arXiv subjects

Geng Li

Publications and source records attributed to Geng Li.

At least 55 records · Page 3Linked to original sources

arXiv2025

Decoupling Coherent and Particle-like Phonon Transport through Bonding Hierarchy in Soft Superionic Crystals

Within the framework of the unified theory thermal transport model, the competing contributions of coherent and incoherent terms create a trade-off relationship, posing substantial challenges to achieving a reduction in overall $\rm κ_L$. In this work, we theoretically demonstrate that the superionic crystals X$_6$Re$_6$S$_8$I$_8$ (X = Rb, Cs) exhibit ultralow glass-like and particle-like thermal conductivities. The weak interactions between free alkali metal ions X$^+$ (X = Rb, Cs) and I$^-$ anions induce pronounced lattice anharmonicity, which enhances phonon scattering and suppresses group velocities, thereby reducing the particle-like thermal conductivity ($\rm κ_p$). Concurrently, the significant bonding heterogeneity within the [Re$_6$S$_8$I$_6$]$^{4-}$ clusters promotes phonon dispersion flattening and low-frequency phonon localization. The resulting discretized phonon flat bands substantially diminish the glass-like thermal conductivity ($\rm κ_c$). At room temperature, the total $\rm κ_L$ of X$_6$Re$_6$S$_8$I$_8$ (X = Rb, Cs) falls below 0.2 Wm$^{-1}$K$^{-1}$. Furthermore, the bonding characteristics between X$^+$ and I$^{-1}$ anions induce an anomalous cation mass-independent stiffening of low-frequency phonon branches in this system, resulting in counterintuitive thermal transport behavior. This work elucidates fundamental mechanisms governing heat transfer in ultralow $\rm κ_L$ materials and establishes novel pathways for transcending conventional thermal conductivity limitations.

cond-mat.mtrl-sci↗

arXiv2025

Analyzing and Boosting the Power of Fine-Grained Visual Recognition for Multi-modal Large Language Models

Multi-modal large language models (MLLMs) have shown remarkable abilities in various visual understanding tasks. However, MLLMs still struggle with fine-grained visual recognition (FGVR), which aims to identify subordinate-level categories from images. This can negatively impact more advanced capabilities of MLLMs, such as object-centric visual question answering and reasoning. In our study, we revisit three quintessential capabilities of MLLMs for FGVR, including object information extraction, category knowledge reserve, object-category alignment, and position of the root cause as a misalignment problem. To address this issue, we present Finedefics, an MLLM that enhances the model's FGVR capability by incorporating informative attribute descriptions of objects into the training phase. We employ contrastive learning on object-attribute pairs and attribute-category pairs simultaneously and use examples from similar but incorrect categories as hard negatives, naturally bringing representations of visual objects and category names closer. Extensive evaluations across multiple popular FGVR datasets demonstrate that Finedefics outperforms existing MLLMs of comparable parameter sizes, showcasing its remarkable efficacy. The code is available at https://github.com/PKU-ICST-MIPL/Finedefics_ICLR2025.

cs.CV↗

arXiv2025

In situ engineering hexagonal boron nitride in van der Waals heterostructures with selective SF6 etching

Van der Waals heterostructures are at the forefront in materials heterostructure engineering, offering the ultimate control in layer selectivity and capability to combine virtually any material. Hexagonal boron nitride (hBN), the most commonly used dielectric material, has proven indispensable in this field, allowing the encapsulation of active 2D materials preserving their exceptional electronic quality. However, not all device applications require full encapsulation but rather require open surfaces, or even selective patterning of hBN layers. Here, we report on a procedure to engineer top hBN layers within van der Waals heterostructures while preserving the underlying active 2D layers. Using a soft selective SF6 etching combined with a series of pre and post-etching treatments, we demonstrate that pristine surfaces can be exposed with atomic flatness while preserving the active layers electronic quality. We benchmark our technique using graphene encapsulated with hBN Hall bar devices. Using Raman spectroscopy combined with quantum transport, we show high quality can be preserved in etched regions by demonstrating low temperature carrier mobilities of 200,000 cm2Vs-1, ballistic transport probed through magnetic focusing, and intrinsic room temperature phonon-limited mobilities. Atomic force microscopy brooming and O2 plasma cleaning are identified as key pre-etching steps to obtaining pristine open surfaces while preserving electronic quality. The technique provides a clean method for opening windows into mesoscopic van der Waals devices that can be used for local probe experiments, patterning top hBN in situ, and exposing 2D layers to their environment for sensing applications.

