arXiv subjects

Ao Wang

Publications and source records attributed to Ao Wang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Symmetry-engineered and electrically tunable in-plane anomalous Hall effect in oxide heterostructures

The family of Hall effects has long served as a premier probe of how symmetry, magnetic order, and topology intertwine in solids. Recently, the in-plane anomalous Hall effect (IP-AHE), a transverse Hall response driven by in-plane magnetization, has emerged as a distinct member of this family, offering innovative spintronic functionalities and illuminating intricate interplay between mirror-symmetry breaking and in-plane magnetic order. However, practical routes to deterministically and reversibly control IP-AHE remain limited. Here, we establish a symmetry-engineered IP-AHE platform, CaRuO3/La2/3Ca1/3MnO3/CaRuO3 heterostructure on NdGaO3(110), that turns strict mirror-symmetry breaking constraints into effective tuning knobs. IP-AHE in these epitaxial trilayers unambiguously couples to the CaRuO3-buffer-induced mirror-symmetry breaking and faithfully reproduces the ferromagnetic hysteresis. Ionic liquid gating further enables reversible reconfigurations of the symmetry breaking, thereby achieving electrical modulation and ON/OFF switching of IP-AHE. This highly tunable IP-AHE platform opens pathways for exploring nontrivial magnetic order and developing programmable Hall functionalities in planar geometries.

cond-mat.str-el↗

arXiv2026

FaST: Efficient and Effective Long-Horizon Forecasting for Large-Scale Spatial-Temporal Graphs via Mixture-of-Experts

Spatial-Temporal Graph (STG) forecasting on large-scale networks has garnered significant attention. However, existing models predominantly focus on short-horizon predictions and suffer from notorious computational costs and memory consumption when scaling to long-horizon predictions and large graphs. Targeting the above challenges, we present FaST, an effective and efficient framework based on heterogeneity-aware Mixture-of-Experts (MoEs) for long-horizon and large-scale STG forecasting, which unlocks one-week-ahead (672 steps at a 15-minute granularity) prediction with thousands of nodes. FaST is underpinned by two key innovations. First, an adaptive graph agent attention mechanism is proposed to alleviate the computational burden inherent in conventional graph convolution and self-attention modules when applied to large-scale graphs. Second, we propose a new parallel MoE module that replaces traditional feed-forward networks with Gated Linear Units (GLUs), enabling an efficient and scalable parallel structure. Extensive experiments on real-world datasets demonstrate that FaST not only delivers superior long-horizon predictive accuracy but also achieves remarkable computational efficiency compared to state-of-the-art baselines. Our source code is available at: https://github.com/yijizhao/FaST.

cs.LG↗

arXiv2025

DevPiolt: Operation Recommendation for IoT Devices at Xiaomi Home

Operation recommendation for IoT devices refers to generating personalized device operations for users based on their context, such as historical operations, environment information, and device status. This task is crucial for enhancing user satisfaction and corporate profits. Existing recommendation models struggle with complex operation logic, diverse user preferences, and sensitive to suboptimal suggestions, limiting their applicability to IoT device operations. To address these issues, we propose DevPiolt, a LLM-based recommendation model for IoT device operations. Specifically, we first equip the LLM with fundamental domain knowledge of IoT operations via continual pre-training and multi-task fine-tuning. Then, we employ direct preference optimization to align the fine-tuned LLM with specific user preferences. Finally, we design a confidence-based exposure control mechanism to avoid negative user experiences from low-quality recommendations. Extensive experiments show that DevPiolt significantly outperforms baselines on all datasets, with an average improvement of 69.5% across all metrics. DevPiolt has been practically deployed in Xiaomi Home app for one quarter, providing daily operation recommendations to 255,000 users. Online experiment results indicate a 21.6% increase in unique visitor device coverage and a 29.1% increase in page view acceptance rates.

cs.AI↗

arXiv2025

Wavelet-Optimized Motion Artifact Correction in 3D MRI Using Pre-trained 2D Score Priors

