Search arXiv⌕ Search

arXiv subjects

Ming Zhang

Publications and source records attributed to Ming Zhang.

At least 253 records · Page 14Linked to original sources
arXiv2023

RAHNet: Retrieval Augmented Hybrid Network for Long-tailed Graph Classification

Graph classification is a crucial task in many real-world multimedia applications, where graphs can represent various multimedia data types such as images, videos, and social networks. Previous efforts have applied graph neural networks (GNNs) in balanced situations where the class distribution is balanced. However, real-world data typically exhibit long-tailed class distributions, resulting in a bias towards the head classes when using GNNs and limited generalization ability over the tail classes. Recent approaches mainly focus on re-balancing different classes during model training, which fails to explicitly introduce new knowledge and sacrifices the performance of the head classes. To address these drawbacks, we propose a novel framework called Retrieval Augmented Hybrid Network (RAHNet) to jointly learn a robust feature extractor and an unbiased classifier in a decoupled manner. In the feature extractor training stage, we develop a graph retrieval module to search for relevant graphs that directly enrich the intra-class diversity for the tail classes. Moreover, we innovatively optimize a category-centered supervised contrastive loss to obtain discriminative representations, which is more suitable for long-tailed scenarios. In the classifier fine-tuning stage, we balance the classifier weights with two weight regularization techniques, i.e., Max-norm and weight decay. Experiments on various popular benchmarks verify the superiority of the proposed method against state-of-the-art approaches.

cs.LG↗
arXiv2023

Towards Long-Tailed Recognition for Graph Classification via Collaborative Experts

Graph classification, aiming at learning the graph-level representations for effective class assignments, has received outstanding achievements, which heavily relies on high-quality datasets that have balanced class distribution. In fact, most real-world graph data naturally presents a long-tailed form, where the head classes occupy much more samples than the tail classes, it thus is essential to study the graph-level classification over long-tailed data while still remaining largely unexplored. However, most existing long-tailed learning methods in visions fail to jointly optimize the representation learning and classifier training, as well as neglect the mining of the hard-to-classify classes. Directly applying existing methods to graphs may lead to sub-optimal performance, since the model trained on graphs would be more sensitive to the long-tailed distribution due to the complex topological characteristics. Hence, in this paper, we propose a novel long-tailed graph-level classification framework via Collaborative Multi-expert Learning (CoMe) to tackle the problem. To equilibrate the contributions of head and tail classes, we first develop balanced contrastive learning from the view of representation learning, and then design an individual-expert classifier training based on hard class mining. In addition, we execute gated fusion and disentangled knowledge distillation among the multiple experts to promote the collaboration in a multi-expert framework. Comprehensive experiments are performed on seven widely-used benchmark datasets to demonstrate the superiority of our method CoMe over state-of-the-art baselines.

cs.LG↗
arXiv2023

Kinematic topologies of black holes

We investigate the kinematic topologies of light rings (LRs) and massive particle rings (PRs) encircling spherical and axisymmetric black holes. Our results demonstrate that the global topology number of LRs is consistently -1 for asymptotically flat and (Anti-)de Sitter spacetime. Additionally, we show that the global topology of PRs varies, with a value of 0 in asymptotically flat and Anti-de Sitter spacetime but -1 in asymptotic de Sitter spacetime.

gr-qc↗
arXiv2023

Analytical approximations to charged black hole solutions in Einstein-Maxwell-Weyl gravity

The Homotopy Analysis Method (HAM) is a useful method to derive analytical approximate solutions of black holes in modified gravity theories. In this paper, we study the Einstein-Weyl gravity coupled with Maxwell field, and obtain analytical approximation solutions for charged black holes by using the HAM. It is found that the analytical approximate solutions are sufficiently accurate in the entire spacetime outside the black hole's event horizon, and also consistent with numerical ones for charged black holes in the Einstein-Maxwell-Weyl gravity.

gr-qc↗
arXiv2023

Criterion-based Heterogeneous Collaborative Filtering for Multi-behavior Implicit Recommendation

