arXiv subjects

Xuan Wang

Publications and source records attributed to Xuan Wang.

At least 217 records · Page 12Linked to original sources

arXiv2024

Do We Really Need a Complex Agent System? Distill Embodied Agent into a Single Model

With the power of large language models (LLMs), open-ended embodied agents can flexibly understand human instructions, generate interpretable guidance strategies, and output executable actions. Nowadays, Multi-modal Language Models~(MLMs) integrate multi-modal signals into LLMs, further bringing richer perception to entity agents and allowing embodied agents to perceive world-understanding tasks more delicately. However, existing works: 1) operate independently by agents, each containing multiple LLMs, from perception to action, resulting in gaps between complex tasks and execution; 2) train MLMs on static data, struggling with dynamics in open-ended scenarios; 3) input prior knowledge directly as prompts, suppressing application flexibility. We propose STEVE-2, a hierarchical knowledge distillation framework for open-ended embodied tasks, characterized by 1) a hierarchical system for multi-granular task division, 2) a mirrored distillation method for parallel simulation data, and 3) an extra expert model for bringing additional knowledge into parallel simulation. After distillation, embodied agents can complete complex, open-ended tasks without additional expert guidance, utilizing the performance and knowledge of a versatile MLM. Extensive evaluations on navigation and creation tasks highlight the superior performance of STEVE-2 in open-ended tasks, with $1.4 \times$ - $7.3 \times$ in performance.

cs.AI↗

arXiv2024

AI for Biomedicine in the Era of Large Language Models

The capabilities of AI for biomedicine span a wide spectrum, from the atomic level, where it solves partial differential equations for quantum systems, to the molecular level, predicting chemical or protein structures, and further extending to societal predictions like infectious disease outbreaks. Recent advancements in large language models, exemplified by models like ChatGPT, have showcased significant prowess in natural language tasks, such as translating languages, constructing chatbots, and answering questions. When we consider biomedical data, we observe a resemblance to natural language in terms of sequences: biomedical literature and health records presented as text, biological sequences or sequencing data arranged in sequences, or sensor data like brain signals as time series. The question arises: Can we harness the potential of recent large language models to drive biomedical knowledge discoveries? In this survey, we will explore the application of large language models to three crucial categories of biomedical data: 1) textual data, 2) biological sequences, and 3) brain signals. Furthermore, we will delve into large language model challenges in biomedical research, including ensuring trustworthiness, achieving personalization, and adapting to multi-modal data representation

cs.CL↗

arXiv2024

iComMa: Inverting 3D Gaussian Splatting for Camera Pose Estimation via Comparing and Matching

We present a method named iComMa to address the 6D camera pose estimation problem in computer vision. Conventional pose estimation methods typically rely on the target's CAD model or necessitate specific network training tailored to particular object classes. Some existing methods have achieved promising results in mesh-free object and scene pose estimation by inverting the Neural Radiance Fields (NeRF). However, they still struggle with adverse initializations such as large rotations and translations. To address this issue, we propose an efficient method for accurate camera pose estimation by inverting 3D Gaussian Splatting (3DGS). Specifically, a gradient-based differentiable framework optimizes camera pose by minimizing the residual between the query image and the rendered image, requiring no training. An end-to-end matching module is designed to enhance the model's robustness against adverse initializations, while minimizing pixel-level comparing loss aids in precise pose estimation. Experimental results on synthetic and complex real-world data demonstrate the effectiveness of the proposed approach in challenging conditions and the accuracy of camera pose estimation.

cs.CV↗

arXiv2024

Scaling Team Coordination on Graphs with Reinforcement Learning

This paper studies Reinforcement Learning (RL) techniques to enable team coordination behaviors in graph environments with support actions among teammates to reduce the costs of traversing certain risky edges in a centralized manner. While classical approaches can solve this non-standard multi-agent path planning problem by converting the original Environment Graph (EG) into a Joint State Graph (JSG) to implicitly incorporate the support actions, those methods do not scale well to large graphs and teams. To address this curse of dimensionality, we propose to use RL to enable agents to learn such graph traversal and teammate supporting behaviors in a data-driven manner. Specifically, through a new formulation of the team coordination on graphs with risky edges problem into Markov Decision Processes (MDPs) with a novel state and action space, we investigate how RL can solve it in two paradigms: First, we use RL for a team of agents to learn how to coordinate and reach the goal with minimal cost on a single EG. We show that RL efficiently solves problems with up to 20/4 or 25/3 nodes/agents, using a fraction of the time needed for JSG to solve such complex problems; Second, we learn a general RL policy for any $N$-node EGs to produce efficient supporting behaviors. We present extensive experiments and compare our RL approaches against their classical counterparts.

