arXiv subjects

Hui Zhang

Publications and source records attributed to Hui Zhang.

At least 253 records · Page 14Linked to original sources

arXiv2024

All-In-One Medical Image Restoration via Task-Adaptive Routing

Although single-task medical image restoration (MedIR) has witnessed remarkable success, the limited generalizability of these methods poses a substantial obstacle to wider application. In this paper, we focus on the task of all-in-one medical image restoration, aiming to address multiple distinct MedIR tasks with a single universal model. Nonetheless, due to significant differences between different MedIR tasks, training a universal model often encounters task interference issues, where different tasks with shared parameters may conflict with each other in the gradient update direction. This task interference leads to deviation of the model update direction from the optimal path, thereby affecting the model's performance. To tackle this issue, we propose a task-adaptive routing strategy, allowing conflicting tasks to select different network paths in spatial and channel dimensions, thereby mitigating task interference. Experimental results demonstrate that our proposed \textbf{A}ll-in-one \textbf{M}edical \textbf{I}mage \textbf{R}estoration (\textbf{AMIR}) network achieves state-of-the-art performance in three MedIR tasks: MRI super-resolution, CT denoising, and PET synthesis, both in single-task and all-in-one settings. The code and data will be available at \href{https://github.com/Yaziwel/All-In-One-Medical-Image-Restoration-via-Task-Adaptive-Routing.git}{https://github.com/Yaziwel/AMIR}.

cs.CV↗

arXiv2024

Hyperon Polarization from the Vortical Fluid in Low Energy Nuclear Collisions

In 2017, STAR Collaboration reported the measurements of hyperon global polarization in heavy ion collisions, suggesting the subatomic fireball fluid created in these collisions as the most vortical fluid. There remains the interesting question: at which beam energy the truly most vortical fluid will be located. In this work we perform a systematic study on the beam energy dependence of hyperon global polarization phenomenon, especially in the interesting $\hat{O}(1\sim 10)\ \rm GeV$ region. We find a non-monotonic trend, with the global polarization to first increase and then decrease when beam energy is lowered from $27~\rm GeV$ down to $3~\rm GeV$. The maximum polarization signal has been identified around $\sqrt{s_{NN}} = 7.7~\rm GeV$, where the heavy ion collisions presumably create the most vortical fluid. Detailed experimental measurements in the $\hat{O}(1\sim 10)\ \rm GeV$ beam energy region are expected to test the prediction very soon.

nucl-th↗

arXiv2024

HD 21520 b: a warm sub-Neptune transiting a bright G dwarf

We report the discovery and validation of HD 21520 b, a transiting planet found with TESS and orbiting a bright G dwarf (V=9.2, $T_{eff} = 5871 \pm 62$ K, $R_{\star} = 1.04\pm 0.02\, R_{\odot}$). HD 21520 b was originally alerted as a system (TOI-4320) consisting of two planet candidates with periods of 703.6 and 46.4 days. However, our analysis supports instead a single-planet system with an orbital period of $25.1292\pm0.0001$ days and radius of $2.70 \pm 0.09\, R_{\oplus}$. Three full transits in sectors 4, 30 and 31 match this period and have transit depths and durations in agreement with each other, as does a partial transit in sector 3. We also observe transits using CHEOPS and LCOGT. SOAR and Gemini high-resolution imaging do not indicate the presence of any nearby companions, and MINERVA-Australis and CORALIE radial velocities rule out an on-target spectroscopic binary. Additionally, we use ESPRESSO radial velocities to obtain a tentative mass measurement of $7.9^{+3.2}_{-3.0}\, M_{\oplus}$, with a 3-$σ$ upper limit of 17.7 $M_{\oplus}$. Due to the bright nature of its host and likely significant gas envelope of the planet, HD 21520 b is a promising candidate for further mass measurements and for atmospheric characterization.

astro-ph.EP↗

arXiv2024

Teaching with Uncertainty: Unleashing the Potential of Knowledge Distillation in Object Detection

