Search arXiv⌕ Search

arXiv subjects

Qian Zhang

Publications and source records attributed to Qian Zhang.

At least 181 records · Page 10Linked to original sources
arXiv2025

Alignment between Brains and AI: Evidence for Convergent Evolution across Modalities, Scales and Training Trajectories

Artificial and biological systems may evolve similar computational solutions despite fundamental differences in architecture and learning mechanisms -- a form of convergent evolution. We demonstrate this phenomenon through large-scale analysis of alignment between human brain activity and internal representations of over 600 AI models spanning language and vision domains, from 1.33M to 72B parameters. Analyzing 60 million alignment measurements reveals that higher-performing models spontaneously develop stronger brain alignment without explicit neural constraints, with language models showing markedly stronger correlation (r=0.89, p<7.5e-13) than vision models (r=0.53, p<2.0e-44). Crucially, longitudinal analysis demonstrates that brain alignment consistently precedes performance improvements during training, suggesting that developing brain-like representations may be a necessary stepping stone toward higher capabilities. We find systematic patterns: language models exhibit strongest alignment with limbic and integrative regions, while vision models show progressive alignment with visual cortices; deeper processing layers converge across modalities; and as representational scale increases, alignment systematically shifts from primary sensory to higher-order associative regions. These findings provide compelling evidence that optimization for task performance naturally drives AI systems toward brain-like computational strategies, offering both fundamental insights into principles of intelligent information processing and practical guidance for developing more capable AI systems.

q-bio.NC↗
arXiv2025

ScEdit: Script-based Assessment of Knowledge Editing

Knowledge Editing (KE) has gained increasing attention, yet current KE tasks remain relatively simple. Under current evaluation frameworks, many editing methods achieve exceptionally high scores, sometimes nearing perfection. However, few studies integrate KE into real-world application scenarios (e.g., recent interest in LLM-as-agent). To support our analysis, we introduce a novel script-based benchmark -- ScEdit (Script-based Knowledge Editing Benchmark) -- which encompasses both counterfactual and temporal edits. We integrate token-level and text-level evaluation methods, comprehensively analyzing existing KE techniques. The benchmark extends traditional fact-based ("What"-type question) evaluation to action-based ("How"-type question) evaluation. We observe that all KE methods exhibit a drop in performance on established metrics and face challenges on text-level metrics, indicating a challenging task. Our benchmark is available at https://github.com/asdfo123/ScEdit.

cs.CL↗
arXiv2025

A novel parameter-free and locking-free enriched Galerkin method for linear elasticity

We propose a novel parameter-free and locking-free enriched Galerkin (EG) method for solving the linear elasticity problem in both two and three dimensions. Unlike existing locking-free EG methods, our method enriches the first-order continuous Galerkin (CG) space with piecewise constants along edges in two dimensions or faces in three dimensions. This enrichment acts as a correction to the normal component of the CG space, ensuring the locking-free property and delivering an oscillation-free stress approximation without requiring post-processing. Our theoretical analysis establishes the well-posedness of the method and derives optimal error estimates. Numerical experiments further demonstrate the accuracy, efficiency, and robustness of the proposed method.

math.NA↗
arXiv2025

4,500 Seconds: Small Data Training Approaches for Deep UAV Audio Classification

Unmanned aerial vehicle (UAV) usage is expected to surge in the coming decade, raising the need for heightened security measures to prevent airspace violations and security threats. This study investigates deep learning approaches to UAV classification focusing on the key issue of data scarcity. To investigate this we opted to train the models using a total of 4,500 seconds of audio samples, evenly distributed across a 9-class dataset. We leveraged parameter efficient fine-tuning (PEFT) and data augmentations to mitigate the data scarcity. This paper implements and compares the use of convolutional neural networks (CNNs) and attention-based transformers. Our results show that, CNNs outperform transformers by 1-2\% accuracy, while still being more computationally efficient. These early findings, however, point to potential in using transformers models; suggesting that with more data and further optimizations they could outperform CNNs. Future works aims to upscale the dataset to better understand the trade-offs between these approaches.