Motion artifacts in magnetic resonance imaging (MRI) remain a major challenge, as they degrade image quality and compromise diagnostic reliability. Score-based generative models (SGMs) have recently shown promise for artifact removal. However, existing 3D SGM-based approaches are limited in two key aspects: (1) their strong dependence on known forward operators makes them ineffective for correcting MRI motion artifacts, and (2) their slow inference speed hinders clinical translation. To overcome these challenges, we propose a wavelet-optimized end-to-end framework for 3D MRI motion correct using pre-trained 2D score priors (3D-WMoCo). Specifically, two orthogonal 2D score priors are leveraged to guide the 3D distribution prior, while a mean-reverting stochastic differential equation (SDE) is employed to model the restoration process of motion-corrupted 3D volumes to motion-free 3D distribution. Furthermore, wavelet diffusion is introduced to accelerate inference, and wavelet convolution is applied to enhance feature extraction. We validate the effectiveness of our approach through both simulated motion artifact experiments and real-world clinical motion artifact correction tests. The proposed method achieves robust performance improvements over existing techniques. Implementation details and source code are available at: https://github.com/ZG-yuan/3D-WMoCo.

cs.CE↗

arXiv2025

CAIT: Triple-Win Compression towards High Accuracy, Fast Inference, and Favorable Transferability For ViTs

Vision Transformers (ViTs) have emerged as state-of-the-art models for various vision tasks recently. However, their heavy computation costs remain daunting for resource-limited devices. To address this, researchers have dedicated themselves to compressing redundant information in ViTs for acceleration. However, existing approaches generally sparsely drop redundant image tokens by token pruning or brutally remove channels by channel pruning, leading to a sub-optimal balance between model performance and inference speed. Moreover, they struggle when transferring compressed models to downstream vision tasks that require the spatial structure of images, such as semantic segmentation. To tackle these issues, we propose CAIT, a joint \underline{c}ompression method for ViTs that achieves a harmonious blend of high \underline{a}ccuracy, fast \underline{i}nference speed, and favorable \underline{t}ransferability to downstream tasks. Specifically, we introduce an asymmetric token merging (ATME) strategy to effectively integrate neighboring tokens. It can successfully compress redundant token information while preserving the spatial structure of images. On top of it, we further design a consistent dynamic channel pruning (CDCP) strategy to dynamically prune unimportant channels in ViTs. Thanks to CDCP, insignificant channels in multi-head self-attention modules of ViTs can be pruned uniformly, significantly enhancing the model compression. Extensive experiments on multiple benchmark datasets show that our proposed method can achieve state-of-the-art performance across various ViTs.

cs.CV↗

arXiv2025

PrefixKV: Adaptive Prefix KV Cache is What Vision Instruction-Following Models Need for Efficient Generation

Recently, large vision-language models (LVLMs) have rapidly gained popularity for their strong generation and reasoning capabilities given diverse multimodal inputs. However, these models incur significant computational and memory overhead during inference, which greatly hinders the efficient deployment in practical scenarios. The extensive key-value (KV) cache, necessitated by the lengthy input and output sequences, notably contributes to the high inference cost. Based on this, recent works have investigated ways to reduce the KV cache size for higher efficiency. Although effective, they generally overlook the distinct importance distributions of KV vectors across layers and maintain the same cache size for each layer during the next token prediction. This results in the significant contextual information loss for certain layers, leading to notable performance decline. To address this, we present PrefixKV, where "Prefix" means the top-ranked KV based on importance rather than position in the original sequence. It reframes the challenge of determining KV cache sizes for all layers into the task of searching for the optimal global prefix configuration. With an adaptive layer-wise KV retention recipe based on binary search, the maximum contextual information can thus be preserved in each layer, facilitating the generation. Extensive experiments demonstrate that our method achieves the state-of-the-art performance compared with others. It exhibits superior inference efficiency and generation quality trade-offs, showing promising potential for practical applications. Code is available at https://github.com/THU-MIG/PrefixKV.

