free editors freeze pretrained weights, yet most still treat M_obj as the entire editable support and steer sampling with CLIP or DINO energies that do not predict the occluded scene. We present PredErase, a training-free inference procedure on frozen FLUX.2 and I-JEPA. The method separates where Fill may rewrite pixels from what structure should occupy the hole. A contact-band expansion M_flux of M_obj exposes local residuals on the supporting plane. I-JEPA, pretrained for masked token prediction, supplies a context-conditioned hole target in representation space; sparse projected gradients align decoded Fill completions with that target inside the instance, while coordinates outside the packed support stay locked. Under instance-only masks on RemovalBench, RORD-Val, and DEFACTO-Val, PredErase improves the native FLUX.2 backbone. Supervised removers remain stronger on several full-image appearance metrics; the supported claim is training-free object-and-effect editing of frozen Fill, not replacement of paired-data erasers.

cs.CV↗

arXiv2026

DDR-Net: Haze-Aware Dual-Domain Refinement for Single-Image Dehazing

Single-image dehazing aims to recover clear scenes from haze-degraded images. It remains challenging due to the atmospheric scattering and the complexity of real-world haze distributions. Although recent end-to-end networks have achieved promising performance, two issues still limit their effectiveness: insufficient feature refinement at the bottleneck stage and weak local structural representation in encoder-decoder architectures. Thus, we propose a Haze-Aware Dual-Domain Refinement Network (DDR-Net) for single-image dehazing. Our method is built upon three modules: Haze Prior Extractor (HPE) provides multi-scale haze-aware priors by operating directly on downsampled hazy images; Detail-Enhanced Blocks (DE Blocks) serve as the core feature extraction units, capturing multi-scale structural information and enhancing edge and texture recovery via gradient-aware convolutions; and Spatial-Frequency Bottleneck Refinement (SFBR) at the bottleneck jointly exploits spatial and frequency information to refine bottleneck features. DDR-Net achieves more effective feature representation and reconstruction for haze removal. Extensive experiments on real-world benchmarks demonstrate that our method outperforms existing dehazing approaches. It achieves competitive performance on synthetic datasets.

cs.CV↗

arXiv2026

Beyond Scene Priors: Fine-Grained Traffic Scene Reasoning with Benchmarking and Query-Guided Small-Object Focus

In safety-critical traffic scenarios, answering complex questions relies on minute, localized visual cues. However, standard Multimodal Large Language Models (MLLMs) tend to over-attend to backgrounds, overwhelming crucial small objects during visual-language alignment, a failure mode we term 'critical evidence dilution.' Furthermore, existing visual question answering (VQA) datasets rarely expose this flaw, as they lack large-scale, distractor-heavy evaluations that require pinpointing local evidence. To bridge this evaluation and architecture gap, we introduce the Fine-Grained Traffic Reasoning Benchmark (FGTR-Bench) and the Text-Guided Small-Object Reasoning MLLM (TSR-MLLM). FGTR-Bench comprises 40,236 single-image Multiple-Choice Questions (MCQs) created via multi-agent generation, consistency checks, and expert audits, alongside a disjoint 4,947-sample blind test split. To resolve evidence dilution, TSR-MLLM, built on Qwen3-VL-4B, uses a query-conditioned Text-Guided Small-Object Focus (TG-SOF) map. Applied once at the decoder boundary, the map adds sparse Top-K gated residuals to the most question-relevant vision slots while leaving text tokens unchanged. Together with lightweight decoder adaptation, TSR-MLLM preserves single-pass inference without external detectors or image re-encoding. Under matched settings, TSR-MLLM outperforms the strongest 4B baseline by 2.1 points on FGTR-Bench (74.1% overall), with larger gains on evidence-local tracks. Furthermore, it remains competitive on DriveQA-V (CARLA Signs) under greedy decoding without task-specific fine-tuning.

cs.CV↗

arXiv2025

End-to-End Multi-Modal Diffusion Mamba

Current end-to-end multi-modal models utilize different encoders and decoders to process input and output information. This separation hinders the joint representation learning of various modalities. To unify multi-modal processing, we propose a novel architecture called MDM (Multi-modal Diffusion Mamba). MDM utilizes a Mamba-based multi-step selection diffusion model to progressively generate and refine modality-specific information through a unified variational autoencoder for both encoding and decoding. This innovative approach allows MDM to achieve superior performance when processing high-dimensional data, particularly in generating high-resolution images and extended text sequences simultaneously. Our evaluations in areas such as image generation, image captioning, visual question answering, text comprehension, and reasoning tasks demonstrate that MDM significantly outperforms existing end-to-end models (MonoFormer, LlamaGen, and Chameleon etc.) and competes effectively with SOTA models like GPT-4V, Gemini Pro, and Mistral. Our results validate MDM's effectiveness in unifying multi-modal processes while maintaining computational efficiency, establishing a new direction for end-to-end multi-modal architectures.