Recent years have witnessed the explosive growth of interaction behaviors in multimedia information systems, where multi-behavior recommender systems have received increasing attention by leveraging data from various auxiliary behaviors such as tip and collect. Among various multi-behavior recommendation methods, non-sampling methods have shown superiority over negative sampling methods. However, two observations are usually ignored in existing state-of-the-art non-sampling methods based on binary regression: (1) users have different preference strengths for different items, so they cannot be measured simply by binary implicit data; (2) the dependency across multiple behaviors varies for different users and items. To tackle the above issue, we propose a novel non-sampling learning framework named Criterion-guided Heterogeneous Collaborative Filtering (CHCF). CHCF introduces both upper and lower thresholds to indicate selection criteria, which will guide user preference learning. Besides, CHCF integrates criterion learning and user preference learning into a unified framework, which can be trained jointly for the interaction prediction of the target behavior. We further theoretically demonstrate that the optimization of Collaborative Metric Learning can be approximately achieved by the CHCF learning framework in a non-sampling form effectively. Extensive experiments on three real-world datasets show the effectiveness of CHCF in heterogeneous scenarios.

cs.IR↗
arXiv2023

Collisional Penrose process of extended test particles near an extremal Kerr black hole

We investigate the collisional Penrose process of extended test particles near extremal Kerr black holes using the pole-dipole-quadrupole approximation. We analyze the motion of the test particles and examine the dynamics and maximum efficiency of energy extraction in this process. Our results demonstrate that the maximum extracted energy in the collisional Penrose process is influenced by the spin s and quadrupolar parameter CES2 of the test particles. Specifically, we observe that, at a fixed collisional position, the energy extraction efficiency decreases as the spin increases for either the pole-dipole or the pole-dipole-quadrupole approximation case. Furthermore, for a fixed spin, the energy extraction efficiency is higher in the pole-dipole-quadrupole approximation compared to the pole-dipole approximation. These findings provide insight into the role of the internal structures of the test particles in the collisional Penrose process.

gr-qc↗
arXiv2023

Tracking Berry curvature effect in molecular dynamics by ultrafast magnetic x-ray scattering

The spin-dependent Berry force is a genuine effect of Berry curvature in molecular dynamics, which can dramatically result in spatial spin separation and change of reaction pathways. However, the way to probe the effect of Berry force remains challenging, because the time-reversal (TR) symmetry required for opposite Berry forces conflicts with TR symmetry breaking spin alignment needed to observe the effect, and the net effect could be transient for a molecular wave packet. We demonstrate that in molecular photodissociation, the dissociation rates can be different for molecules with opposite initial spin directions due to Berry force. We showcase that the spatially separated spin density, which is transiently induced by Berry force as the molecular wave packet passes through conical intersection, can be reconstructed from the circular dichroism (CD) of ultrafast non-resonant magnetic x-ray scattering using free electron lasers.

physics.atom-ph↗
arXiv2023

Strong Cosmic Censorship in accelerating spacetime

It has been established that Strong Cosmic Censorship Conjecture (SCCC) is respected by an asymptotically flat black hole with charge, but violated by a charged asymptotically de Sitter black hole. We studied the instability of the Cauchy horizon for an accelerating black hole in Einstein theory conformally coupled with a scalar field. The black hole is uncharged while having inner, outer, and acceleration horizons. In the limit of vanishing acceleration, denoted by $A \to 0$, it becomes asymptotically flat and extremal, with its acceleration horizon vanishing as well. By exploring the perturbation of the massless scalar field upon the accelerating black hole, whose decay rate is governed by the quasinormal mode spectra, we show that the SCCC is violated in the near-extremal regime and also in the $A\to 0^+$ limit. Our result is the first observation of a black hole violating the SCCC in an almost asymptotically Minkowskian flat regime, as well as the first example of a black hole violating the SCCC with neither charge nor rotation.

gr-qc↗
arXiv2023

HI intensity mapping with MeerKAT: forecast for delay power spectrum measurement using interferometer mode

Neutral hydrogen (HI) intensity mapping (IM) surveys are considered a promising tool for investigating the expansion history of the Universe. In this work, we explore the potential of MeerKAT HI IM observations in interferometer mode to estimate the power spectrum and constrain cosmological parameters within typical dark energy models. We employ an approach called the "delay spectrum," which allows us to separate the weak HI signal from foreground contamination in the frequency domain. Our findings indicate that the choice of survey fields significantly impacts the fractional errors on the power spectrum ($ΔP/P$) within a limited observational time of 10 hours. As the integration time increases from 10 hours to 10,000 hours, $ΔP/P$ progressively decreases until cosmic variance begins to dominate. For a total observation time of 10,000 hours, the lowest $ΔP/P$ at low $k$ can be achieved by tracking 100 points for MeerKAT L-band (900-1200 MHz) and 10 points for MeerKAT UHF-band (580-1000 MHz). Next, we assess the performance of HI IM in constraining typical dark energy models. We find that MeerKAT HI IM survey in interferometer mode demonstrates limited capability in constraining the dark-energy equation of state, even when combined with Planck data.