cs.CV↗

arXiv2025

YOLOE: Real-Time Seeing Anything

Object detection and segmentation are widely employed in computer vision applications, yet conventional models like YOLO series, while efficient and accurate, are limited by predefined categories, hindering adaptability in open scenarios. Recent open-set methods leverage text prompts, visual cues, or prompt-free paradigm to overcome this, but often compromise between performance and efficiency due to high computational demands or deployment complexity. In this work, we introduce YOLOE, which integrates detection and segmentation across diverse open prompt mechanisms within a single highly efficient model, achieving real-time seeing anything. For text prompts, we propose Re-parameterizable Region-Text Alignment (RepRTA) strategy. It refines pretrained textual embeddings via a re-parameterizable lightweight auxiliary network and enhances visual-textual alignment with zero inference and transferring overhead. For visual prompts, we present Semantic-Activated Visual Prompt Encoder (SAVPE). It employs decoupled semantic and activation branches to bring improved visual embedding and accuracy with minimal complexity. For prompt-free scenario, we introduce Lazy Region-Prompt Contrast (LRPC) strategy. It utilizes a built-in large vocabulary and specialized embedding to identify all objects, avoiding costly language model dependency. Extensive experiments show YOLOE's exceptional zero-shot performance and transferability with high inference efficiency and low training cost. Notably, on LVIS, with 3$\times$ less training cost and 1.4$\times$ inference speedup, YOLOE-v8-S surpasses YOLO-Worldv2-S by 3.5 AP. When transferring to COCO, YOLOE-v8-L achieves 0.6 AP$^b$ and 0.4 AP$^m$ gains over closed-set YOLOv8-L with nearly 4$\times$ less training time. Code and models are available at https://github.com/THU-MIG/yoloe.

cs.CV↗

arXiv2025

AgentMental: An Interactive Multi-Agent Framework for Explainable and Adaptive Mental Health Assessment

Mental health assessment is crucial for early intervention and effective treatment, yet traditional clinician-based approaches are limited by the shortage of qualified professionals. Recent advances in artificial intelligence have sparked growing interest in automated psychological assessment, yet most existing approaches are constrained by their reliance on static text analysis, limiting their ability to capture deeper and more informative insights that emerge through dynamic interaction and iterative questioning. Therefore, in this paper, we propose a multi-agent framework for mental health evaluation that simulates clinical doctor-patient dialogues, with specialized agents assigned to questioning, adequacy evaluation, scoring, and updating. We introduce an adaptive questioning mechanism in which an evaluation agent assesses the adequacy of user responses to determine the necessity of generating targeted follow-up queries to address ambiguity and missing information. Additionally, we employ a tree-structured memory in which the root node encodes the user's basic information, while child nodes (e.g., topic and statement) organize key information according to distinct symptom categories and interaction turns. This memory is dynamically updated throughout the interaction to reduce redundant questioning and further enhance the information extraction and contextual tracking capabilities. Experimental results on the DAIC-WOZ dataset illustrate the effectiveness of our proposed method, which achieves better performance than existing approaches.

cs.CL↗

arXiv2025

Torpor: GPU-Enabled Serverless Computing for Low-Latency, Resource-Efficient Inference

Serverless computing offers a compelling cloud model for online inference services. However, existing serverless platforms lack efficient support for GPUs, hindering their ability to deliver high-performance inference. In this paper, we present Torpor, a serverless platform for GPU-efficient, low-latency inference. To enable efficient sharing of a node's GPUs among numerous inference functions, Torpor maintains models in main memory and dynamically swaps them onto GPUs upon request arrivals (i.e., late binding with model swapping). Torpor uses various techniques, including asynchronous API redirection, GPU runtime sharing, pipelined model execution, and efficient GPU memory management, to minimize latency overhead caused by model swapping. Additionally, we design an interference-aware request scheduling algorithm that utilizes high-speed GPU interconnects to meet latency service-level objectives (SLOs) for individual inference functions. We have implemented Torpor and evaluated its performance in a production environment. Utilizing late binding and model swapping, Torpor can concurrently serve hundreds of inference functions on a worker node with 4 GPUs, while achieving latency performance comparable to native execution, where each model is cached exclusively on a GPU. Pilot deployment in a leading commercial serverless cloud shows that Torpor reduces the GPU provisioning cost by 70% and 65% for users and the platform, respectively.