cs.RO↗

arXiv2024

Decoupling Degradations with Recurrent Network for Video Restoration in Under-Display Camera

Under-display camera (UDC) systems are the foundation of full-screen display devices in which the lens mounts under the display. The pixel array of light-emitting diodes used for display diffracts and attenuates incident light, causing various degradations as the light intensity changes. Unlike general video restoration which recovers video by treating different degradation factors equally, video restoration for UDC systems is more challenging that concerns removing diverse degradation over time while preserving temporal consistency. In this paper, we introduce a novel video restoration network, called D$^2$RNet, specifically designed for UDC systems. It employs a set of Decoupling Attention Modules (DAM) that effectively separate the various video degradation factors. More specifically, a soft mask generation function is proposed to formulate each frame into flare and haze based on the diffraction arising from incident light of different intensities, followed by the proposed flare and haze removal components that leverage long- and short-term feature learning to handle the respective degradations. Such a design offers an targeted and effective solution to eliminating various types of degradation in UDC systems. We further extend our design into multi-scale to overcome the scale-changing of degradation that often occur in long-range videos. To demonstrate the superiority of D$^2$RNet, we propose a large-scale UDC video benchmark by gathering HDR videos and generating realistically degraded videos using the point spread function measured by a commercial UDC system. Extensive quantitative and qualitative evaluations demonstrate the superiority of D$^2$RNet compared to other state-of-the-art video restoration and UDC image restoration methods. Code is available at https://github.com/ChengxuLiu/DDRNet.git

cs.CV↗

arXiv2024

MGE: A Training-Free and Efficient Model Generation and Enhancement Scheme

To provide a foundation for the research of deep learning models, the construction of model pool is an essential step. This paper proposes a Training-Free and Efficient Model Generation and Enhancement Scheme (MGE). This scheme primarily considers two aspects during the model generation process: the distribution of model parameters and model performance. Experiments result shows that generated models are comparable to models obtained through normal training, and even superior in some cases. Moreover, the time consumed in generating models accounts for only 1\% of the time required for normal model training. More importantly, with the enhancement of Evolution-MGE, generated models exhibits competitive generalization ability in few-shot tasks. And the behavioral dissimilarity of generated models has the potential of adversarial defense.

cs.CV↗

arXiv2024

Flares hunting in hot subdwarf and white dwarf stars from Cycles 1-5 of TESS photometry

Stellar flares are critical phenomena on stellar surfaces, which are closely tied to stellar magnetism. While extensively studied in main-sequence (MS) stars, their occurrence in evolved compact stars, specifically hot subdwarfs and white dwarfs (WDs), remains scarcely explored. Based on Cycles 1-5 of TESS photometry, we conducted a pioneering survey of flare events in $\sim12,000$ compact stars, corresponding to $\sim38,000$ light curves with 2-minute cadence. Through dedicated techniques for detrending light curves, identifying preliminary flare candidates, and validating them via machine learning, we established a catalog of 1016 flares from 193 compact stars, including 182 from 58 sdB/sdO stars and 834 from 135 WDs, respectively. However, all flaring compact stars showed signs of contamination from nearby objects or companion stars, preventing sole attribution of the detected flares. For WDs, it is highly probable that the flares originated from their cool MS companions. In contrast, the higher luminosities of sdB/sdO stars diminish companion contributions, suggesting that detected flares originated from sdB/sdO stars themselves or through close magnetic interactions with companions. Focusing on a refined sample of 23 flares from 13 sdB/sdO stars, we found their flare frequency distributions were slightly divergent from those of cool MS stars; instead, they resemble those of hot B/A-type MS stars having radiative envelopes. This similarity implies the flares on sdB/sdO stars, if these flares did originate from them, may share underlying mechanisms with hot MS stars, which warrants further investigation.