cs.SD↗
arXiv2025

On the Input-Output Monotonicity of Voltage Dynamics of Power System with Grid-Forming Converters

Integration of renewable resources is profoundly reshaping the dynamics of modern power systems. This study shows that the voltage dynamics of power systems with multiple grid-forming (GFM) converters often enjoys a desirable property called input-output monotonicity. A systematic approach for computing the derivatives of the voltage subsystem is presented first, which provides insight into the structural characteristics of these models. Next, the sign pattern of the trajectory Jacobian matrix associated with the voltage subsystem is analyzed and revealed. The analysis indicates that the voltage dynamics of power systems often exhibits the so-called input-output monotonicity property. The theoretical results are then validated through several simulation examples, underscoring their practical implications.

eess.SY↗
arXiv2025

FireFly-T: High-Throughput Sparsity Exploitation for Spiking Transformer Acceleration with Dual-Engine Overlay Architecture

Spiking transformers are emerging as a promising architecture that combines the energy efficiency of Spiking Neural Networks (SNNs) with the powerful attention mechanisms of transformers. However, existing hardware accelerators lack support for spiking attention, exhibit limited throughput in exploiting fine-grained sparsity, and struggle with scalable parallelism in sparse computation. To address these, we propose FireFly-T, a dual-engine overlay architecture that integrates a sparse engine for activation sparsity and a binary engine for spiking attention. In the sparse engine, we propose a highthroughput sparse decoder that exploits fine-grained sparsity by concurrently extracting multiple non-zero spikes. To complement this, we introduce a scalable load balancing mechanism with weight dispatch and out-of-order execution, eliminating bank conflicts to support scalable multidimensional parallelism. In the binary engine, we leverage the byte-level write capability of SRAMs to efficiently manipulate the 3D dataflows required for spiking attention with minimal resource overhead. We also optimize the core AND-PopCount operation in spiking attention through a LUT6-based implementation, improving timing closure and reducing LUT utilization on Xilinx FPGAs. As an overlay architecture, FireFly-T further incorporates an orchestrator that dynamically manipulates input dataflows with flexible adaptation for diverse network topologies, while ensuring efficient resource utilization and maintaining high throughput. Experimental results demonstrate that our accelerator achieves $1.39\times$ and $2.40\times$ higher energy efficiency, as well as $4.21\times$ and $7.10\times$ greater DSP efficiency, compared to FireFly v2 and the transformer-enabled SpikeTA, respectively. These results highlight its potential as an efficient hardware platform for spiking transformer.

cs.AR↗
arXiv2025

CityNavAgent: Aerial Vision-and-Language Navigation with Hierarchical Semantic Planning and Global Memory

Aerial vision-and-language navigation (VLN), requiring drones to interpret natural language instructions and navigate complex urban environments, emerges as a critical embodied AI challenge that bridges human-robot interaction, 3D spatial reasoning, and real-world deployment. Although existing ground VLN agents achieved notable results in indoor and outdoor settings, they struggle in aerial VLN due to the absence of predefined navigation graphs and the exponentially expanding action space in long-horizon exploration. In this work, we propose \textbf{CityNavAgent}, a large language model (LLM)-empowered agent that significantly reduces the navigation complexity for urban aerial VLN. Specifically, we design a hierarchical semantic planning module (HSPM) that decomposes the long-horizon task into sub-goals with different semantic levels. The agent reaches the target progressively by achieving sub-goals with different capacities of the LLM. Additionally, a global memory module storing historical trajectories into a topological graph is developed to simplify navigation for visited targets. Extensive benchmark experiments show that our method achieves state-of-the-art performance with significant improvement. Further experiments demonstrate the effectiveness of different modules of CityNavAgent for aerial VLN in continuous city environments. The code is available at \href{https://github.com/VinceOuti/CityNavAgent}{link}.

cs.RO↗
arXiv2025

Perceive-IR: Learning to Perceive Degradation Better for All-in-One Image Restoration

