Searcharxiv⌕ Search

arXiv subjects

Gang Li

Publications and source records attributed to Gang Li.

At least 307 records · Page 17Linked to original sources
arXiv2023

Modelling two Energetic Storm Particle Events Observed by Solar Orbiter Using the Combined EUHFORIA and iPATH Models

By coupling the EUropean Heliospheric FORcasting Information Asset (EUHFORIA) and the improved Particle Acceleration and Transport in the Heliosphere (iPATH) model, two energetic storm particle (ESP) events, originating from the same active region (AR 13088) and observed by Solar Orbiter (SolO) on August 31 2022 and September 05 2022, are modelled. While both events originated from the same active region, they exhibited notable differences, including: 1) the August ESP event lasted for 7 hours, while the September event persisted for 16 hours; 2) The time intensity profiles for the September event showed a clear cross-over upstream of the shock where the intensity of higher energy protons exceeds those of lower energy protons, leading to positive (``reverse'') spectral indices prior to the shock passage. For both events, our simulations replicate the observed duration of the shock sheath, depending on the deceleration history of the CME. Imposing different choices of escaping length scale, which is related to the decay of upstream turbulence, the modelled time intensity profiles prior to the shock arrival also agree with observations. In particular, the cross-over of this time profile in the September event is well reproduced. We show that a ``reverse'' upstream spectrum is the result of the interplay between two length scales. One characterizes the decay of upstream shock accelerated particles, which are controlled by the energy-dependent diffusion coefficient, and the other characterizes the decay of upstream turbulence power, which is related to the process of how streaming protons upstream of the shock excite Alfvén waves. Simulations taking into account real-time background solar wind, the dynamics of the CME propagation, and upstream turbulence at the shock front are necessary to thoroughly understand the ESP phase of large SEP events.

astro-ph.SR↗
arXiv2023

Beam test of a 180 nm CMOS Pixel Sensor for the CEPC vertex detector

The proposed Circular Electron Positron Collider (CEPC) imposes new challenges for the vertex detector in terms of pixel size and material budget. A Monolithic Active Pixel Sensor (MAPS) prototype called TaichuPix, based on a column drain readout architecture, has been developed to address the need for high spatial resolution. In order to evaluate the performance of the TaichuPix-3 chips, a beam test was carried out at DESY II TB21 in December 2022. Meanwhile, the Data Acquisition (DAQ) for a muti-plane configuration was tested during the beam test. This work presents the characterization of the TaichuPix-3 chips with two different processes, including cluster size, spatial resolution, and detection efficiency. The analysis results indicate the spatial resolution better than 5 $μm$ and the detection efficiency exceeds 99.5 % for both TaichuPix-3 chips with the two different processes.

physics.ins-det↗
arXiv2023

Holistic Evaluation of GPT-4V for Biomedical Imaging

In this paper, we present a large-scale evaluation probing GPT-4V's capabilities and limitations for biomedical image analysis. GPT-4V represents a breakthrough in artificial general intelligence (AGI) for computer vision, with applications in the biomedical domain. We assess GPT-4V's performance across 16 medical imaging categories, including radiology, oncology, ophthalmology, pathology, and more. Tasks include modality recognition, anatomy localization, disease diagnosis, report generation, and lesion detection. The extensive experiments provide insights into GPT-4V's strengths and weaknesses. Results show GPT-4V's proficiency in modality and anatomy recognition but difficulty with disease diagnosis and localization. GPT-4V excels at diagnostic report generation, indicating strong image captioning skills. While promising for biomedical imaging AI, GPT-4V requires further enhancement and validation before clinical deployment. We emphasize responsible development and testing for trustworthy integration of biomedical AGI. This rigorous evaluation of GPT-4V on diverse medical images advances understanding of multimodal large language models (LLMs) and guides future work toward impactful healthcare applications.

eess.IV↗
arXiv2023

On Uniqueness And Existence of Conformally Compact Einstein Metrics with Homogeneous Conformal Infinity. II