cond-mat.mes-hall↗

arXiv2024

UniRiT: Towards Few-Shot Non-Rigid Point Cloud Registration

Non-rigid point cloud registration is a critical challenge in 3D scene understanding, particularly in surgical navigation. Although existing methods achieve excellent performance when trained on large-scale, high-quality datasets, these datasets are prohibitively expensive to collect and annotate, e.g., organ data in authentic medical scenarios. With insufficient training samples and data noise, existing methods degrade significantly since non-rigid patterns are more flexible and complicated than rigid ones, and the distributions across samples are more distinct, leading to higher difficulty in representation learning with few data. In this work, we aim to deal with this challenging few-shot non-rigid point cloud registration problem. Based on the observation that complex non-rigid transformation patterns can be decomposed into rigid and small non-rigid transformations, we propose a novel and effective framework, UniRiT. UniRiT adopts a two-step registration strategy that first aligns the centroids of the source and target point clouds and then refines the registration with non-rigid transformations, thereby significantly reducing the problem complexity. To validate the performance of UniRiT on real-world datasets, we introduce a new dataset, MedMatch3D, which consists of real human organs and exhibits high variability in sample distribution. We further establish a new challenging benchmark for few-shot non-rigid registration. Extensive empirical results demonstrate that UniRiT achieves state-of-the-art performance on MedMatch3D, improving the existing best approach by 94.22%.

cs.CV↗

arXiv2024

Terahertz photocurrent probe of quantum geometry and interactions in magic-angle twisted bilayer graphene

Moiré materials represent strongly interacting electron systems bridging topological and correlated physics. Despite significant advances, decoding wavefunction properties underlying the quantum geometry remains challenging. Here, we utilize polarization-resolved photocurrent measurements to probe magic-angle twisted bilayer graphene, leveraging its sensitivity to the Berry connection that encompasses quantum "textures" of electron wavefunctions. Using terahertz light resonant with optical transitions of its flat bands, we observe bulk photocurrents driven by broken symmetries and reveal the interplay between electron interactions and quantum geometry. We observe inversion-breaking gapped states undetectable through quantum transport, sharp changes in the polarization axes caused by interaction-induced band renormalization, and recurring photocurrent patterns at integer fillings of the moiré unit cell that track the evolution of quantum geometry through the cascade of phase transitions. The large and tunable terahertz response intrinsic to flat-band systems offers direct insights into the quantum geometry of interacting electrons and paves the way for innovative terahertz quantum technologies.

cond-mat.mes-hall↗

arXiv2024

GERA: Geometric Embedding for Efficient Point Registration Analysis

Point cloud registration aims to provide estimated transformations to align point clouds, which plays a crucial role in pose estimation of various navigation systems, such as surgical guidance systems and autonomous vehicles. Despite the impressive performance of recent models on benchmark datasets, many rely on complex modules like KPConv and Transformers, which impose significant computational and memory demands. These requirements hinder their practical application, particularly in resource-constrained environments such as mobile robotics. In this paper, we propose a novel point cloud registration network that leverages a pure MLP architecture, constructing geometric information offline. This approach eliminates the computational and memory burdens associated with traditional complex feature extractors and significantly reduces inference time and resource consumption. Our method is the first to replace 3D coordinate inputs with offline-constructed geometric encoding, improving generalization and stability, as demonstrated by Maximum Mean Discrepancy (MMD) comparisons. This efficient and accurate geometric representation marks a significant advancement in point cloud analysis, particularly for applications requiring fast and reliability.