Existing All-in-One image restoration methods often fail to perceive degradation types and severity levels simultaneously, overlooking the importance of fine-grained quality perception. Moreover, these methods often utilize highly customized backbones, which hinder their adaptability and integration into more advanced restoration networks. To address these limitations, we propose Perceive-IR, a novel backbone-agnostic All-in-One image restoration framework designed for fine-grained quality control across various degradation types and severity levels. Its modular structure allows core components to function independently of specific backbones, enabling seamless integration into advanced restoration models without significant modifications. Specifically, Perceive-IR operates in two key stages: 1) multi-level quality-driven prompt learning stage, where a fine-grained quality perceiver is meticulously trained to discern three tier quality levels by optimizing the alignment between prompts and images within the CLIP perception space. This stage ensures a nuanced understanding of image quality, laying the groundwork for subsequent restoration; 2) restoration stage, where the quality perceiver is seamlessly integrated with a difficulty-adaptive perceptual loss, forming a quality-aware learning strategy. This strategy not only dynamically differentiates sample learning difficulty but also achieves fine-grained quality control by driving the restored image toward the ground truth while pulling it away from both low- and medium-quality samples.

cs.CV↗
arXiv2025

Existence and multiplicity of solutions for a critical Kirchhoff type elliptic equation with a logarithmic perturbation

In this paper, we are interested in the following critical Kirchhoff type elliptic equation with a logarithmic perturbation \begin{equation}\label{eq0} \begin{cases} -\left(1+b\int_Ω|\nabla{u}|^2\mathrm{d}x\right) Δ{u}=λu+μu\log{u^2}+|u|^{2^{*}-2}u, &x\inΩ,\\ u=0,&x\in\partialΩ, \end{cases} \end{equation} where $Ω$ is a bounded domain in $\mathbb{R}^{N}(N\geq3)$ with smooth boundary $\partial Ω$, $b$, $λ$ and $μ$ are parameters and $2^{*}=\frac{2N}{N-2}$ is the critical Sobolev exponent. The presence of a nonlocal term, together with a critical nonlinearity and a logarithmic term, prevents to apply in a straightforward way the classical critical point theory. Moreover, the geometry structure of the energy functional changes as the space dimension $N$ varies, which has a crucial influence on the existence of solutions to the problem. On the basis of some careful analysis on the structure of the energy functional, existence and (or) multiplicity results are obtained by using variational methods. More precisely, if $N=3$, problem (0.1) admits a local minimum solution, a ground state solution and a sequence of solutions with their $H_0^1(Ω)$-norms converging to $0$. If $N=4$, the existence of infinitely many solutions is also obtained. When $N\geq5$, problem (0.1) admits a local minimum solution with negative energy. Sufficient conditions are also derived for the local minimum solution to be a ground state solution.

math.AP↗
arXiv2025

Existence of nontrivial solutions to a critical Kirchhoff equation with a logarithmic type perturbation in dimension four

In this paper, a critical Kirchhoff equation with a logarithmic type subcritical term is considered in a bounded domain in $\mathbb{R}^4$. We view this problem as a critical elliptic equation with a nonlocal perturbation, and investigate how the nonlocal term affects the existence of weak solutions to the problem. By means of Ekeland's variational principle, Brézis-Lieb's lemma and some convergence tricks for nonlocal problems, we show that this problem admits a local minimum solution and a least energy solution under some appropriate assumptions on the parameters. Moreover, under some further assumptions, the local minimum solution is also a least energy solution. Compared with the ones obtained in [3] and [8], our results show that the introduction of the nonlocal term enlarges the ranges of the parameters such that the problem admits weak solutions, which implies that the nonlocal term has a positive effect on the existence of weak solutions.