cs.CV↗

arXiv2025

Traffic-MLLM: Curiosity-Regularized Supervised Learning for Traffic Scenario Case-Based Reasoning

For safe and robust autonomous driving, decision-making systems must effectively leverage past experiences to handle the inherent long-tail of traffic scenarios. Case-Based Reasoning (CBR) provides a natural paradigm for this by adapting solutions from prior cases. However, in complex and dynamic traffic environments, traditional CBR methods struggle to effectively abstract and adapt knowledge under uncertainty. Meanwhile, although multimodal large language models (MLLMs) exhibit strong perceptual and linguistic capabilities, their reasoning behavior often relies on empirical pattern fitting, limiting robustness under distribution shift and long-tail scenarios. We propose Traffic-MLLM, a retrieval-free neural case modeling framework for multimodal traffic reasoning. Instead of performing explicit case retrieval at inference time, Traffic-MLLM learns a structured and generalizable case space directly during training. To support this learning process, we construct a multi-source case base by integrating dynamic traffic videos and large-scale static visual question-answering data, serving as a unified training substrate for learning structured case representations. To further improve representation quality near knowledge boundaries, we introduce a curiosity-driven refinement mechanism based on Random Network Distillation (RND), encouraging the model to internalize cross-case structural regularities rather than surface correlations. Experiments on the SUTD-TrafficQA and DriveQA benchmarks demonstrate consistent improvements in dynamic reasoning, regulatory understanding, and cross-domain transfer. Traffic-MLLM achieves 50.8% accuracy on SUTD-TrafficQA, 74.8% on the CARLA-based DriveQA split, and 83.1% on the real-world Mapillary split, indicating that representation-level case-space refinement provides an effective alternative to explicit retrieval for scalable multimodal case adaptation.

cs.CV↗

arXiv2025

Contrastive Learning-Driven Traffic Sign Perception: Multi-Modal Fusion of Text and Vision

Traffic sign recognition, as a core component of autonomous driving perception systems, directly influences vehicle environmental awareness and driving safety. Current technologies face two significant challenges: first, the traffic sign dataset exhibits a pronounced long-tail distribution, resulting in a substantial decline in recognition performance of traditional convolutional networks when processing low-frequency and out-of-distribution classes; second, traffic signs in real-world scenarios are predominantly small targets with significant scale variations, making it difficult to extract multi-scale features.To overcome these issues, we propose a novel two-stage framework combining open-vocabulary detection and cross-modal learning. For traffic sign detection, our NanoVerse YOLO model integrates a reparameterizable vision-language path aggregation network (RepVL-PAN) and an SPD-Conv module to specifically enhance feature extraction for small, multi-scale targets. For traffic sign classification, we designed a Traffic Sign Recognition Multimodal Contrastive Learning model (TSR-MCL). By contrasting visual features from a Vision Transformer with semantic features from a rule-based BERT, TSR-MCL learns robust, frequency-independent representations, effectively mitigating class confusion caused by data imbalance. On the TT100K dataset, our method achieves a state-of-the-art 78.4% mAP in the long-tail detection task for all-class recognition. The model also obtains 91.8% accuracy and 88.9% recall, significantly outperforming mainstream algorithms and demonstrating superior accuracy and generalization in complex, open-world scenarios.

cs.CV↗

arXiv2025

Astra: Toward General-Purpose Mobile Robots via Hierarchical Multimodal Learning

Modern robot navigation systems encounter difficulties in diverse and complex indoor environments. Traditional approaches rely on multiple modules with small models or rule-based systems and thus lack adaptability to new environments. To address this, we developed Astra, a comprehensive dual-model architecture, Astra-Global and Astra-Local, for mobile robot navigation. Astra-Global, a multimodal LLM, processes vision and language inputs to perform self and goal localization using a hybrid topological-semantic graph as the global map, and outperforms traditional visual place recognition methods. Astra-Local, a multitask network, handles local path planning and odometry estimation. Its 4D spatial-temporal encoder, trained through self-supervised learning, generates robust 4D features for downstream tasks. The planning head utilizes flow matching and a novel masked ESDF loss to minimize collision risks for generating local trajectories, and the odometry head integrates multi-sensor inputs via a transformer encoder to predict the relative pose of the robot. Deployed on real in-house mobile robots, Astra achieves high end-to-end mission success rate across diverse indoor environments.