astro-ph.SR↗

arXiv2024

Hypergraph product code with 0.2 constant coding rate and high code capacity noise threshold

The low coding rate of quantum stabilizer codes results in formidable physical qubit overhead when realizing quantum error correcting in engineering. In this letter, we propose a new class of hypergraph-product code called TGRE-hypergraph-product code. This code has constant coding rate 0.2, which is the highest constant coding rate of quantum stabilizer codes to our best knowledge. We perform simulations to test the error correcting capability TGRE-hypergraph-product code and find their code capacity noise threshold in depolarizing noise channel is around 0.096.

quant-ph↗

arXiv2024

Quantum polar stabilizer codes based on polarization of pure quantum channel don't work for quantum computing

Inspired by classical polar codes, whose coding rate can asymptotically achieve the Shannon capacity, researchers are trying to find its analogue in quantum information field, which are called quantum polar codes. However, no one has designed a quantum polar coding scheme which applies to quantum computing yet. There are two intuitions in previous research. The first is that directly converting classical polar coding circuits to quantum ones will produce polarization phenomenon of pure quantum channel, which has been proved in our previous work. The second is that based on this quantum polarization phenomenon one can design a quantum polar coding scheme that applies to quantum computing. There are several previous work following the second intuition, none of which has been verified by experiments. In this paper, we follow the second intuition and propose a more reasonable quantum polar stabilizer code construction algorithm than any previous ones by using the theory of stabilizer codes. Unfortunately, simulation experiments show that even the stabilizer codes obtained from this more reasonable construction algorithm don't work, which implies that the second intuition leads to a dead end. Based on the analysis on why the second intuition don't work, we provide a possible future direction of designing quantum stabilizer codes with high coding rate by borrowing the idea of classical polar codes. following this direction, we find a class of quantum stabilizer codes with coding rate 0.5 for pure Pauli X, Z and Y noise.

quant-ph↗

arXiv2024

Determining the upper bound of code distance of quantum stabilizer codes through Monte Carlo method based on fully decoupled belief propagation

Code distance is an important parameter for quantum stabilizer codes (QSCs). Directly precisely computing it is an NP-complete problem. However, the upper bound of code distance can be computed by some efficient methods. In this paper, employing the idea of Monte Carlo method, we propose the algorithm of determining the upper bound of code distance of QSCs based on fully decoupled belief propagation. Our algorithm shows high precision - the upper bound of code distance determined by the algorithm of a variety of QSCs whose code distance is known is consistent with actual code distance. Besides, we explore the upper bound of logical X operators of Z-type Tanner-graph-recursive-expansion (Z-TGRE) code and Chamon code, which is a kind of XYZ product code constructed by three repetition codes. The former is consistent with the theoretical analysis, and the latter implies the code distance of XYZ product codes can very likely achieve $O(N^{2/3})$, which supports the conjecture of Leverrier et al..

quant-ph↗

arXiv2024

Saccade-Contingent Rendering

Battery-constrained power consumption, compute limitations, and high frame rate requirements in head-mounted displays present unique challenges in the drive to present increasingly immersive and comfortable imagery in virtual reality. However, humans are not equally sensitive to all regions of the visual field, and perceptually-optimized rendering techniques are increasingly utilized to address these bottlenecks. Many of these techniques are gaze-contingent and often render reduced detail away from a user's fixation. Such techniques are dependent on spatio-temporally-accurate gaze tracking and can result in obvious visual artifacts when eye tracking is inaccurate. In this work we present a gaze-contingent rendering technique which only requires saccade detection, bypassing the need for highly-accurate eye tracking. In our first experiment, we show that visual acuity is reduced for several hundred milliseconds after a saccade. In our second experiment, we use these results to reduce the rendered image resolution after saccades in a controlled psychophysical setup, and find that observers cannot discriminate between saccade-contingent reduced-resolution rendering and full-resolution rendering. Finally, in our third experiment, we introduce a 90 pixels per degree headset and validate our saccade-contingent rendering method under typical VR viewing conditions.