math.AP↗
arXiv2025

DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving

Recently, the diffusion model has emerged as a powerful generative technique for robotic policy learning, capable of modeling multi-mode action distributions. Leveraging its capability for end-to-end autonomous driving is a promising direction. However, the numerous denoising steps in the robotic diffusion policy and the more dynamic, open-world nature of traffic scenes pose substantial challenges for generating diverse driving actions at a real-time speed. To address these challenges, we propose a novel truncated diffusion policy that incorporates prior multi-mode anchors and truncates the diffusion schedule, enabling the model to learn denoising from anchored Gaussian distribution to the multi-mode driving action distribution. Additionally, we design an efficient cascade diffusion decoder for enhanced interaction with conditional scene context. The proposed model, DiffusionDrive, demonstrates 10$\times$ reduction in denoising steps compared to vanilla diffusion policy, delivering superior diversity and quality in just 2 steps. On the planning-oriented NAVSIM dataset, with the aligned ResNet-34 backbone, DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new record, while running at a real-time speed of 45 FPS on an NVIDIA 4090. Qualitative results on challenging scenarios further confirm that DiffusionDrive can robustly generate diverse plausible driving actions. Code and model will be available at https://github.com/hustvl/DiffusionDrive.

cs.CV↗
arXiv2025

LGPS: A Lightweight GAN-Based Approach for Polyp Segmentation in Colonoscopy Images

Colorectal cancer (CRC) is a major global cause of cancer-related deaths, with early polyp detection and removal during colonoscopy being crucial for prevention. While deep learning methods have shown promise in polyp segmentation, challenges such as high computational costs, difficulty in segmenting small or low-contrast polyps, and limited generalizability across datasets persist. To address these issues, we propose LGPS, a lightweight GAN-based framework for polyp segmentation. LGPS incorporates three key innovations: (1) a MobileNetV2 backbone enhanced with modified residual blocks and Squeeze-and-Excitation (ResE) modules for efficient feature extraction; (2) Convolutional Conditional Random Fields (ConvCRF) for precise boundary refinement; and (3) a hybrid loss function combining Binary Cross-Entropy, Weighted IoU Loss, and Dice Loss to address class imbalance and enhance segmentation accuracy. LGPS is validated on five benchmark datasets and compared with state-of-the-art(SOTA) methods. On the largest and challenging PolypGen test dataset, LGPS achieves a Dice of 0.7299 and an IoU of 0.7867, outperformed all SOTA works and demonstrating robust generalization. With only 1.07 million parameters, LGPS is 17 times smaller than the smallest existing model, making it highly suitable for real-time clinical applications. Its lightweight design and strong performance underscore its potential for improving early CRC diagnosis. Code is available at https://github.com/Falmi/LGPS/.

cs.CV↗
arXiv2025

Multimodal Mamba: Decoder-only Multimodal State Space Model via Quadratic to Linear Distillation

Recent Multimodal Large Language Models (MLLMs) have achieved remarkable performance but face deployment challenges due to their quadratic computational complexity, growing Key-Value cache requirements, and reliance on separate vision encoders. We propose mmMamba, a framework for developing linear-complexity native multimodal state space models through progressive distillation from existing MLLMs using moderate academic computational resources. Our approach enables the direct conversion of trained decoder-only MLLMs to linear-complexity architectures without requiring pre-trained RNN-based LLM or vision encoders. We propose an seeding strategy to carve Mamba from trained Transformer and a three-stage distillation recipe, which can effectively transfer the knowledge from Transformer to Mamba while preserving multimodal capabilities. Our method also supports flexible hybrid architectures that combine Transformer and Mamba layers for customizable efficiency-performance trade-offs. Distilled from the Transformer-based decoder-only HoVLE, mmMamba-linear achieves competitive performance against existing linear and quadratic-complexity VLMs, while mmMamba-hybrid further improves performance significantly, approaching HoVLE's capabilities. At 103K tokens, mmMamba-linear demonstrates 20.6$\times$ speedup and 75.8% GPU memory reduction compared to HoVLE, while mmMamba-hybrid achieves 13.5$\times$ speedup and 60.2% memory savings. Code and models are released at https://github.com/hustvl/mmMamba