Knowledge distillation (KD) is a widely adopted and effective method for compressing models in object detection tasks. Particularly, feature-based distillation methods have shown remarkable performance. Existing approaches often ignore the uncertainty in the teacher model's knowledge, which stems from data noise and imperfect training. This limits the student model's ability to learn latent knowledge, as it may overly rely on the teacher's imperfect guidance. In this paper, we propose a novel feature-based distillation paradigm with knowledge uncertainty for object detection, termed "Uncertainty Estimation-Discriminative Knowledge Extraction-Knowledge Transfer (UET)", which can seamlessly integrate with existing distillation methods. By leveraging the Monte Carlo dropout technique, we introduce knowledge uncertainty into the training process of the student model, facilitating deeper exploration of latent knowledge. Our method performs effectively during the KD process without requiring intricate structures or extensive computational resources. Extensive experiments validate the effectiveness of our proposed approach across various distillation strategies, detectors, and backbone architectures. Specifically, following our proposed paradigm, the existing FGD method achieves state-of-the-art (SoTA) performance, with ResNet50-based GFL achieving 44.1% mAP on the COCO dataset, surpassing the baselines by 3.9%.

cs.CV↗

arXiv2024

Nacala-Roof-Material: Drone Imagery for Roof Detection, Classification, and Segmentation to Support Mosquito-borne Disease Risk Assessment

As low-quality housing and in particular certain roof characteristics are associated with an increased risk of malaria, classification of roof types based on remote sensing imagery can support the assessment of malaria risk and thereby help prevent the disease. To support research in this area, we release the Nacala-Roof-Material dataset, which contains high-resolution drone images from Mozambique with corresponding labels delineating houses and specifying their roof types. The dataset defines a multi-task computer vision problem, comprising object detection, classification, and segmentation. In addition, we benchmarked various state-of-the-art approaches on the dataset. Canonical U-Nets, YOLOv8, and a custom decoder on pretrained DINOv2 served as baselines. We show that each of the methods has its advantages but none is superior on all tasks, which highlights the potential of our dataset for future research in multi-task learning. While the tasks are closely related, accurate segmentation of objects does not necessarily imply accurate instance separation, and vice versa. We address this general issue by introducing a variant of the deep ordinal watershed (DOW) approach that additionally separates the interior of objects, allowing for improved object delineation and separation. We show that our DOW variant is a generic approach that improves the performance of both U-Net and DINOv2 backbones, leading to a better trade-off between semantic segmentation and instance segmentation.

cs.CV↗

arXiv2024

The RoboDrive Challenge: Drive Anytime Anywhere in Any Condition

In the realm of autonomous driving, robust perception under out-of-distribution conditions is paramount for the safe deployment of vehicles. Challenges such as adverse weather, sensor malfunctions, and environmental unpredictability can severely impact the performance of autonomous systems. The 2024 RoboDrive Challenge was crafted to propel the development of driving perception technologies that can withstand and adapt to these real-world variabilities. Focusing on four pivotal tasks -- BEV detection, map segmentation, semantic occupancy prediction, and multi-view depth estimation -- the competition laid down a gauntlet to innovate and enhance system resilience against typical and atypical disturbances. This year's challenge consisted of five distinct tracks and attracted 140 registered teams from 93 institutes across 11 countries, resulting in nearly one thousand submissions evaluated through our servers. The competition culminated in 15 top-performing solutions, which introduced a range of innovative approaches including advanced data augmentation, multi-sensor fusion, self-supervised learning for error correction, and new algorithmic strategies to enhance sensor robustness. These contributions significantly advanced the state of the art, particularly in handling sensor inconsistencies and environmental variability. Participants, through collaborative efforts, pushed the boundaries of current technologies, showcasing their potential in real-world scenarios. Extensive evaluations and analyses provided insights into the effectiveness of these solutions, highlighting key trends and successful strategies for improving the resilience of driving perception systems. This challenge has set a new benchmark in the field, providing a rich repository of techniques expected to guide future research in this field.

cs.CV↗

arXiv2024

Momentum-resolved electronic structures and strong electronic correlations in graphene-like nitride superconductors

Although transition-metal nitrides have been widely applied for several decades, experimental investigations of their high-resolution electronic band structures are rare due to the lack of high-quality single-crystalline samples. Here, we report on the first momentum-resolved electronic band structures of titanium nitride (TiN) films, a remarkable nitride superconductor. The measurements of crystal structures and electrical transport properties confirmed the high quality of these films. More importantly, with a combination of high-resolution angle-resolved photoelectron spectroscopy and the first-principles calculations, the extracted Coulomb interaction strength of TiN films can be as large as 8.5 eV, whereas resonant photoemission spectroscopy yields a value of 6.26 eV. These large values of Coulomb interaction strength indicate that superconducting TiN is a strongly correlated system. Our results uncover the unexpected electronic correlations in transition-metal nitrides, potentially providing a perspective not only to understand their emergent quantum states but also to develop their applications in quantum devices.