cs.GR↗

arXiv2024

Eye Motion Matters for 3D Face Reconstruction

Recent advances in single-image 3D face reconstruction have shown remarkable progress in various applications. Nevertheless, prevailing techniques tend to prioritize the global facial contour and expression, often neglecting the nuanced dynamics of the eye region. In response, we introduce an Eye Landmark Adjustment Module, complemented by a Local Dynamic Loss, designed to capture the dynamic features of the eyes area. Our module allows for flexible adjustment of landmarks, resulting in accurate recreation of various eye states. In this paper, we present a comprehensive evaluation of our approach, conducting extensive experiments on two datasets. The results underscore the superior performance of our approach, highlighting its significant contributions in addressing this particular challenge.

cs.CV↗

arXiv2023

CiT-Net: Convolutional Neural Networks Hand in Hand with Vision Transformers for Medical Image Segmentation

The hybrid architecture of convolutional neural networks (CNNs) and Transformer are very popular for medical image segmentation. However, it suffers from two challenges. First, although a CNNs branch can capture the local image features using vanilla convolution, it cannot achieve adaptive feature learning. Second, although a Transformer branch can capture the global features, it ignores the channel and cross-dimensional self-attention, resulting in a low segmentation accuracy on complex-content images. To address these challenges, we propose a novel hybrid architecture of convolutional neural networks hand in hand with vision Transformers (CiT-Net) for medical image segmentation. Our network has two advantages. First, we design a dynamic deformable convolution and apply it to the CNNs branch, which overcomes the weak feature extraction ability due to fixed-size convolution kernels and the stiff design of sharing kernel parameters among different inputs. Second, we design a shifted-window adaptive complementary attention module and a compact convolutional projection. We apply them to the Transformer branch to learn the cross-dimensional long-term dependency for medical images. Experimental results show that our CiT-Net provides better medical image segmentation results than popular SOTA methods. Besides, our CiT-Net requires lower parameters and less computational costs and does not rely on pre-training. The code is publicly available at https://github.com/SR0920/CiT-Net.

eess.IV↗

arXiv2023

VectorTalker: SVG Talking Face Generation with Progressive Vectorisation

High-fidelity and efficient audio-driven talking head generation has been a key research topic in computer graphics and computer vision. In this work, we study vector image based audio-driven talking head generation. Compared with directly animating the raster image that most widely used in existing works, vector image enjoys its excellent scalability being used for many applications. There are two main challenges for vector image based talking head generation: the high-quality vector image reconstruction w.r.t. the source portrait image and the vivid animation w.r.t. the audio signal. To address these, we propose a novel scalable vector graphic reconstruction and animation method, dubbed VectorTalker. Specifically, for the highfidelity reconstruction, VectorTalker hierarchically reconstructs the vector image in a coarse-to-fine manner. For the vivid audio-driven facial animation, we propose to use facial landmarks as intermediate motion representation and propose an efficient landmark-driven vector image deformation module. Our approach can handle various styles of portrait images within a unified framework, including Japanese manga, cartoon, and photorealistic images. We conduct extensive quantitative and qualitative evaluations and the experimental results demonstrate the superiority of VectorTalker in both vector graphic reconstruction and audio-driven animation.

cs.CV↗

arXiv2023

Resilience for Distributed Consensus with Constraints

This paper proposes a new approach that enables multi-agent systems to achieve resilient \textit{constrained} consensus in the presence of Byzantine attacks, in contrast to existing literature that is only applicable to \textit{unconstrained} resilient consensus problems. The key enabler for our approach is a new device called a \textit{$(γ_i,α_i)$-resilient convex combination}, which allows normal agents in the network to utilize their locally available information to automatically isolate the impact of the Byzantine agents. Such a resilient convex combination is computable through linear programming, whose complexity scales well with the size of the overall system. By applying this new device to multi-agent systems, we introduce network and constraint redundancy conditions under which resilient constrained consensus can be achieved with an exponential convergence rate. We also provide insights on the design of a network such that the redundancy conditions are satisfied. Finally, numerical simulations and an example of safe multi-agent learning are provided to demonstrate the effectiveness of the proposed results.

