arXiv subjects

Lei Geng

Publications and source records attributed to Lei Geng.

At least 19 recordsLinked to original sources

arXiv2026

Quantics tensor cross interpolation for high-order strong-coupling expansions

Real-time impurity solvers enable the study of transport phenomena and the description of nonequilibrium lattice systems within the framework of dynamical mean-field theory (DMFT). They also provide direct access to the spectral functions of both equilibrium and nonequilibrium systems. A widely used approach is the self-consistent strong-coupling expansion, whose lowest-order implementation corresponds to the non-crossing approximation. Higher-order implementations, however, are computationally demanding because the number of diagram topologies grows factorially with expansion order, while the evaluation of self-energies and Green's functions requires increasingly high-dimensional integrations. Here, we demonstrate that the latter challenge can be mitigated by employing quantics tensor cross interpolation in a variable-separated framework. Compared with the previously used scale-separated approach, the new scheme yields substantially lower bond dimensions and capacitates self-consistent steady-state DMFT calculations up to fourth order. We illustrate its performance with representative results for both equilibrium and photo-doped systems. In addition, we analyze the convergence of the strong-coupling expansion in the challenging noninteracting limit by computing diagrams up to sixth order. At this order, the onset of the asymptotic regime of the strong-coupling expansion becomes apparent, which allows the application of extrapolation techniques.

cond-mat.str-el↗

arXiv2026

SAOT: Self-Supervised Continual Graph Learning with Structure-Aware Optimal Transport

Self-supervised Continual Graph Learning (CGL) aims to successively learn from a graph sequence with different tasks without label supervision - a paradigm that has attracted widespread attention. Most existing self-supervised CGL methods rely on instance-level consistency objectives that enforce stability of individual node (or node-pair) embeddings. Due to optimizing nodes in isolation, these methods fail to maintain global relational structure, causing inter-node correspondences to progressively distort under continual learning. To this end, we propose a novel Structure-Aware Optimal Transport (SAOT) framework that explicitly captures and preserves relational structure within graph representations across sequential tasks. Specifically, SAOT leverages optimal transport theory to capture global inter-node correspondences, thereby facilitating and enhancing graph representation learning. Simultaneously, SAOT incorporates a cross-task knowledge distillation mechanism to preserve the previous structural knowledge. Extensive experiments on four CGL benchmark datasets demonstrate that SAOT outperforms existing self-supervised baselines. In particular, SAOT achieves significant performance gains, improving average accuracy by up to 5% on CoraFull-CL and over 15% on Products-CL compared with state-of-the-art methods in the Class-IL setting.

cs.LG↗

arXiv2026

Nonperturbative Nonlinear Hall Effect in Nonequilibrium Steady States

The nonlinear Hall effect in quantum materials has attracted broad interest, yet most existing studies focus on the weak-field, perturbative regime. Here we develop a nonperturbative approach based on nonequilibrium steady-state Green's functions for dc-field-driven lattice systems, with dissipation and interactions incorporated through self-energies beyond the constant relaxation-time approximation and interband transitions treated alongside their intraband counterparts. Applied to a two-band semimetal model, our approach provides direct access to the strong-field Hall response beyond the nonperturbative crossover where the edge of the nonequilibrium distribution reaches Berry-curvature hot spots, a regime in which constant relaxation-time estimates and Berry curvature dipole calculations become unreliable. We further demonstrate that interaction and electron-phonon self-energies within dynamical mean-field theory can substantially change the Hall signal. Our framework enables quantitative simulations of nonequilibrium nonlinear Hall phenomena and provides guidance for strong-field transport experiments.