cs.DC↗

arXiv2025

Lattice thermal conductivity of 16 elemental metals from molecular dynamics simulations with a unified neuroevolution potential

Metals play a crucial role in heat management in electronic devices, such as integrated circuits, making it vital to understand heat transport in elementary metals and alloys. In this work, we systematically study phonon thermal transport in 16 metals using the efficient homogeneous nonequilibrium molecular dynamics (HNEMD) method and the recently developed unified neuroevolution potential version 1 (UNEP-v1) for 16 metals and their alloys. We compare our results with existing ones based on the Boltzmann transport equation (BTE) approach and find that our HNEMD results align well with BTE results obtained by considering phonon-phonon scattering only. By contrast, HNEMD results based on the conventional embedded-atom method potential show less satisfactory agreement with BTE ones. Given the high accuracy of the UNEP-v1 model demonstrated in various metal alloys, we anticipate that the HNEMD method combined with the UNEP-v1 model will be a promising tool for exploring phonon thermal transport properties in complex systems such as high-entropy alloys.

cond-mat.mtrl-sci↗

arXiv2025

Photometric analysis of asteroids in the Phocaea region

The Phocaea asteroid family, one of the large ancient families located in the inner main belt, may be the sources of near-Earth asteroids (NEAs) due to the nearby 3:1 mean motion resonance with Jupiter, the v6 secular resonance, and the Yarkovsky and YORP effects. Thus, understanding the influence of the Yarkovsky and YORP effects on the Phocaea family is one of the keys to figuring out the source of NEAs. However, the physical properties of most of the Phocaea family members are unknown at present. We perform a photometric analysis for 44 asteroids in the Phocaea region using photometric data obtained by ground-based and space-based telescopes (i.e., TESS and Gaia). Based on the derived physical properties, we find significant footprints of the Yarkovsky and YORP effects on the Phocaea family members. Selecting five asteroids nearby the inside boundary of the V-shape in the absolute-magnitude semimajor-axis (H, a) space, we estimate their densities considering their migration in semimajor-axis arises from the Yarkovsky effect. The bulk density of (852) Wladilena ({3.54 g/cm3) suggests a link to the H chondrite meteorites. Incorporating the grain density of the H chondrites, we estimate the macroporosities of the asteroids (290) Bruna, (1164) Kobolda, and (587) Hypsipyle, respectively 41%, 47%, and 65%, implying rubble pile structures. Considering the H chondrites link to asteroid (25) Phocaea, we suggest the parent body of the Phocaea family has been composed of H chondrite like material and the Phocaea family may be one of the sources of H chondrite meteorites.

astro-ph.EP↗

arXiv2025

Attentional Graph Meta-Learning for Indoor Localization Using Extremely Sparse Fingerprints

Fingerprint-based indoor localization is often labor-intensive due to the need for dense grids and repeated measurements across time and space. Maintaining high localization accuracy with extremely sparse fingerprints remains a persistent challenge. Existing benchmark methods primarily rely on the measured fingerprints, while neglecting valuable spatial and environmental characteristics. In this paper, we propose a systematic integration of an Attentional Graph Neural Network (AGNN) model, capable of learning spatial adjacency relationships and aggregating information from neighboring fingerprints, and a meta-learning framework that utilizes datasets with similar environmental characteristics to enhance model training. To minimize the labor required for fingerprint collection, we introduce two novel data augmentation strategies: 1) unlabeled fingerprint augmentation using moving platforms, which enables the semi-supervised AGNN model to incorporate information from unlabeled fingerprints, and 2) synthetic labeled fingerprint augmentation through environmental digital twins, which enhances the meta-learning framework through a practical distribution alignment, which can minimize the feature discrepancy between synthetic and real-world fingerprints effectively. By integrating these novel modules, we propose the Attentional Graph Meta-Learning (AGML) model. This novel model combines the strengths of the AGNN model and the meta-learning framework to address the challenges posed by extremely sparse fingerprints. To validate our approach, we collected multiple datasets from both consumer-grade WiFi devices and professional equipment across diverse environments. Extensive experiments conducted on both synthetic and real-world datasets demonstrate that the AGML model-based localization method consistently outperforms all baseline methods using sparse fingerprints across all evaluated metrics.