cs.CV↗
arXiv2025

Brain-inspired Action Generation with Spiking Transformer Diffusion Policy Model

Spiking Neural Networks (SNNs) has the ability to extract spatio-temporal features due to their spiking sequence. While previous research has primarily foucus on the classification of image and reinforcement learning. In our paper, we put forward novel diffusion policy model based on Spiking Transformer Neural Networks and Denoising Diffusion Probabilistic Model (DDPM): Spiking Transformer Modulate Diffusion Policy Model (STMDP), a new brain-inspired model for generating robot action trajectories. In order to improve the performance of this model, we develop a novel decoder module: Spiking Modulate De coder (SMD), which replaces the traditional Decoder module within the Transformer architecture. Additionally, we explored the substitution of DDPM with Denoising Diffusion Implicit Models (DDIM) in our frame work. We conducted experiments across four robotic manipulation tasks and performed ablation studies on the modulate block. Our model consistently outperforms existing Transformer-based diffusion policy method. Especially in Can task, we achieved an improvement of 8%. The proposed STMDP method integrates SNNs, dffusion model and Transformer architecture, which offers new perspectives and promising directions for exploration in brain-inspired robotics.

cs.RO↗
arXiv2025

MTDP: A Modulated Transformer based Diffusion Policy Model

Recent research on robot manipulation based on Behavior Cloning (BC) has made significant progress. By combining diffusion models with BC, diffusion policiy has been proposed, enabling robots to quickly learn manipulation tasks with high success rates. However, integrating diffusion policy with high-capacity Transformer presents challenges, traditional Transformer architectures struggle to effectively integrate guiding conditions, resulting in poor performance in manipulation tasks when using Transformer-based models. In this paper, we investigate key architectural designs of Transformers and improve the traditional Transformer architecture by proposing the Modulated Transformer Diffusion Policy (MTDP) model for diffusion policy. The core of this model is the Modulated Attention module we proposed, which more effectively integrates the guiding conditions with the main input, improving the generative model's output quality and, consequently, increasing the robot's task success rate. In six experimental tasks, MTDP outperformed existing Transformer model architectures, particularly in the Toolhang experiment, where the success rate increased by 12\%. To verify the generality of Modulated Attention, we applied it to the UNet architecture to construct Modulated UNet Diffusion Policy model (MUDP), which also achieved higher success rates than existing UNet architectures across all six experiments. The Diffusion Policy uses Denoising Diffusion Probabilistic Models (DDPM) as the diffusion model. Building on this, we also explored Denoising Diffusion Implicit Models (DDIM) as the diffusion model, constructing the MTDP-I and MUDP-I model, which nearly doubled the generation speed while maintaining performance.

cs.RO↗
arXiv2025

PassiveBLE: Towards Fully Commodity-Compatible BLE Backscatter

Bluetooth Low Energy (BLE) backscatter is a promising candidate for battery-free Internet of Things (IoT) applications. Unlike existing commodity-level BLE backscatter systems that only enable one-shot communication through BLE advertising packets, we propose PassiveBLE, a backscatter system that can establish authentic and fully compatible BLE connections on data channels. The key enabling techniques include (i) a synchronization circuit that can wake up tags and activate backscatter communications with symbol-level accuracy to facilitate BLE data packet generation; (ii) a distributed coding scheme that offloads the major encoding and processing burdens from tags to the excitation source while achieving high throughput; (iii) a BLE connection scheduler to enable fully compatible BLE connection interactions, including connection establishment, maintenance and termination for multiple backscatter tags. We prototype PassiveBLE tags with off-the-shelf components and also convert the circuits and control logic into ASIC design sketch, whose power consumptions are 491 uW and 9.9 uW, respectively. Experimental results demonstrate that PassiveBLE achieves a success rate of over 99.9% in establishing commodity BLE connections. PassiveBLE also achieves commodity-compatible BLE communication with a high goodput of up to 974 kbps in LE 2M PHY mode and 532 kbps in LE 1M PHY mode, which is about 63.3 times higher than the previous commodity-level BLE backscatter system in the same mode.

cs.NI↗