cs.CV↗

arXiv2024

Thermodynamic Geometric Control of Active Matter

Active matter represents a class of non-equilibrium systems that constantly dissipate energy to produce directed motion. The thermodynamic control of active matter holds great potential for advancements in synthetic molecular motors, targeted drug delivery, and adaptive smart materials. However, the inherently non-equilibrium nature of active matter poses a significant challenge in achieving optimal control with minimal energy cost. In this work, we extend the concept of thermodynamic geometry, traditionally applied to passive systems, to active matter, proposing a systematic geometric framework for minimizing energy cost in non-equilibrium driving processes. We derive a cost metric that defines a Riemannian manifold for control parameters, enabling the use of powerful geometric tools to determine optimal control protocols. The geometric perspective reveals that, unlike in passive systems, minimizing energy cost in active systems involves a trade-off between intrinsic and external dissipation, leading to an optimal transportation speed that coincides with the self-propulsion speed of active matter. This insight enriches the broader concept of thermodynamic geometry. We demonstrate the application of this approach by optimizing the performance of an active monothermal engine within this geometric framework.

cond-mat.stat-mech↗

arXiv2024

Ferromagnetism and correlated insulating states in monolayer Mo33Te56

Kagome lattices have an inherent two-dimensional nature. Despite previous realizations in the monolayer limit, their abilities to drive emergent electronic states such as correlated insulators have remained unobserved. Here, we report the experimental realization of a new structural phase of monolayer Mo33Te56, characterized by its virtually global uniformity as a mirror-twin boundary loop superlattice embedded in an H-MoTe2 monolayer. Through a combination of scanning tunnelling microscopy (STM) and theoretical calculations, we unveil a kagome geometry along with multiple associated sets of kagome flat bands. Crucially, the partial filling of these kagome bands induces ferromagnetism as revealed by spin-polarized STM, and leads to a correlated insulating state exhibiting a hard gap as large as 15 meV. Our findings represent a major advance in kagome materials, offering a framework with clearer band structures and more intrinsic two-dimensional properties for exploring flat-band physics.

cond-mat.mtrl-sci↗

arXiv2024

A deep neural network framework for dynamic multi-valued mapping estimation and its applications

This paper addresses the problem of modeling and estimating dynamic multi-valued mappings. While most mathematical models provide a unique solution for a given input, real-world applications often lack deterministic solutions. In such scenarios, estimating dynamic multi-valued mappings is necessary to suggest different reasonable solutions for each input. This paper introduces a deep neural network framework incorporating a generative network and a classification component. The objective is to model the dynamic multi-valued mapping between the input and output by providing a reliable uncertainty measurement. Generating multiple solutions for a given input involves utilizing a discrete codebook comprising finite variables. These variables are fed into a generative network along with the input, producing various output possibilities. The discreteness of the codebook enables efficient estimation of the output's conditional probability distribution for any given input using a classifier. By jointly optimizing the discrete codebook and its uncertainty estimation during training using a specially designed loss function, a highly accurate approximation is achieved. The effectiveness of our proposed framework is demonstrated through its application to various imaging problems, using both synthetic and real imaging data. Experimental results show that our framework accurately estimates the dynamic multi-valued mapping with uncertainty estimation.

cs.CV↗

arXiv2024

Orbital Stability Study of the Taiji Space Gravitational Wave Detector

Space-based gravitational wave detection is extremely sensitive to disturbances. The Keplerian configuration cannot accurately reflect the variations in spacecraft configuration. Planetary gravitational disturbances are one of the main sources. Numerical simulation is an effective method to investigate the impact of perturbation on spacecraft orbits. This study shows that, in the context of the Taiji project, Earth's gravity is an essential factor in the change in heliocentric formation configuration, contributing to the relative acceleration between spacecrafts in the order of $\mathcal O(10^{-6})\,{\rm m\cdot s^{-2}}$. Considering 00:00:00 on 27 October 2032 as the initial orbiting moment, under the influence of Earth's gravitational perturbation, the maximum relative change in armlengths and variation rates of armlengths for Taiji is $1.6\times 10^{5}\,{\rm km}$, $32\,{\rm m\cdot s^{-1}}$, respectively, compared with the unperturbed Keplerian orbit. Additionally, by considering the gravitational perturbations of Venus and Jupiter, the armlength and relative velocity for Taiji are reduced by $16.01\%$ and $17.45\%$, respectively, compared with when only considering that of Earth. The maximum amplitude of the formation motion indicator changes with the orbit entry time. Results show that the relative velocity increase between the spacecrafts is minimal when the initial orbital moment occurs in July. Moreover, the numerical simulation results are inconsistent when using different ephemerides. The differences between ephemerides DE440 and DE430 are smaller than those between DE440 and DE421.