cond-mat.str-el↗

arXiv2026

$\eta$-pairing in metallic and particle-hole asymmetric systems

Light-induced superconducting-like states have been reported in several classes of correlated materials. From a theoretical point of view, the induction of $\eta$-pairing is a promising route to nonthermal superconductivity. Numerical studies of photo-doped Mott systems revealed $\eta$-pairing states with very high effective critical temperatures. These investigations were however restricted to particle-hole symmetric states in large-gap Mott insulators, while the experiments were performed on strongly correlated metallic systems. It is thus relevant to explore if $\eta$-pairing also exists in non-particle-hole symmetric setups and in photo-excited metallic states. Here we use steady-state nonequilibrium dynamical mean field theory combined with a strong-coupling impurity solver up to third order to investigate this issue. We find that in the strongly correlated regime with large Mott gap, and for low effective doublon and holon temperatures, $\eta$-pairing is robust against changes in the total filling and an imbalance in the doublon and holon density. An asymmetry in the effective doublon and holon temperatures can however strongly suppress the order parameter. In photo-doped metallic systems with a three-peak structure in the local density of states, $\eta$-pairing can be realized in set-ups with positive doublon and holon temperatures and a population inversion in the low-energy quasi-particle band.

cond-mat.str-el↗

arXiv2026

High-temperature $\eta$-pairing superconductivity in the photodoped Hubbard model

We investigate superconductivity emerging in the photodoped Mott insulating Hubbard model using steady-state dynamical mean-field theory implemented on the real-frequency axis. By employing high-order strong-coupling impurity solvers, we obtain the nonequilibrium phase diagram for photoinduced $\eta$-pairing superconductivity with a remarkably high effective critical temperature. We further identify a superconducting gap in the momentum-resolved spectral function and optical conductivity, providing spectroscopic signatures accessible to experiments. Our results highlight a route to a controllable form of high-temperature superconductivity in nonequilibrium strongly correlated systems, fundamentally distinct from the equilibrium $s$-wave pairing state in the attractive Hubbard model or cuprate-like $d$-wave superconductors.

cond-mat.str-el↗

arXiv2025

Photoinduced excitonic magnetism in a multiorbital Hubbard system

Multiorbital Hubbard models with Hund coupling and crystal-field splitting exhibit an instability toward spin-triplet excitonic order in the parameter regime characterized by strong local spin fluctuations. Upon chemical doping, two distinct types of excitonic ferromagnetism have been reported. Using steady-state nonequilibrium dynamical mean-field theory, we demonstrate that photo-doped half-filled systems can host nonthermal counterparts of these excitonic phases and exhibit a rich phase diagram in the space of photo-doping and crystal field splitting. Photo-doping a spin-triplet excitonic insulator provides a route towards nonequilibrium control of magnetic order.

cond-mat.str-el↗

arXiv2025

Adaptive Sparse Softmax: An Effective and Efficient Softmax Variant

Softmax with the cross entropy loss is the standard configuration for current neural classification models. The gold score for a target class is supposed to be 1, but it is never reachable under the softmax schema. Such a problem makes the training process continue forever and leads to overfitting. Moreover, the "target-approach-1" training goal forces the model to continuously learn all samples, leading to a waste of time in handling some samples which have already been classified correctly with high confidence, while the test goal simply requires the target class of each sample to hold the maximum score. To solve the above weaknesses, we propose the Adaptive Sparse softmax (AS-Softmax) which designs a reasonable and test-matching transformation on top of softmax. For more purposeful learning, we discard the classes with far smaller scores compared with the actual class during training. Then the model could focus on learning to distinguish the target class from its strong opponents, which is also the great challenge in test. In addition, since the training losses of easy samples will gradually drop to 0 in AS-Softmax, we develop an adaptive gradient accumulation strategy based on the masked sample ratio to speed up training. We verify the proposed AS-Softmax on a variety of text multi-class, text multi-label, text token classification, image classification and audio classification tasks with class sizes ranging from 5 to 5000+. The results show that AS-Softmax consistently outperforms softmax and its variants, and the loss of AS-Softmax is remarkably correlated with classification performance in validation. Furthermore, adaptive gradient accumulation strategy can bring about 1.2x training speedup comparing with the standard softmax while maintaining classification effectiveness.