cs.LG↗

arXiv2025

LSNet: See Large, Focus Small

Vision network designs, including Convolutional Neural Networks and Vision Transformers, have significantly advanced the field of computer vision. Yet, their complex computations pose challenges for practical deployments, particularly in real-time applications. To tackle this issue, researchers have explored various lightweight and efficient network designs. However, existing lightweight models predominantly leverage self-attention mechanisms and convolutions for token mixing. This dependence brings limitations in effectiveness and efficiency in the perception and aggregation processes of lightweight networks, hindering the balance between performance and efficiency under limited computational budgets. In this paper, we draw inspiration from the dynamic heteroscale vision ability inherent in the efficient human vision system and propose a ``See Large, Focus Small'' strategy for lightweight vision network design. We introduce LS (\textbf{L}arge-\textbf{S}mall) convolution, which combines large-kernel perception and small-kernel aggregation. It can efficiently capture a wide range of perceptual information and achieve precise feature aggregation for dynamic and complex visual representations, thus enabling proficient processing of visual information. Based on LS convolution, we present LSNet, a new family of lightweight models. Extensive experiments demonstrate that LSNet achieves superior performance and efficiency over existing lightweight networks in various vision tasks. Codes and models are available at https://github.com/jameslahm/lsnet.

cs.CV↗

arXiv2025

Axion-like Universal Gravitational Wave Interpretation of Pulsar Timing Array Data

Formation of cosmological solitons is generically accompanied by production of gravitational waves (GWs), with a universal GW background expected at frequency scales below that of non-linear dynamics. Beginning with a general phenomenological description of GWs associated with soliton formation, we demonstrate that universal GW background from axion-like particle (ALP) solitonic oscillons provides a viable interpretation to the recent NANOGrav 15 year pulsar timing array data, which does not suffer from the overproduction of primordial black holes. We show that pulsar timing array data displays preference for models where formed solitons do not strongly interact or cluster. Coincidence observations with Nancy Roman telescope will allow to discriminate between distinct scenarios of cosmological solitons.

astro-ph.CO↗

arXiv2024

YOLO-UniOW: Efficient Universal Open-World Object Detection

Traditional object detection models are constrained by the limitations of closed-set datasets, detecting only categories encountered during training. While multimodal models have extended category recognition by aligning text and image modalities, they introduce significant inference overhead due to cross-modality fusion and still remain restricted by predefined vocabulary, leaving them ineffective at handling unknown objects in open-world scenarios. In this work, we introduce Universal Open-World Object Detection (Uni-OWD), a new paradigm that unifies open-vocabulary and open-world object detection tasks. To address the challenges of this setting, we propose YOLO-UniOW, a novel model that advances the boundaries of efficiency, versatility, and performance. YOLO-UniOW incorporates Adaptive Decision Learning to replace computationally expensive cross-modality fusion with lightweight alignment in the CLIP latent space, achieving efficient detection without compromising generalization. Additionally, we design a Wildcard Learning strategy that detects out-of-distribution objects as "unknown" while enabling dynamic vocabulary expansion without the need for incremental learning. This design empowers YOLO-UniOW to seamlessly adapt to new categories in open-world environments. Extensive experiments validate the superiority of YOLO-UniOW, achieving achieving 34.6 AP and 30.0 APr on LVIS with an inference speed of 69.6 FPS. The model also sets benchmarks on M-OWODB, S-OWODB, and nuScenes datasets, showcasing its unmatched performance in open-world object detection. Code and models are available at https://github.com/THU-MIG/YOLO-UniOW.