In this paper we show that for an $\text{Sp}(k+1)$ invariant metric $\hat{g}$ on $\mathbb{S}^{4k+3}$ $(k\geq 1)$ close to the round metric, the conformally compact Einstein (CCE) manifold $(M, g)$ with $(\mathbb{S}^{4k+3}, [\hat{g}])$ as its conformal infinity is unique up to isometries. Moreover, by the result in [LiQingShi], $g$ is the Graham-Lee metric on the unit ball $B_1\subset \mathbb{R}^{4k+4}$. We also give an a priori estimate on the Einstein metric $g$. Based on the estimate and Graham-Lee and Lee's seminal perturbation result, we use the continuity method directly to obtain an existence result of the non-positively curved CCE metric with prescribed conformal infinity $(\mathbb{S}^{4k+3}, [\hat{g}])$ when the metric $\hat{g}$ is $\text{Sp}(k+1)$-invariant. We also generalize the results to the case of conformal infinity $(\mathbb{S}^{15},[\hat{g}])$ with $\hat{g}$ a Spin$(9)$-invariant metric in the appendix.

math.DG↗
arXiv2023

Cross-correlation of cosmic voids with thermal Sunyaev-Zel'dovich data

We provide a measurement of the deficit in the Sunyaev-Zel'dovich Compton-$y$ signal towards cosmic voids, by stacking a catalogue of 97,090 voids constructed with BOSS-DR12 data, on the $y$ maps built on data from the Atacama Cosmology Telescope (ACT) DR4 and the Planck satellite. We detect the void signal with a significance of $7.3\,σ$ with ACT and $9.7\,σ$ with Planck, obtaining agreements in the associated void radial $y$ profiles extracted from both maps. The inner-void profile (for angular separations within the void angular radius) is reconstructed with significances of $4.7σ$ and $6.1σ$ with ACT and Planck, respectively; we model such profile using a simple model that assumes uniform gas (under)density and temperature, which enables us to place constraints on the product $(-δ_{\rm v}T_{\rm e})$ of the void density contrast (negative) and the electron temperature. The best-fit values from the two data sets are $(-δ_{\rm v}T_{\rm e})=(6.5\pm 2.3)\times 10^{5}\,\text{K}$ for ACT and $(8.6 \pm 2.1)\times 10^{5}\,\text{K}$ for Planck ($68\%$ C.L.), which are in good agreement under uncertainty. The data allow us to place lower limits on the expected void electron temperature at $2.7\times10^5\,\text{K}$ with ACT and $5.1\times10^5\,\text{K}$ with Planck ($95\%$ C.L.); these results can transform into upper limits for the ratio between the void electron density and the cosmic mean as $n^{\rm v}_{\rm e}/\bar{n}_{\rm e}\leqslant 0.73$ and $0.49$ ($95\%$ C.L.), respectively. Our findings prove the feasibility of using tSZ observations to constrain the gas properties inside cosmic voids, and confirm that voids are under-pressured regions compared to their surroundings.

astro-ph.CO↗
arXiv2023

Progress and outlook on advanced fly scans based on Mamba

Development related to PandABox-based fly scans is an important part of the active work on Mamba, the software framework for beamline experiments at the High Energy Photon Source (HEPS); presented in this paper is the progress of our development, and some outlook for advanced fly scans based on knowledge learned during the process. By treating fly scans as a collaboration between a few loosely coupled subsystems - motors / mechanics, detectors / data processing, sequencer devices like PandABox - systematic analyses of issues in fly scans are conducted. Interesting products of these analyses include a general-purpose software-based fly-scan mechanism, a general way to design undulator-monochromator fly scans, a sketch of how to practically implement online tuning of fly-scan behaviours based on processing of the data acquired, and many more. Based on the results above, an architectural discussion on >=10kHz fly scans is given.

physics.ins-det↗
arXiv2023

Variability of interplanetary shock and associated energetic particle properties as a function of the time window around the shock

We study the effect of sampling windows on derived shock and associated energetic storm particle (ESP) properties in 296 fast-forward interplanetary shocks using ACE measurements at 1 au between 02/1998 - 08/2013. We vary the time windows from 2-mins to 20-mins for the shock properties and from 2-mins to 540-mins for ESP properties. Variability is quantified by the median absolute deviation (MAD) statistic. We find that the magnetic, density, and temperature compression ratios vary from their median values by 17.03%, 20.05%, 25.91%, respectively; shock speed by 16.26%, speed jump by 45.46%, Alfvenic Mach number by 31.53%, and shock obliquity by 24.25%. Spectral indices in the 2-min to 540-min windows downstream of the shock vary from the median value of 1.79 by 26.05%, and by 30.53% from the 1.70 median value upstream of the shock. Similarity of ESP spectral indices upstream and downstream of the shock suggest that these ESP populations are likely locally accelerated at the shock. Furthermore, we find that for a moving sampling window around the shock, values for the density ratio hold for ~10-mins; the magnetic ratio and shock speed jump hold for ~30-mins, and ~60-mins, respectively. Fixing the upstream window to 2-mins and moving only in the downstream direction, then the density ratio holds for ~60-mins downstream, magnetic ratio for ~30-mins, and the shock speed jump holds for ~110-mins. Beyond these time windows, derived shock properties no longer representative of shock properties. These results provide constraints for modeling and forecasting efforts of shock and ESP-associated properties.