gr-qc↗

arXiv2024

Infrared Spectroscopy for Diagnosing Superlattice Minibands in Magic-angle Twisted Bilayer Graphene

Twisted bilayer graphene (TBG) represents a highly tunable, strongly correlated electron system owed to its unique flat electronic bands. However, understanding the single-particle band structure alone has been challenging due to complex lattice reconstruction effects and a lack of spectroscopic measurements over a broad energy range. Here, we probe the band structure of TBG around the magic angle using infrared spectroscopy. Our measurements reveal spectral features originating from interband transitions whose energies are uniquely defined by the twist angle. By combining with quantum transport, we connect spectral features over a broad energy range (10 to 700 meV) spanning several superlattice minibands and track their evolution with twist angle. We compare our data with calculations of the band structures obtained via the continuum model and find good agreement only when considering a variation of interlayer/intralayer tunnelling parameters with the twist angle. Our analysis suggests that the magic angle also shifts due to lattice relaxation, and is better defined for a wide angular range from 0.9° to 1.1°. Our work provides spectroscopic insights into TBG's band structure and offers an optical fingerprint of the magic angle for screening heterostructures before nanofabrication.

cond-mat.mes-hall↗

arXiv2023

FCNC charmed-hadron decays with invisible singlet particles in light of recent data

The flavor-changing neutral current (FCNC) decays of charmed hadrons with missing energy $(\not\!\!E)$ can serve as potentially promising hunting grounds for hints of new physics, as the standard-model backgrounds are very suppressed. A few of such processes have been searched for in recent experiments, specifically $D^0\to\,\not\!\!E$ by Belle and $D^0\toπ^0$$\not\!\!E$ and $Λ_c^+\to p\!\not\!\!E$ by BESIII, resulting in upper bounds on their branching fractions. We consider them to illuminate the possible contributions of the quark transition $c\to u\!\not\!\!E$ with a couple of invisible spinless bosons carrying away the missing energy, assuming that they are not charge conjugates of each other and hence can have unequal masses. We find that these data are complementary in that they constrain different sets of the underlying operators and do not cover the same ranges of the bosons' masses, but there are regions not yet accessible. From the allowed parameter space, we show that other $D$-meson decays, such as $D\toρ$$\not\!\!E$, and the charmed-baryon ones $Ξ_c\to(Σ,Λ)$$\not\!\!E$ can have sizable branching fractions and therefore may offer further probes of the new-physics interactions. We point out the importance of $D^0\toγ\!\not\!\!E$ which are not yet searched for but could access parts of the parameter space beyond the reach of the other modes. In addition, we look at a scenario where the invisibles are instead fermionic, namely sterile neutrinos, and a scalar leptoquark mediates $c\to u\!\not\!\!E$. We discuss the implications of the aforesaid bounds for this model. The predictions we make for the various charmed-hadron decays in the different scenarios may be testable in the near future by BESIII and Belle II.

hep-ph↗

arXiv2023

Qubit Reset with a Shortcut-to-Isothermal Scheme

Landauer's principle shows that the minimum energy cost to reset a classical bit in a bath with temperature $T$ is $k_{B}T\ln2$ in the infinite time. However, the task to reset the bit in finite time has posted a new challenge, especially for quantum bit (qubit) where both the operation time and controllability are limited. We design a shortcut-to-isothermal scheme to reset a qubit in finite time $τ$ with limited controllability. The energy cost is minimized with the optimal control scheme with and without nonholonomic constraint. This optimal control scheme can provide a reference to realize qubit reset with minimum energy cost for the limited time.