cond-mat.mtrl-sci↗

arXiv2024

Curvatures of metric Jordan algebras

In this paper, we investigate metric Jordan algebras, and follow the lines of the paper (J. Milnor: Curvatures of left invariant metrics on Lie groups. Adv. Math. (1976)). Firstly, we define the Jordan-Levi-Civita connection, then we show that every metric Jordan algebra admits a unique Jordan-Levi-Civita connection. Secondly, using the Jordan-Levi-Civita connection, we introduce three natural curvature tensors on metric Jordan algebras, and obtain the corresponding curvature formulas. Thirdly, based on these curvature formulas, we prove that every formally real Jordan algebra admits both a metric of non-positive Jordan curvature, and a Jordan-Einstein metric of negative Jordan scalar curvature. Besides, for nilpotent Jordan algebras, we prove that they admit no Jordan-Einstein metrics.

math.DG↗

arXiv2024

Augmenting Sports Videos with VisCommentator

Visualizing data in sports videos is gaining traction in sports analytics, given its ability to communicate insights and explicate player strategies engagingly. However, augmenting sports videos with such data visualizations is challenging, especially for sports analysts, as it requires considerable expertise in video editing. To ease the creation process, we present a design space that characterizes augmented sports videos at an element-level (what the constituents are) and clip-level (how those constituents are organized). We do so by systematically reviewing 233 examples of augmented sports videos collected from TV channels, teams, and leagues. The design space guides selection of data insights and visualizations for various purposes. Informed by the design space and close collaboration with domain experts, we design VisCommentator, a fast prototyping tool, to eases the creation of augmented table tennis videos by leveraging machine learning-based data extractors and design space-based visualization recommendations. With VisCommentator, sports analysts can create an augmented video by selecting the data to visualize instead of manually drawing the graphical marks. Our system can be generalized to other racket sports (e.g., tennis, badminton) once the underlying datasets and models are available. A user study with seven domain experts shows high satisfaction with our system, confirms that the participants can reproduce augmented sports videos in a short period, and provides insightful implications into future improvements and opportunities.

cs.HC↗

arXiv2024

Deep Lead Optimization: Leveraging Generative AI for Structural Modification

The idea of using deep-learning-based molecular generation to accelerate discovery of drug candidates has attracted extraordinary attention, and many deep generative models have been developed for automated drug design, termed molecular generation. In general, molecular generation encompasses two main strategies: de novo design, which generates novel molecular structures from scratch, and lead optimization, which refines existing molecules into drug candidates. Among them, lead optimization plays an important role in real-world drug design. For example, it can enable the development of me-better drugs that are chemically distinct yet more effective than the original drugs. It can also facilitate fragment-based drug design, transforming virtual-screened small ligands with low affinity into first-in-class medicines. Despite its importance, automated lead optimization remains underexplored compared to the well-established de novo generative models, due to its reliance on complex biological and chemical knowledge. To bridge this gap, we conduct a systematic review of traditional computational methods for lead optimization, organizing these strategies into four principal sub-tasks with defined inputs and outputs. This review delves into the basic concepts, goals, conventional CADD techniques, and recent advancements in AIDD. Additionally, we introduce a unified perspective based on constrained subgraph generation to harmonize the methodologies of de novo design and lead optimization. Through this lens, de novo design can incorporate strategies from lead optimization to address the challenge of generating hard-to-synthesize molecules; inversely, lead optimization can benefit from the innovations in de novo design by approaching it as a task of generating molecules conditioned on certain substructures.

q-bio.BM↗

arXiv2024

Reconstructing Intrinsic Stellar Noise with Stellar Atmospheric Parameters and Chromospheric Activity

Accurately characterizing intrinsic stellar photometric noise induced by stellar astrophysics, such as stellar activity, granulation, and oscillations, is of crucial importance for detecting transiting exoplanets. In this study, we investigate the relation between the intrinsic stellar photometric noise, as quantified by the Kepler rrmsCDPP measurement, and the level of stellar chromospheric activity, as indicated by the S-index of Ca II HK lines derived from the LAMOST spectra. Our results reveal a clear positive correlation between S-index and rrmsCDPP, and the correlation becomes more significant at higher activity levels and on longer timescales. We have therefore built an empirical relation between rrmsCDPP and S-index as well as Teff, logg, [Fe/H], and apparent magnitude with the XGBoost regression algorithm, using the LAMOST-Kepler common star sample as the training set. This method achieves a precision of ~20 ppm for inferring the intrinsic noise from the S-index and other stellar labels on a 6-hour integration duration. We have applied this empirical relation to the full LAMOST DR7 spectra database, and obtained the intrinsic noise predictions for 1,358,275 stars. The resultant catalog is publicly available and expected to be valuable for optimizing target selection for future exoplanet-hunting space missions, such as the Earth 2.0 mission.