cs.RO↗

arXiv2024

Improved PCRLB for radar tracking in clutter with geometry-dependent target measurement uncertainty and application to radar trajectory control

In realistic radar tracking, target measurement uncertainty (TMU) in terms of both detection probability and measurement error covariance is significantly affected by the target-to-radar (T2R) geometry. However, existing posterior Cramer-Rao Lower Bounds (PCRLBs) rarely investigate the fundamental impact of T2R geometry on target measurement uncertainty and eventually on mean square error (MSE) of state estimate, inevitably resulting in over-conservative lower bound. To address this issue, this paper firstly derives the generalized model of target measurement error covariance for bistatic radar with moving receiver and transmitter illuminating any type of signal, along with its approximated solution to specify the impact of T2R geometry on error covariance. Based upon formulated TMU model, an improved PCRLB (IPCRLB) fully accounting for both measurement origin uncertainty and geometry-dependent TMU is then re-derived, both detection probability and measurement error covariance are treated as state-dependent parameters when differentiating log-likelihood with respect to target state. Compared to existing PCRLBs that partially or completely ignore the dependence of target measurement uncertainty on T2R geometry, proposed IPCRLB provides a much accurate (less-conservative) lower bound for radar tracking in clutter with geometry-dependent TMU. The new bound is then applied to radar trajectory control to effectively optimize T2R geometry and exhibits least uncertainty of acquired target measurement and more accurate state estimate for bistatic radar tracking in clutter, compared to state-of-the-art trajectory control methods.

eess.SP↗

arXiv2024

Prove Symbolic Regression is NP-hard by Symbol Graph

Symbolic regression (SR) is the task of discovering a symbolic expression that fits a given data set from the space of mathematical expressions. Despite the abundance of research surrounding the SR problem, there's a scarcity of works that confirm its NP-hard nature. Therefore, this paper introduces the concept of a symbol graph as a comprehensive representation of the entire mathematical expression space, effectively illustrating the NP-hard characteristics of the SR problem. Leveraging the symbol graph, we establish a connection between the SR problem and the task of identifying an optimally fitted degree-constrained Steiner Arborescence (DCSAP). The complexity of DCSAP, which is proven to be NP-hard, directly implies the NP-hard nature of the SR problem.

cs.CC↗

arXiv2022

Taylor Genetic Programming for Symbolic Regression

Genetic programming (GP) is a commonly used approach to solve symbolic regression (SR) problems. Compared with the machine learning or deep learning methods that depend on the pre-defined model and the training dataset for solving SR problems, GP is more focused on finding the solution in a search space. Although GP has good performance on large-scale benchmarks, it randomly transforms individuals to search results without taking advantage of the characteristics of the dataset. So, the search process of GP is usually slow, and the final results could be unstable.To guide GP by these characteristics, we propose a new method for SR, called Taylor genetic programming (TaylorGP) (Code and appendix at https://kgae-cup.github.io/TaylorGP/). TaylorGP leverages a Taylor polynomial to approximate the symbolic equation that fits the dataset. It also utilizes the Taylor polynomial to extract the features of the symbolic equation: low order polynomial discrimination, variable separability, boundary, monotonic, and parity. GP is enhanced by these Taylor polynomial techniques. Experiments are conducted on three kinds of benchmarks: classical SR, machine learning, and physics. The experimental results show that TaylorGP not only has higher accuracy than the nine baseline methods, but also is faster in finding stable results.

cs.NE↗

arXiv2022

Exploring Hidden Semantics in Neural Networks with Symbolic Regression

Many recent studies focus on developing mechanisms to explain the black-box behaviors of neural networks (NNs). However, little work has been done to extract the potential hidden semantics (mathematical representation) of a neural network. A succinct and explicit mathematical representation of a NN model could improve the understanding and interpretation of its behaviors. To address this need, we propose a novel symbolic regression method for neural works (called SRNet) to discover the mathematical expressions of a NN. SRNet creates a Cartesian genetic programming (NNCGP) to represent the hidden semantics of a single layer in a NN. It then leverages a multi-chromosome NNCGP to represent hidden semantics of all layers of the NN. The method uses a (1+$\lambda$) evolutionary strategy (called MNNCGP-ES) to extract the final mathematical expressions of all layers in the NN. Experiments on 12 symbolic regression benchmarks and 5 classification benchmarks show that SRNet not only can reveal the complex relationships between each layer of a NN but also can extract the mathematical representation of the whole NN. Compared with LIME and MAPLE, SRNet has higher interpolation accuracy and trends to approximate the real model on the practical dataset.