physics.space-ph↗
arXiv2023

A Zero-Shot Language Agent for Computer Control with Structured Reflection

Large language models (LLMs) have shown increasing capacity at planning and executing a high-level goal in a live computer environment (e.g. MiniWoB++). To perform a task, recent works often require a model to learn from trace examples of the task via either supervised learning or few/many-shot prompting. Without these trace examples, it remains a challenge how an agent can autonomously learn and improve its control on a computer, which limits the ability of an agent to perform a new task. We approach this problem with a zero-shot agent that requires no given expert traces. Our agent plans for executable actions on a partially observed environment, and iteratively progresses a task by identifying and learning from its mistakes via self-reflection and structured thought management. On the easy tasks of MiniWoB++, we show that our zero-shot agent often outperforms recent SoTAs, with more efficient reasoning. For tasks with more complexity, our reflective agent performs on par with prior best models, even though previous works had the advantages of accessing expert traces or additional screen information.

cs.CL↗
arXiv2023

Enhancing the quantum entanglement and EPR steering of a coupled optomechanical system with a squeezed vacuum field

Quantum entanglement and Einstein-Podolsky-Rosen (EPR) steering are valuable resources in quantum information processing. How to enhance the quantum entanglement and EPR steering of coupled optomechanical systems with a weak squeezed vacuum field are studied when the displacement of detuning induced by the mechanical mode is considered. Compared with the condition that the system interacts with a vacuum environment, the quantum entanglement and EPR steering are stronger when the squeezed vacuum field is applied. A squeezed vacuum field with a large degree is not beneficial to enhance the quantum entanglement and EPR steering. Rather than the squeezing parameter of the squeezed vacuum field, the reference phase plays a vital role in this model.

quant-ph↗
arXiv2023

AUC-mixup: Deep AUC Maximization with Mixup

While deep AUC maximization (DAM) has shown remarkable success on imbalanced medical tasks, e.g., chest X-rays classification and skin lesions classification, it could suffer from severe overfitting when applied to small datasets due to its aggressive nature of pushing prediction scores of positive data away from that of negative data. This paper studies how to improve generalization of DAM by mixup data augmentation -- an approach that is widely used for improving generalization of the cross-entropy loss based deep learning methods. %For overfitting issues arising from limited data, the common approach is to employ mixup data augmentation to boost the models' generalization performance by enriching the training data. However, AUC is defined over positive and negative pairs, which makes it challenging to incorporate mixup data augmentation into DAM algorithms. To tackle this challenge, we employ the AUC margin loss and incorporate soft labels into the formulation to effectively learn from data generated by mixup augmentation, which is referred to as the AUC-mixup loss. Our experimental results demonstrate the effectiveness of the proposed AUC-mixup methods on imbalanced benchmark and medical image datasets compared to standard DAM training methods.

cs.LG↗
arXiv2023

The Road to On-board Change Detection: A Lightweight Patch-Level Change Detection Network via Exploring the Potential of Pruning and Pooling

Existing satellite remote sensing change detection (CD) methods often crop original large-scale bi-temporal image pairs into small patch pairs and then use pixel-level CD methods to fairly process all the patch pairs. However, due to the sparsity of change in large-scale satellite remote sensing images, existing pixel-level CD methods suffer from a waste of computational cost and memory resources on lots of unchanged areas, which reduces the processing efficiency of on-board platform with extremely limited computation and memory resources. To address this issue, we propose a lightweight patch-level CD network (LPCDNet) to rapidly remove lots of unchanged patch pairs in large-scale bi-temporal image pairs. This is helpful to accelerate the subsequent pixel-level CD processing stage and reduce its memory costs. In our LPCDNet, a sensitivity-guided channel pruning method is proposed to remove unimportant channels and construct the lightweight backbone network on basis of ResNet18 network. Then, the multi-layer feature compression (MLFC) module is designed to compress and fuse the multi-level feature information of bi-temporal image patch. The output of MLFC module is fed into the fully-connected decision network to generate the predicted binary label. Finally, a weighted cross-entropy loss is utilized in the training process of network to tackle the change/unchange class imbalance problem. Experiments on two CD datasets demonstrate that our LPCDNet achieves more than 1000 frames per second on an edge computation platform, i.e., NVIDIA Jetson AGX Orin, which is more than 3 times that of the existing methods without noticeable CD performance loss. In addition, our method reduces more than 60% memory costs of the subsequent pixel-level CD processing stage.