cs.LG↗

arXiv2025

Third-order strong-coupling impurity solver for real-frequency DMFT: Accurate spectral functions for antiferromagnetic and photo-doped states

We present a real-frequency third-order strong-coupling impurity solver which employs quantics tensor cross interpolation (QTCI) for an efficient evaluation of the diagram weights. Applying the method to dynamical mean-field theory (DMFT) calculations of the single-band Hubbard model on the Bethe lattice, we clarify the interaction and temperature range in which the third-order approach yields accurate results. Since the calculations are implemented on the real-time/frequency axis, the detailed structure of spectral functions can be obtained without analytical continuation, as we demonstrate with examples for paramagnetic, antiferromagnetic and photo-doped states. Our work establishes a viable path toward high-order, real-frequency impurity solvers for both equilibrium and non-equilibrium DMFT studies.

cond-mat.str-el↗

arXiv2025

HGOT: Self-supervised Heterogeneous Graph Neural Network with Optimal Transport

Heterogeneous Graph Neural Networks (HGNNs), have demonstrated excellent capabilities in processing heterogeneous information networks. Self-supervised learning on heterogeneous graphs, especially contrastive self-supervised strategy, shows great potential when there are no labels. However, this approach requires the use of carefully designed graph augmentation strategies and the selection of positive and negative samples. Determining the exact level of similarity between sample pairs is non-trivial.To solve this problem, we propose a novel self-supervised Heterogeneous graph neural network with Optimal Transport (HGOT) method which is designed to facilitate self-supervised learning for heterogeneous graphs without graph augmentation strategies. Different from traditional contrastive self-supervised learning, HGOT employs the optimal transport mechanism to relieve the laborious sampling process of positive and negative samples. Specifically, we design an aggregating view (central view) to integrate the semantic information contained in the views represented by different meta-paths (branch views). Then, we introduce an optimal transport plan to identify the transport relationship between the semantics contained in the branch view and the central view. This allows the optimal transport plan between graphs to align with the representations, forcing the encoder to learn node representations that are more similar to the graph space and of higher quality. Extensive experiments on four real-world datasets demonstrate that our proposed HGOT model can achieve state-of-the-art performance on various downstream tasks. In particular, in the node classification task, HGOT achieves an average of more than 6% improvement in accuracy compared with state-of-the-art methods.

cs.LG↗

arXiv2025

Real-frequency TPSC+DMFT investigation of the square-lattice Hubbard model

We investigate the two-dimensional Hubbard model using a real-frequency implementation of the TPSC+DMFT approach. This hybrid method combines the nonlocal correlations captured by the Two-Particle Self-Consistent (TPSC) approach with the local dynamical correlations of Dynamical Mean-Field Theory (DMFT). The results demonstrate that TPSC+DMFT effectively describes pseudogap physics and nonlocal fluctuations in the moderately correlated regime, while also reproducing the Mott insulating state at larger interaction strengths. For doped Mott insulators, we find that TPSC+DMFT captures the evolution of Fermi pockets into Fermi arcs, consistent with the results from cluster DMFT and photoemission studies. These findings highlight the capability of TPSC+DMFT to bridge the gap between weak and strong coupling physics in Hubbard models, providing insights into spin and charge fluctuations, as well as their role in the pseudogap formation.

cond-mat.str-el↗

arXiv2024

Direct May Not Be the Best: An Incremental Evolution View of Pose Generation

Pose diversity is an inherent representative characteristic of 2D images. Due to the 3D to 2D projection mechanism, there is evident content discrepancy among distinct pose images. This is the main obstacle bothering pose transformation related researches. To deal with this challenge, we propose a fine-grained incremental evolution centered pose generation framework, rather than traditional direct one-to-one in a rush. Since proposed approach actually bypasses the theoretical difficulty of directly modeling dramatic non-linear variation, the incurred content distortion and blurring could be effectively constrained, at the same time the various individual pose details, especially clothes texture, could be precisely maintained. In order to systematically guide the evolution course, both global and incremental evolution constraints are elaborately designed and merged into the overall framework. And a novel triple-path knowledge fusion structure is worked out to take full advantage of all available valuable knowledge to conduct high-quality pose synthesis. In addition, our framework could generate a series of valuable byproducts, namely the various intermediate poses. Extensive experiments have been conducted to verify the effectiveness of the proposed approach. Code is available at https://github.com/Xiaofei-CN/Incremental-Evolution-Pose-Generation.