cs.LG↗

arXiv2020

Formal Scenario-Based Testing of Autonomous Vehicles: From Simulation to the Real World

We present a new approach to automated scenario-based testing of the safety of autonomous vehicles, especially those using advanced artificial intelligence-based components, spanning both simulation-based evaluation as well as testing in the real world. Our approach is based on formal methods, combining formal specification of scenarios and safety properties, algorithmic test case generation using formal simulation, test case selection for track testing, executing test cases on the track, and analyzing the resulting data. Experiments with a real autonomous vehicle at an industrial testing facility support our hypotheses that (i) formal simulation can be effective at identifying test cases to run on the track, and (ii) the gap between simulated and real worlds can be systematically evaluated and bridged.

eess.SY↗

arXiv2020

LGSVL Simulator: A High Fidelity Simulator for Autonomous Driving

Testing autonomous driving algorithms on real autonomous vehicles is extremely costly and many researchers and developers in the field cannot afford a real car and the corresponding sensors. Although several free and open-source autonomous driving stacks, such as Autoware and Apollo are available, choices of open-source simulators to use with them are limited. In this paper, we introduce the LGSVL Simulator which is a high fidelity simulator for autonomous driving. The simulator engine provides end-to-end, full-stack simulation which is ready to be hooked up to Autoware and Apollo. In addition, simulator tools are provided with the core simulation engine which allow users to easily customize sensors, create new types of controllable objects, replace some modules in the core simulator, and create digital twins of particular environments.

cs.RO↗

arXiv2018

Safe and Efficient Intersection Control of Connected and Autonomous Intersection Traffic

In this dissertation, we address a problem of safe and efficient intersection crossing traffic management of autonomous and connected ground traffic. Toward this objective, an algorithm that is called the Discrete-time occupancies trajectory based Intersection traffic Coordination Algorithm (DICA) is proposed. All vehicles in the system are Connected and Autonomous Vehicles (CAVs) and capable of wireless Vehicle-to-Intersection communication. In the proposed framework, an intersection coordinates the motions of CAVs based on their proposed DTOTs to let them cross the intersection efficiently while avoiding collisions. In case when there is a collision between vehicles' DTOTs, the intersection modifies conflicting DTOTs to avoid the collision and requests CAVs to approach and cross the intersection according to the modified DTOTs. We then prove that the basic DICA is deadlock free and also starvation free. We also show that the basic DICA is conservative in computational complexity and improve it by several computational approaches. Next, we addressed the problem of evacuating emergency vehicles as quickly as possible through autonomous and connected intersection traffic in this dissertation. The proposed intersection control algorithm Reactive DICA aims to determine an efficient vehicle-passing sequence which allows the emergency vehicle to cross an intersection as soon as possible while the travel times of other vehicles are minimally affected. When there are no emergency vehicles within the intersection area, the vehicles are controlled by DICA. When there are emergency vehicles entering communication range, we prioritize emergency vehicles through optimal ordering of vehicles. A genetic algorithm is proposed to solve the optimization problem which finds the optimal vehicle sequence that gives the emergency vehicles the highest priority.

eess.SY↗

arXiv2017

Paving the Way to Smart Micro Energy Internet: Concepts, Design Principles, and Engineering Practices

The energy internet is one of the most promising future energy infrastructures that could both enhance energy efficiency and improve its operating flexibility. Analogous to the micro-grid, the micro energy internet emphasizes the distribution level and demand side. This paper proposes concepts and design principles of a smart micro energy internet for accommodating micro-grids, distributed poly-generation systems, energy storage facilities, and associated energy distribution infrastructures. Since the dispatch and control system of the smart micro energy internet is responsible for external disturbances, it should be able to approach a satisfactory operating point while supporting multiple criteria, such as safety, economy, and environmental protection. To realize the vision of a smart micro energy internet, an engineering game theory based energy management system with self-approaching-optimum capability is investigated. Based on the proposed concepts, design principles, and energy management system, this paper presents a prototype of China's first conceptual solar-based smart micro energy internet, established in Qinghai University.