eess.SY↗

arXiv2023

Meta-Path Based Attentional Graph Learning Model for Vulnerability Detection

In recent years, deep learning (DL)-based methods have been widely used in code vulnerability detection. The DL-based methods typically extract structural information from source code, e.g., code structure graph, and adopt neural networks such as Graph Neural Networks (GNNs) to learn the graph representations. However, these methods fail to consider the heterogeneous relations in the code structure graph, i.e., the heterogeneous relations mean that the different types of edges connect different types of nodes in the graph, which may obstruct the graph representation learning. Besides, these methods are limited in capturing long-range dependencies due to the deep levels in the code structure graph. In this paper, we propose a Meta-path based Attentional Graph learning model for code vulNErability deTection, called MAGNET. MAGNET constructs a multi-granularity meta-path graph for each code snippet, in which the heterogeneous relations are denoted as meta-paths to represent the structural information. A meta-path based hierarchical attentional graph neural network is also proposed to capture the relations between distant nodes in the graph. We evaluate MAGNET on three public datasets and the results show that MAGNET outperforms the best baseline method in terms of F1 score by 6.32%, 21.50%, and 25.40%, respectively. MAGNET also achieves the best performance among all the baseline methods in detecting Top-25 most dangerous Common Weakness Enumerations (CWEs), further demonstrating its effectiveness in vulnerability detection.

cs.SE↗

arXiv2023

Improved belief propagation decoding algorithm based on decoupling representation of Pauli operators for quantum LDPC codes

We propose a new method called decoupling representation to represent Pauli operators as vectors over $GF(2)$, based on which we propose partially decoupled belief propagation and fully decoupled belief propagation decoding algorithm for quantum low density parity-check codes. These two algorithms have the capability to deal with the correlations between the $X$ part and the $Z$ part of the vectors in symplectic representation, which are introduced by Pauli $Y$ errors. Hence, they can not only apply to CSS codes, but also to non-CSS codes. Under the assumption that there is no measurement error, compared with traditional belief propagation algorithm in symplectic representation over $GF(2)$, within the same number of iterations, the decoding accuracy of partially decoupled belief propagation and fully decoupled belief propagation algorithm is significantly improved in pure $Y$ noise and depolarizing noise, which supports that decoding algorithms of quantum error correcting codes might have better performance in decoupling representation than in symplectic representation. The impressive performance of fully decoupled belief propagation algorithm might promote the realization of quantum error correcting codes in engineering.

quant-ph↗

arXiv2023

Semi-supervised Estimation of Event Rate with Doubly-censored Survival Data

Electronic Health Record (EHR) has emerged as a valuable source of data for translational research. To leverage EHR data for risk prediction and subsequently clinical decision support, clinical endpoints are often time to onset of a clinical condition of interest. Precise information on clinical event times is often not directly available and requires labor-intensive manual chart review to ascertain. In addition, events may occur outside of the hospital system, resulting in both left and right censoring often termed double censoring. On the other hand, proxies such as time to the first diagnostic code are readily available yet with varying degrees of accuracy. Using error-prone event times derived from these proxies can lead to biased risk estimates while only relying on manually annotated event times, which are typically only available for a small subset of patients, can lead to high variability. This signifies the need for semi-supervised estimation methods that can efficiently combine information from both the small subset of labeled observations and a large size of surrogate proxies. While semi-supervised estimation methods have been recently developed for binary and right-censored data, no methods currently exist in the presence of double censoring. This paper fills the gap by developing a robust and efficient Semi-supervised Estimation of Event rate with Doubly-censored Survival data (SEEDS) by leveraging a small set of gold standard labels and a large set of surrogate features. Under regularity conditions, we demonstrate that the proposed SEEDS estimator is consistent and asymptotically normal. Simulation results illustrate that SEEDS performs well in finite samples and can be substantially more efficient compared to the supervised counterpart. We apply the SEEDS to estimate the age-specific survival rate of type 2 diabetes using EHR data from Mass General Brigham.

stat.ME↗