astro-ph.CO↗
arXiv2023

Erasing-based lossless compression method for streaming floating-point time series

There are a prohibitively large number of floating-point time series data generated at an unprecedentedly high rate. An efficient, compact and lossless compression for time series data is of great importance for a wide range of scenarios. Most existing lossless floating-point compression methods are based on the XOR operation, but they do not fully exploit the trailing zeros, which usually results in an unsatisfactory compression ratio. This paper proposes an Erasing-based Lossless Floating-point compression algorithm, i.e., Elf. The main idea of Elf is to erase the last few bits (i.e., set them to zero) of floating-point values, so the XORed values are supposed to contain many trailing zeros. The challenges of the erasing-based method are three-fold. First, how to quickly determine the erased bits? Second, how to losslessly recover the original data from the erased ones? Third, how to compactly encode the erased data? Through rigorous mathematical analysis, Elf can directly determine the erased bits and restore the original values without losing any precision. To further improve the compression ratio, we propose a novel encoding strategy for the XORed values with many trailing zeros. Furthermore, observing the values in a time series usually have similar significand counts, we propose an upgraded version of Elf named Elf+ by optimizing the significand count encoding strategy, which improves the compression ratio and reduces the running time further. Both Elf and Elf+ work in a streaming fashion. They take only O(N) (where N is the length of a time series) in time and O(1) in space, and achieve a notable compression ratio with a theoretical guarantee. Extensive experiments using 22 datasets show the powerful performance of Elf and Elf+ compared with 9 advanced competitors for both double-precision and single-precision floating-point values.

cs.DS↗
arXiv2023

Surface superconductor-insulator transition induced by an electric field

It is well-known that the electric field can induce phase transitions between superconducting, metallic and insulating states in thin-film materials due to its control of the charge carrier density. Since a similar effect on the charge carriers can also be expected for surfaces of bulk samples, here we investigate the transformation of the surface states in a superconductor under an applied screened electric field. Our study is performed by numerically solving the self-consistent Bogoliubov-de Gennes equations for the one-dimensional attractive Hubbard model. It is found that the surface insulating regime occurs at sufficiently large (but still experimentally accessible) electric fields. Our calculations yield the phase diagram of the surface superconducting, metallic, and insulating states for a wide range of temperatures and applied fields. Our results are in qualitative agreement with the phase diagram obtained by the transport measurements for (Li, Fe)OHFeSe thin flakes [Sci. Bull. 64, 653 (2019); ACS Nano 14, 7513 (2020)].

cond-mat.supr-con↗
arXiv2023

Holographic thermodynamics of rotating black holes

We provide mass/energy formulas for the extended thermodynamics, mixed thermodynamics, and holographic conformal field theory (CFT) thermodynamics for the charged and rotating Kerr-Newman Anti-de Sitter black holes. Then for the CFT thermal states dual to the black hole, we find the first-order phase transitions and criticality phenomena in the canonical ensemble with fixed angular momentum, volume, and central charge. We observe that the CFT states cannot be analogous to the Van der Waals fluids, despite the critical exponents falling into the universality class predicted by the mean field theory. Additionally, we examine the (de)confinement phase transitions within the grand canonical ensemble with fixed angular velocity, volume, and central charge of the CFT. Our findings suggest that the near zero temperature (de)confinement phase transitions can occur with the angular velocity of the CFT that solely depends on the CFT volume.

hep-th↗
arXiv2023

Bulk-boundary thermodynamic equivalence: a topology viewpoint

Setting the cosmological constant to be dynamical, we study the bulk and boundary thermodynamics of charged Anti-de Sitter black holes. We develop mass/energy formulas in terms of thermodynamic state functions for the extended thermodynamics, mixed thermodynamics, and boundary conformal field theory thermodynamics. We employ the residue method to study the topological properties of the phase transitions. Our analysis reveals that the bulk and boundary thermodynamics are topologically equivalent for both criticalities and first-order phase transitions in the canonical ensembles, as well as for the Hawking-Page(-like) phase transitions in the grand canonical ensembles. Additionally, those three kinds of phase transitions are shown to be distinguished by their unique topological charges. Our results exemplify the gravity-gauge duality in terms of topology.