astro-ph.SR↗

arXiv2024

Single-peak and multi-peak solutions for Hamiltonian elliptic systems in dimension two

This paper is concerned with the Hamiltonian elliptic system in dimension two\begin{equation*}\aligned \left\{ \begin{array}{lll} -ε^2Δu+V(x)u=g(v)\ & \text{in}\quad \mathbb{R}^2,\\ -ε^2Δv+V(x)v=f(u)\ & \text{in}\quad \mathbb{R}^2, \end{array}\right.\endaligned \end{equation*} where $V\in C(\mathbb{R}^2)$ has local minimum points, and $f,g\in C^1(\mathbb{R})$ are assumed to be of exponential growth in the sense of Trudinger-Moser inequality. When $V$ admits one or several local strict minimum points, we show the existence and concentration of single-peak and multi-peak semiclassical states respectively, as well as strong convergence and exponential decay. In addition, positivity of solutions and uniqueness of local maximum points of solutions are also studied. Our theorems extend the results of Ramos and Tavares [Calc. Var. 31 (2008) 1-25], where $f$ and $g$ have polynomial growth. It seems that it is the first attempt to obtain multi-peak semiclassical states for Hamiltonian elliptic system with exponential growth.

math.AP↗

arXiv2024

Mutual Occurrence Ratio of Planets. I. New Clues to Reveal Origins of Hot- and Warm-Jupiter from the RV Sample

Many studies have analyzed planetary occurrence rates and their dependence on the host's properties to provide clues to planet formation, but few have focused on the mutual occurrence ratio of different kinds of planets. Such relations reveal whether and how one type of planet evolves into another, e.g. from a cold Jupiter to a warm or even hot Jupiter, and demonstrate how stellar properties impact the evolution history of planetary systems. We propose a new classification of giant planets, i.e. cold Jupiter(CJ), warm Jupiter(WJ), and hot Jupiter(HJ), according to their position relative to the snow line in the system. Then, we derive their occurrence rates($η_{\rm HJ}$, $η_{\rm WJ}$, $η_{\rm CJ}$) with the detection completeness of RV(Radial Velocity) surveys(HARPS$\&$ CORALIE) considered. Finally, we analyze the correlation between the mutual occurrence ratios, i.e. $η{_{\rm CJ}} / η_{\rm WJ}$, $η{_{\rm CJ}} / η_{\rm HJ}$ or $η{_{\rm WJ}}/η_{\rm HJ}$, and various stellar properties, e.g. effective temperature $T_{\rm eff}$. Our results show that the $η_{\rm HJ}$, $η_{\rm WJ}$ and $η_{\rm CJ}$ are increasing with the increasing $T_{\rm eff}$ when $T_{\rm eff}\in (4600,6600] K$. Furthermore, the mutual occurrence ratio between CJ and WJ, i.e. $η{_{\rm CJ}} /η_{\rm WJ}$, shows a decreasing trend with the increasing $T_{\rm eff}$. But, both $η{_{\rm CJ}}/η_{\rm HJ}$ and $η{_{\rm WJ}}/η_{\rm HJ}$ are increasing when the $T_{\rm eff}$ increases. Further consistency tests reveal that the formation processes of WJ and HJ may be dominated by orbital change mechanisms rather than the in-situ model. However, unlike WJ, which favors gentle disk migration, HJ favors a more violent mechanism that requires further investigation.

astro-ph.EP↗

arXiv2024

A cut-and-project perspective for linearized Bregman iterations

The linearized Bregman iterations (LBreI) and its variants are powerful tools for finding sparse or low-rank solutions to underdetermined linear systems. In this study, we propose a cut-and-project perspective for the linearized Bregman method via a bilevel optimization formulation, along with a new unified algorithmic framework. The new perspective not only encompasses various existing linearized Bregman iteration variants as specific instances, but also allows us to extend the linearized Bregman method to solve more general inverse problems. We provide a completed convergence result of the proposed algorithmic framework, including convergence guarantees to feasible points and optimal solutions, and the sublinear convergence rate. Moreover, we introduce the Bregman distance growth condition to ensure linear convergence. At last, our findings are illustrated via numerical tests.