cs.CV↗

arXiv2024

Promptable Anomaly Segmentation with SAM Through Self-Perception Tuning

Segment Anything Model (SAM) has made great progress in anomaly segmentation tasks due to its impressive generalization ability. However, existing methods that directly apply SAM through prompting often overlook the domain shift issue, where SAM performs well on natural images but struggles in industrial scenarios. Parameter-Efficient Fine-Tuning (PEFT) offers a promising solution, but it may yield suboptimal performance by not adequately addressing the perception challenges during adaptation to anomaly images. In this paper, we propose a novel \textbf{S}elf-\textbf{P}erceptinon \textbf{T}uning (\textbf{SPT}) method, aiming to enhance SAM's perception capability for anomaly segmentation. The SPT method incorporates a self-drafting tuning strategy, which generates an initial coarse draft of the anomaly mask, followed by a refinement process. Additionally, a visual-relation-aware adapter is introduced to improve the perception of discriminative relational information for mask generation. Extensive experimental results on several benchmark datasets demonstrate that our SPT method can significantly outperform baseline methods, validating its effectiveness.

cs.CV↗

arXiv2024

[CLS] Token Tells Everything Needed for Training-free Efficient MLLMs

Multimodal Large Language Models (MLLMs) have recently demonstrated strong performance across a wide range of vision-language tasks, garnering significant attention in the computer vision. However, their efficient deployment remains a substantial challenge due to high computational costs and memory requirements. Recognizing the redundancy of information within the vision modality, recent studies have explored methods for compressing visual tokens in MLLMs to enhance efficiency in a training-free manner. Despite their effectiveness, existing methods like Fast rely on the attention between visual tokens and prompt text tokens as the importance indicator, overlooking the relevance to response text and thus introducing perception bias. In this paper, we demonstrate that in MLLMs, the [CLS] token in the visual encoder inherently knows which visual tokens are important for MLLMs. Building on this prior, we introduce a simple yet effective method for train-free visual token compression, called VTC-CLS. Firstly, it leverages the attention score of the [CLS] token on visual tokens as an importance indicator for pruning visual tokens. Besides, we also explore ensembling the importance scores derived by the [CLS] token from different layers to capture the key visual information more comprehensively. Extensive experiments demonstrate that our VTC-CLS achieves the state-of-the-art performance across various tasks compared with baseline methods. It also brings notably less computational costs in a training-free manner, highlighting its effectiveness and superiority. Code and models are available at \url{https://github.com/THU-MIG/VTC-CLS}.

cs.CV↗

arXiv2024

ReDAN: An Empirical Study on Remote DoS Attacks against NAT Networks

In this paper, we conduct an empirical study on remote DoS attacks targeting NAT networks. We show that Internet attackers operating outside local NAT networks can remotely identify a NAT device and subsequently terminate TCP connections initiated from the identified NAT device to external servers. Our attack involves two steps. First, we identify NAT devices on the Internet by exploiting inadequacies in the PMTUD mechanism within NAT specifications. This deficiency creates a fundamental side channel that allows Internet attackers to distinguish if a public IPv4 address serves a NAT device or a separate IP host, aiding in the identification of target NAT devices. Second, we launch a remote DoS attack to terminate TCP connections on the identified NAT devices. While recent NAT implementations may include protective measures, such as packet legitimacy validation to prevent malicious manipulations on NAT mappings, we discover that these safeguards are not widely adopted in real world. Consequently, attackers can send crafted packets to deceive NAT devices into erroneously removing innocent TCP connection mappings, thereby disrupting the NATed clients to access remote TCP servers. Our experimental results reveal widespread security vulnerabilities in existing NAT devices. After testing 8 types of router firmware and 30 commercial NAT devices from 14 vendors, we identify vulnerabilities in 6 firmware types and 29 NAT devices. Moreover, our measurements reveal a stark reality: 166 out of 180 (over 92%) tested real-world NAT networks, comprising 90 4G LTE/5G networks, 60 public Wi-Fi networks, and 30 cloud VPS networks, are susceptible to exploitation. We responsibly disclosed the vulnerabilities to affected vendors and received a significant number of acknowledgments. Finally, we propose our countermeasures against the identified DoS attack.

cs.CR↗