hep-th↗
arXiv2023

Learning on Graphs under Label Noise

Node classification on graphs is a significant task with a wide range of applications, including social analysis and anomaly detection. Even though graph neural networks (GNNs) have produced promising results on this task, current techniques often presume that label information of nodes is accurate, which may not be the case in real-world applications. To tackle this issue, we investigate the problem of learning on graphs with label noise and develop a novel approach dubbed Consistent Graph Neural Network (CGNN) to solve it. Specifically, we employ graph contrastive learning as a regularization term, which promotes two views of augmented nodes to have consistent representations. Since this regularization term cannot utilize label information, it can enhance the robustness of node representations to label noise. Moreover, to detect noisy labels on the graph, we present a sample selection technique based on the homophily assumption, which identifies noisy nodes by measuring the consistency between the labels with their neighbors. Finally, we purify these confident noisy labels to permit efficient semantic graph learning. Extensive experiments on three well-known benchmark datasets demonstrate the superiority of our CGNN over competing approaches.

cs.LG↗
arXiv2023

Towards Semi-supervised Universal Graph Classification

Graph neural networks have pushed state-of-the-arts in graph classifications recently. Typically, these methods are studied within the context of supervised end-to-end training, which necessities copious task-specific labels. However, in real-world circumstances, labeled data could be limited, and there could be a massive corpus of unlabeled data, even from unknown classes as a complementary. Towards this end, we study the problem of semi-supervised universal graph classification, which not only identifies graph samples which do not belong to known classes, but also classifies the remaining samples into their respective classes. This problem is challenging due to a severe lack of labels and potential class shifts. In this paper, we propose a novel graph neural network framework named UGNN, which makes the best of unlabeled data from the subgraph perspective. To tackle class shifts, we estimate the certainty of unlabeled graphs using multiple subgraphs, which facilities the discovery of unlabeled data from unknown categories. Moreover, we construct semantic prototypes in the embedding space for both known and unknown categories and utilize posterior prototype assignments inferred from the Sinkhorn-Knopp algorithm to learn from abundant unlabeled graphs across different subgraph views. Extensive experiments on six datasets verify the effectiveness of UGNN in different settings.

cs.LG↗
arXiv2023

MolXPT: Wrapping Molecules with Text for Generative Pre-training

Generative pre-trained Transformer (GPT) has demonstrates its great success in natural language processing and related techniques have been adapted into molecular modeling. Considering that text is the most important record for scientific discovery, in this paper, we propose MolXPT, a unified language model of text and molecules pre-trained on SMILES (a sequence representation of molecules) wrapped by text. Briefly, we detect the molecule names in each sequence and replace them to the corresponding SMILES. In this way, the SMILES could leverage the information from surrounding text, and vice versa. The above wrapped sequences, text sequences from PubMed and SMILES sequences from PubChem are all fed into a language model for pre-training. Experimental results demonstrate that MolXPT outperforms strong baselines of molecular property prediction on MoleculeNet, performs comparably to the best model in text-molecule translation while using less than half of its parameters, and enables zero-shot molecular generation without finetuning.

cs.CL↗
arXiv2023

Complex-valued neural operator assisted soliton identification

The numerical determination of solitary states is an important topic for such research areas as Bose-Einstein condensates, nonlinear optics, plasma physics, etc. In this paper, we propose a data-driven approach for identifying solitons based on dynamical solutions of real-time differential equations. Our approach combines a machine-learning architecture called the complex-valued neural operator (CNO) with an energy-restricted gradient optimization. The former serves as a generalization of the traditional neural operator to the complex domain, and constructs a smooth mapping between the initial and final states; the latter facilitates the search for solitons by constraining the energy space. We concretely demonstrate this approach on the quasi-one-dimensional Bose-Einstein condensate with homogeneous and inhomogeneous nonlinearities. Our work offers a new idea for data-driven effective modeling and studies of solitary waves in nonlinear physical systems.

cond-mat.quant-gas↗
arXiv2023

Overspinning a rotating black hole in semiclassical gravity with type-A trace anomaly

Recently, Fernandes discovered an analytic solution for rotating black holes in semiclassical gravity induced by the trace anomaly. These solutions exhibit some distinctive characteristics, including a non-spherically symmetric event horizon and violations of the Kerr bound. As a crucial assumption to uphold causality in spacetime, we investigate the validity of the weak cosmic censorship conjecture (WCCC) within this class of solutions with type-A trace anomaly by introducing a test particle on the equatorial plane. Our study reveals three distinct mechanisms that can potentially destroy the event horizon, leading to a violation of the WCCC. Our findings indicate that, with the exception of extremal Kerr, static extremal, and static singular black holes, the WCCC may be violated under the first-order perturbation of the test particle. These results suggest the need for further exploration of modifications to the behavior of the test particle under quantum effects in order to address the violation of the WCCC in this system.

gr-qc↗