math.OC↗

arXiv2017

Autonomous and Connected Intersection Crossing Traffic Management using Discrete-Time Occupancies Trajectory

In this paper, we address a problem of safe and efficient intersection crossing traffic management of autonomous and connected ground traffic. Toward this objective, we propose an algorithm that is called the Discrete-time occupancies trajectory based Intersection traffic Coordination Algorithm (DICA). We first prove that the basic DICA is deadlock free and also starvation free. Then, we show that the basic DICA has a computational complexity of $\mathcal{O}(n^2 L_m^3)$ where $n$ is the number of vehicles granted to cross an intersection and $L_m$ is the maximum length of intersection crossing routes. To improve the overall computational efficiency of the algorithm, the basic DICA is enhanced by several computational approaches that are proposed in this paper. The enhanced algorithm has the computational complexity of $\mathcal{O}(n^2 L_m \log_2 L_m)$. The improved computational efficiency of the enhanced algorithm is validated through simulation using an open source traffic simulator, called the Simulation of Urban MObility (SUMO). The overall throughput as well as the computational efficiency of the enhanced algorithm are also compared with those of an optimized traffic light control.

eess.SY↗

arXiv2015

Fracture resistance of zigzag single walled carbon nanotubes

Brittle fracture is one of the important failure modes of Single-Walled Carbon Nanotube (SWNT) due to mechanical loading. In this paper, the fracture resistance of zigzag SWNTs with preexisting defects is calculated using fracture mechanics concepts based on atomistic simulations. The problem of unstable crack growth at finite temperature, presumably caused by lattice trapping effect, is circumvented by computing the strain energy release rate through a series of displacement-controlled tensile loading of SWNTs (applied through moving the outermost layer of atoms at one end at constant strain rate of 9.4x10-4/ps) with pre-existing crack-like defects of various lengths. The strain energy release rate, G, is computed for (17,0), (28,0) and (35,0) SWNTs (each with aspect ratio 4) with pre-existing cracks up to 29.5Å long. The fracture resistance, Gc, is determined as a function of crack length for each tube at three different temperatures (1K, 300K and 500K). A significant dependence of Gc on crack length is observed reminiscent of the rising R curve behavior of metals at the macroscale: for the zigzag nanotubes Gc increases with crack length at small length, and tends to reach a constant value if the tube diameter is large enough. We suspect that the lattice trapping effect plays the role of crack tip plasticity at the atomic scale. For example, at 300 Kelvin, Gc for the (35,0) tube with aspect ratio 4 converges to 6 Joule/m2 as the crack length exceeds 20 Angstrom. This value is comparable with the fracture toughness of graphite and Silicon. The fracture resistance of the tubes is found to decrease significantly as the temperature increases. To study the length effects, the computations are repeated for zigzag nanotubes with the same three chiralities but with aspect ratio 8 at 1K.

cond-mat.mtrl-sci↗

arXiv2015

Effect of randomly occurring Stone-Wales defects on mechanical properties of carbon nanotubes using atomistic simulation

While CNTs are found to have ultra high stiffness and strength, an enormous scatter is also observed in available laboratory results. This paper studies the effects of randomly distributed Stone Wales (SW or 5 7 7 5) defects on the mechanical properties of single walled nanotubes (SWNTs) using the technique of atomistic simulation (AS). A Matern hard core random field applied on a finite cylindrical surface is used to describe the spatial distribution of the Stone Wales defects. We simulate a set of displacement controlled tensile loading up to fracture of SWNTs with (6,6) armchair and (10, 0) zigzag configurations and aspect ratio around 6. A modified Morse potential is adopted to model the interatomic forces. We found that fracture invariably initiates from a defect if one is present; for a defect-free tube the crack initiates at quite random locations. The force-displacement curve typically behaves almost linearly up to about half way, although there is no obvious yield point. Three mechanical properties - stiffness, ultimate strength and ultimate strain - are calculated from the simulated force and displacement time histories. The randomness in mechanical behavior resulting only from initial velocity distribution was found to be insignificant at room temperature. The mean values of stiffness, ultimate strength and ultimate strain of the tube decrease as the average number of defects increases although the coefficients of variation do not show such monotonic trend. The introduction of an additional defect has the most pronounced effect on the randomness in mechanical properties when the tube is originally defect free. We also found that, for a given mean number of defects in the tube, the zigzag configuration has less strength and less ultimate strain on the average, but more uncertainty in its stiffness and ultimate strain, compared with the armchair tube.

cond-mat.mtrl-sci↗