cs.CV↗
arXiv2023

Self-supervised Fetal MRI 3D Reconstruction Based on Radiation Diffusion Generation Model

Although the use of multiple stacks can handle slice-to-volume motion correction and artifact removal problems, there are still several problems: 1) The slice-to-volume method usually uses slices as input, which cannot solve the problem of uniform intensity distribution and complementarity in regions of different fetal MRI stacks; 2) The integrity of 3D space is not considered, which adversely affects the discrimination and generation of globally consistent information in fetal MRI; 3) Fetal MRI with severe motion artifacts in the real-world cannot achieve high-quality super-resolution reconstruction. To address these issues, we propose a novel fetal brain MRI high-quality volume reconstruction method, called the Radiation Diffusion Generation Model (RDGM). It is a self-supervised generation method, which incorporates the idea of Neural Radiation Field (NeRF) based on the coordinate generation and diffusion model based on super-resolution generation. To solve regional intensity heterogeneity in different directions, we use a pre-trained transformer model for slice registration, and then, a new regionally Consistent Implicit Neural Representation (CINR) network sub-module is proposed. CINR can generate the initial volume by combining a coordinate association map of two different coordinate mapping spaces. To enhance volume global consistency and discrimination, we introduce the Volume Diffusion Super-resolution Generation (VDSG) mechanism. The global intensity discriminant generation from volume-to-volume is carried out using the idea of diffusion generation, and CINR becomes the deviation intensity generation network of the volume-to-volume diffusion model. Finally, the experimental results on real-world fetal brain MRI stacks demonstrate the state-of-the-art performance of our method.

eess.IV↗
arXiv2023

3DDesigner: Towards Photorealistic 3D Object Generation and Editing with Text-guided Diffusion Models

Text-guided diffusion models have shown superior performance in image/video generation and editing. While few explorations have been performed in 3D scenarios. In this paper, we discuss three fundamental and interesting problems on this topic. First, we equip text-guided diffusion models to achieve 3D-consistent generation. Specifically, we integrate a NeRF-like neural field to generate low-resolution coarse results for a given camera view. Such results can provide 3D priors as condition information for the following diffusion process. During denoising diffusion, we further enhance the 3D consistency by modeling cross-view correspondences with a novel two-stream (corresponding to two different views) asynchronous diffusion process. Second, we study 3D local editing and propose a two-step solution that can generate 360-degree manipulated results by editing an object from a single view. Step 1, we propose to perform 2D local editing by blending the predicted noises. Step 2, we conduct a noise-to-text inversion process that maps 2D blended noises into the view-independent text embedding space. Once the corresponding text embedding is obtained, 360-degree images can be generated. Last but not least, we extend our model to perform one-shot novel view synthesis by fine-tuning on a single image, firstly showing the potential of leveraging text guidance for novel view synthesis. Extensive experiments and various applications show the prowess of our 3DDesigner. The project page is available at https://3ddesigner-diffusion.github.io/.

cs.CV↗
arXiv2023

Constraints on primordial black holes in the mixed dark matter scenarios using the ratio $\rm (^3{He}+D)/H$

We derive the upper limit on the dark matter (DM) fraction in primordial black holes (PBHs) in the mixed DM scenarios. In this scenarios, a PBH can accrete weakly interacting massive particles (WIMPs) to form a ultracompact minihalo (UCMH) with a density profile of $ρ_{\rm DM}(r)\sim r^{-9/4}$. The energy released from UCMHs due to dark matter annihilation has influence on the photodissociation of $^{4}{\rm He}$, producing the $^{3}{\rm He}$ and the D. By requiring that the ratio $\rm (^3{He}+D)/H$ caused by UCMHs does not exceed the measured value, we derive the upper limit on the dark matter fraction in PBHs. For the canonical value of DM thermally averaged annihilation cross section $\left<σv\right>=3\times 10^{-26}\rm cm^{3}s^{-1}$, we find that the upper limit is $f_{\rm PBH} < 0.35(0.75)$ for DM mass $m_χ=1(10)~\rm GeV$. Compared with other limits obtained by different astronomical measurements, although our limit is not the strongest, we provide a different way of constraining the cosmological abundance of PBHs.