cs.CV↗

arXiv2024

Multi-Scale Subgraph Contrastive Learning

Graph-level contrastive learning, aiming to learn the representations for each graph by contrasting two augmented graphs, has attracted considerable attention. Previous studies usually simply assume that a graph and its augmented graph as a positive pair, otherwise as a negative pair. However, it is well known that graph structure is always complex and multi-scale, which gives rise to a fundamental question: after graph augmentation, will the previous assumption still hold in reality? By an experimental analysis, we discover the semantic information of an augmented graph structure may be not consistent as original graph structure, and whether two augmented graphs are positive or negative pairs is highly related with the multi-scale structures. Based on this finding, we propose a multi-scale subgraph contrastive learning architecture which is able to characterize the fine-grained semantic information. Specifically, we generate global and local views at different scales based on subgraph sampling, and construct multiple contrastive relationships according to their semantic associations to provide richer self-supervised signals. Extensive experiments and parametric analyzes on eight graph classification real-world datasets well demonstrate the effectiveness of the proposed method.

cs.AI↗

arXiv2024

Anomalous photo-induced band renormalization in correlated materials: Case study of Ta$_2$NiSe$_5$

We investigate the anomalous photo-induced band renormalization in correlated materials, exemplified by the case of Ta$_2$NiSe$_5$. The manifestation of this anomaly is characterized by the alternating direction of band shift in response to changes in the laser parameters or electron momentum. We attribute the phenomena to the band inversion of the material and the selective excitation of a high-lying flat band, leading to the competition between the Hartree shift and the order collapse. These findings are based on {\it ab initio} determined effective model for Ta$_2$NiSe$_5$, in which we incorporate high-lying states and the time-dependent GW simulation to follow the non-equilibrium dynamics induced by the laser. Our findings reveal the sensitivity of the non-equilibrium electronic dynamics to the band structure and laser protocols, providing valuable guidance for the selection of suitable materials and lasers in the engineering of band structures.

cond-mat.str-el↗

arXiv2023

KBioXLM: A Knowledge-anchored Biomedical Multilingual Pretrained Language Model

Most biomedical pretrained language models are monolingual and cannot handle the growing cross-lingual requirements. The scarcity of non-English domain corpora, not to mention parallel data, poses a significant hurdle in training multilingual biomedical models. Since knowledge forms the core of domain-specific corpora and can be translated into various languages accurately, we propose a model called KBioXLM, which transforms the multilingual pretrained model XLM-R into the biomedical domain using a knowledge-anchored approach. We achieve a biomedical multilingual corpus by incorporating three granularity knowledge alignments (entity, fact, and passage levels) into monolingual corpora. Then we design three corresponding training tasks (entity masking, relation masking, and passage relation prediction) and continue training on top of the XLM-R model to enhance its domain cross-lingual ability. To validate the effectiveness of our model, we translate the English benchmarks of multiple tasks into Chinese. Experimental results demonstrate that our model significantly outperforms monolingual and multilingual pretrained models in cross-lingual zero-shot and few-shot scenarios, achieving improvements of up to 10+ points. Our code is publicly available at https://github.com/ngwlh-gl/KBioXLM.