math.OC↗

arXiv2024

The X-LANCE Technical Report for Interspeech 2024 Speech Processing Using Discrete Speech Unit Challenge

Discrete speech tokens have been more and more popular in multiple speech processing fields, including automatic speech recognition (ASR), text-to-speech (TTS) and singing voice synthesis (SVS). In this paper, we describe the systems developed by the SJTU X-LANCE group for the TTS (acoustic + vocoder), SVS, and ASR tracks in the Interspeech 2024 Speech Processing Using Discrete Speech Unit Challenge. Notably, we achieved 1st rank on the leaderboard in the TTS track both with the whole training set and only 1h training data, with the highest UTMOS score and lowest bitrate among all submissions.

eess.AS↗

arXiv2024

SLSM : An Efficient Strategy for Lazy Schema Migration on Shared-Nothing Databases

By introducing intermediate states for metadata changes and ensuring that at most two versions of metadata exist in the cluster at the same time, shared-nothing databases are capable of making online, asynchronous schema changes. However, this method leads to delays in the deployment of new schemas since it requires waiting for massive data backfill. To shorten the service vacuum period before the new schema is available, this paper proposes a strategy named SLSM for zero-downtime schema migration on shared-nothing databases. Based on the lazy migration of stand-alone databases, SLSM keeps the old and new schemas with the same data distribution, reducing the node communication overhead of executing migration transactions for shared-nothing databases. Further, SLSM combines migration transactions with user transactions by extending the distributed execution plan to allow the data involved in migration transactions to directly serve user transactions, greatly reducing the waiting time of user transactions. Experiments demonstrate that our strategy can greatly reduce the latency of user transactions and improve the efficiency of data migration compared to existing schemes.

cs.DB↗

arXiv2024

Recovery Sets of Subspaces from a Simplex Code

Recovery sets for vectors and subspaces are important in the construction of distributed storage system codes. These concepts are also interesting in their own right. In this paper, we consider the following very basic recovery question: what is the maximum number of possible pairwise disjoint recovery sets for each recovered element? The recovered elements in this work are d-dimensional subspaces of a $k$-dimensional vector space over GF(q). Each server stores one representative for each distinct one-dimensional subspace of the k-dimensional vector space, or equivalently a distinct point of PG(k-1,q). As column vectors, the associated vectors of the stored one-dimensional subspaces form the generator matrix of the $[(q^k -1)/(q-1),k,q^{k-1}]$ simplex code over GF(q). Lower bounds and upper bounds on the maximum number of such recovery sets are provided. It is shown that generally, these bounds are either tight or very close to being tight.

cs.IT↗

arXiv2024

UniCATS: A Unified Context-Aware Text-to-Speech Framework with Contextual VQ-Diffusion and Vocoding

The utilization of discrete speech tokens, divided into semantic tokens and acoustic tokens, has been proven superior to traditional acoustic feature mel-spectrograms in terms of naturalness and robustness for text-to-speech (TTS) synthesis. Recent popular models, such as VALL-E and SPEAR-TTS, allow zero-shot speaker adaptation through auto-regressive (AR) continuation of acoustic tokens extracted from a short speech prompt. However, these AR models are restricted to generate speech only in a left-to-right direction, making them unsuitable for speech editing where both preceding and following contexts are provided. Furthermore, these models rely on acoustic tokens, which have audio quality limitations imposed by the performance of audio codec models. In this study, we propose a unified context-aware TTS framework called UniCATS, which is capable of both speech continuation and editing. UniCATS comprises two components, an acoustic model CTX-txt2vec and a vocoder CTX-vec2wav. CTX-txt2vec employs contextual VQ-diffusion to predict semantic tokens from the input text, enabling it to incorporate the semantic context and maintain seamless concatenation with the surrounding context. Following that, CTX-vec2wav utilizes contextual vocoding to convert these semantic tokens into waveforms, taking into consideration the acoustic context. Our experimental results demonstrate that CTX-vec2wav outperforms HifiGAN and AudioLM in terms of speech resynthesis from semantic tokens. Moreover, we show that UniCATS achieves state-of-the-art performance in both speech continuation and editing.

cs.SD↗