astro-ph.CO↗
arXiv2023

Update of hadronic decays of $J/ψ$ and $ψ(2S)$ though virtual photons

The hadronic decay branching ratios of $J/ψ$ and $ψ(2S)$ through virtual photons, $B(J/ψ, ψ(2S) \rightarrow γ^*\rightarrow \text{hadrons})$, are updated by using the latest published measurements of the $R$ value and the branching ratios of $J/ψ, ψ(2S) \rightarrow l^+l^-$. Their respective precision increases by about 4 and 3 times.

hep-ex↗
arXiv2023

Internal magnetic fields in 13 red giants detected by asteroseismology

While surface fields have been measured for stars across the HR diagram, internal magnetic fields remain largely unknown. The recent seismic detection of magnetic fields in the cores of several Kepler red giants has opened a new avenue to understand better the origin of magnetic fields and their impact on stellar structure and evolution. We aim to use asteroseismology to systematically search for internal magnetic fields in red giant stars and to determine the strengths and geometries of these fields. Magnetic fields are known to break the symmetry of rotational multiplets. In red giants, oscillation modes are mixed, behaving as pressure modes in the envelope and as gravity modes in the core. Magnetism-induced asymmetries are expected to be stronger for g-dominated modes than for p-dominated modes and to decrease with frequency. After collecting a sample of 2500 Kepler red giant stars with clear mixed-mode patterns, we specifically searched for targets among 1200 stars with dipole triplets. We identified 13 stars exhibiting clear asymmetric multiplets and measured their parameters, especially the asymmetry parameter and the magnetic frequency shift. By combining these estimates with best-fitting stellar models, we measured average core magnetic fields ranging from 20 to 150kG, corresponding to 5% to 30% of the critical field strengths. We showed that the detected core fields have various horizontal geometries, some of which significantly differ from a dipolar configuration. We found that the field strengths decrease with stellar evolution, despite the fact that the cores of these stars are contracting. Even though these stars have strong internal magnetic fields, they display normal core rotation rates, suggesting no significantly different histories of angular momentum transport compared to other red giant stars. We also discuss the possible origin of the detected fields.

astro-ph.SR↗
arXiv2023

When AUC meets DRO: Optimizing Partial AUC for Deep Learning with Non-Convex Convergence Guarantee

In this paper, we propose systematic and efficient gradient-based methods for both one-way and two-way partial AUC (pAUC) maximization that are applicable to deep learning. We propose new formulations of pAUC surrogate objectives by using the distributionally robust optimization (DRO) to define the loss for each individual positive data. We consider two formulations of DRO, one of which is based on conditional-value-at-risk (CVaR) that yields a non-smooth but exact estimator for pAUC, and another one is based on a KL divergence regularized DRO that yields an inexact but smooth (soft) estimator for pAUC. For both one-way and two-way pAUC maximization, we propose two algorithms and prove their convergence for optimizing their two formulations, respectively. Experiments demonstrate the effectiveness of the proposed algorithms for pAUC maximization for deep learning on various datasets.

cs.LG↗
arXiv2023

Cartoondiff: Training-free Cartoon Image Generation with Diffusion Transformer Models

Image cartoonization has attracted significant interest in the field of image generation. However, most of the existing image cartoonization techniques require re-training models using images of cartoon style. In this paper, we present CartoonDiff, a novel training-free sampling approach which generates image cartoonization using diffusion transformer models. Specifically, we decompose the reverse process of diffusion models into the semantic generation phase and the detail generation phase. Furthermore, we implement the image cartoonization process by normalizing high-frequency signal of the noisy image in specific denoising steps. CartoonDiff doesn't require any additional reference images, complex model designs, or the tedious adjustment of multiple parameters. Extensive experimental results show the powerful ability of our CartoonDiff. The project page is available at: https://cartoondiff.github.io/

cs.CV↗