cs.CL↗

arXiv2023

RSpell: Retrieval-augmented Framework for Domain Adaptive Chinese Spelling Check

Chinese Spelling Check (CSC) refers to the detection and correction of spelling errors in Chinese texts. In practical application scenarios, it is important to make CSC models have the ability to correct errors across different domains. In this paper, we propose a retrieval-augmented spelling check framework called RSpell, which searches corresponding domain terms and incorporates them into CSC models. Specifically, we employ pinyin fuzzy matching to search for terms, which are combined with the input and fed into the CSC model. Then, we introduce an adaptive process control mechanism to dynamically adjust the impact of external knowledge on the model. Additionally, we develop an iterative strategy for the RSpell framework to enhance reasoning capabilities. We conducted experiments on CSC datasets in three domains: law, medicine, and official document writing. The results demonstrate that RSpell achieves state-of-the-art performance in both zero-shot and fine-tuning scenarios, demonstrating the effectiveness of the retrieval-augmented CSC framework. Our code is available at https://github.com/47777777/Rspell.

cs.CL↗

arXiv2022

General and Domain Adaptive Chinese Spelling Check with Error Consistent Pretraining

The lack of label data is one of the significant bottlenecks for Chinese Spelling Check (CSC). Existing researches use the method of automatic generation by exploiting unlabeled data to expand the supervised corpus. However, there is a big gap between the real input scenario and automatic generated corpus. Thus, we develop a competitive general speller ECSpell which adopts the Error Consistent masking strategy to create data for pretraining. This error consistency masking strategy is used to specify the error types of automatically generated sentences which is consistent with real scene. The experimental result indicates our model outperforms previous state-of-the-art models on the general benchmark. Moreover, spellers often work within a particular domain in real life. Due to lots of uncommon domain terms, experiments on our built domain specific datasets show that general models perform terribly. Inspired by the common practice of input methods, we propose to add an alterable user dictionary to handle the zero-shot domain adaption problem. Specifically, we attach a User Dictionary guided inference module (UD) to a general token classification based speller. Our experiments demonstrate that ECSpell$^{UD}$, namely ECSpell combined with UD, surpasses all the other baselines largely, even approaching the performance on the general benchmark.

cs.CL↗

arXiv2022

Laser-induced electron Fresnel diffraction in the tunneling and over-barrier ionization

The photoelectron momentum distribution in the strong-field ionization has a variety of structures that reveal the complicated dynamics of this process. Recently, we identified a low-energy interference structure in the case of a super-intense extreme ultraviolet (XUV) laser pulse and attributed it to the laser-induced electron Fresnel diffraction. This structure is determined by the laser-induced electron displacement [Geng L et al. 2021 Phys. Rev. A 104(2) L021102]. In the present work, we find that the Fresnel diffraction picture is also present in the tunneling and over-barrier regime of ionization by short pulses. However, the electron displacement is now induced by the electric field component of the laser pulse, instead of by the magnetic field component in the case of the superintense XUV pulse. After corresponding modifications to our quantum and semiclassical models, we find the same physical mechanism of the Fresnel diffraction governs the low-energy interference structures along the laser polarization. The results predicted by the two models agree well with the accurate results from the numerical solution to the time-dependent Schrodinger equation.

physics.atom-ph↗

arXiv2022

Laser-induced electron Fresnel diffraction by XUV pulses at extreme intensity

Ionization of atoms and molecules in laser fields can lead to various interesting interference structures in the photoelectron spectrum. For the case of a super-intense extreme ultraviolet laser pulse, we identify a novel petal-like interference structure in the electron momentum distribution along the direction of the laser field propagation. We show that this structure is quite general and can be attributed to the Fresnel diffraction of the electronic wavepacket by the nucleus. Our results are demonstrated by numerically solving the time-dependent Schrodinger equation of the atomic hydrogen beyond the dipole approximation. By building an analytical model, we find that the electron displacement determines the aforementioned interference pattern. In addition, we establish the physical picture of laser-induced electron Fresnel diffraction which is reinforced by both quantum and semiclassical models.

physics.atom-ph↗