quant-ph↗

arXiv2023

IWEK: An Interpretable What-If Estimator for Database Knobs

The knobs of modern database management systems have significant impact on the performance of the systems. With the development of cloud databases, an estimation service for knobs is urgently needed to improve the performance of database. Unfortunately, few attentions have been paid to estimate the performance of certain knob configurations. To fill this gap, we propose IWEK, an interpretable & transferable what-if estimator for database knobs. To achieve interpretable estimation, we propose linear estimator based on the random forest for database knobs for the explicit and trustable evaluation results. Due to its interpretability, our estimator capture the direct relationships between knob configuration and its performance, to guarantee the high availability of database. We design a two-stage transfer algorithm to leverage historical experiences to efficiently build the knob estimator for new scenarios. Due to its lightweight design, our method can largely reduce the overhead of collecting training data and could achieve cold start knob estimation for new scenarios. Extensive experiments on YCSB and TPCC show that our method performs well in interpretable and transferable knob estimation with limited training data. Further, our method could achieve efficient estimator transfer with only 10 samples in TPCC and YSCB.

cs.DB↗

arXiv2023

Theoretical proposal to obtain strong Majorana evidence from scanning tunneling spectroscopy of a vortex with a dissipative environment

It is predicted that a vortex in a topological superconductor contains a Majorana zero mode (MZM). The confirmative Majorana signature, i.e., the $2e^2/h$ quantized conductance, however is easily sabotaged by unavoidable interruptions, e.g. instrument broadening, non-Majorana signal, and extra particle channels. We propose to avoid the signal interruption by introducing disorder-induced dissipation that couples to the tip-sample tunneling. With dissipation involved, we highlight three features, each of which alone can provide a strong evidence to identify MZM. Firstly, dissipation suppresses a finite-energy Caroli-de Gennes-Matricon (CdGM) conductance peak into a valley, while it does not split MZM zero-bias conductance peak. Secondly, we predict a dissipation-dependent scaling feature of the zero-bias conductance peak. Thirdly, the introduced dissipation manifests the MZM signal by suppressing non-topological CdGM modes. Importantly, the observation of these features does not require a quantized conductance value $2e^2/h$.

cond-mat.supr-con↗

arXiv2023

A Shortcut to Finite-time Memory Erasure

To achieve fast computation, it is crucial to reset the memory to a desired state within a limited time. However, the inherent delay in the system's response often prevents reaching the desired state once the control process is completed in finite time. To address this challenge, we propose a shortcut strategy that incorporates an auxiliary control to guide the system towards an equilibrium state that corresponds to the intended control, thus enabling accurate memory reset. Through the application of thermodynamic geometry, we derive an optimal shortcut protocol for erasure processes that minimizes the energy cost. This research provides an effective design principle for realizing the finite-time erasure process while simultaneously reducing the energy cost, thereby alleviating the burden of heat dissipation.

cond-mat.stat-mech↗

arXiv2023

FCNC processes of charmed hadrons with invisible scalar

We study the invisible scalar of $S$ in charmed meson and baryon decays with flavor-changing neutral currents (FCNCs) based on the model-independent effective Lagrangian between the quarks and invisible scalar. From the bounds of the coupling constants extracted from the recent BES III experiment on the decay of $D^{0}\toπ^{0}\barνν$, we predict that the decay branching ratios of $D^{+}\toπ^{+}S$, $D_s^+\to K^+S$, $Λ_c\to pS$, $Ξ_c^{0(+)}\to Σ^{0(+)}S$, and $Ξ_c^0\toΛS$ can be as large as $10.6,~2.53,~2.39,~0.963,~5.77$, and $2.92\times10^{-4}$ with $m_{_S}\approx 1.1~{\rm GeV}$, respectively. Some of these decay modes are accessible to the ongoing experiments, such as BES III, LHCb and Belle II, as well as the future ones, such as FCC-ee.

hep-ph↗

arXiv2022

Bottomed baryon decays with invisible Majorana fermions

We study the invisible Majorana fermions of $χ$ in bottomed baryon decays with flavor-changing neutral currents (FCNCs) based on the model-independent effective Lagrangian between the quarks and invisible particles. From the bounds of the coupling constants extracted from the experiments, we examine the decay branching ratios of $Λ_b\toΛχχ$, $Ξ_b^{0(-)}\toΞ^{0(-)}χχ$, $Λ_b\to n χχ$, $Ξ_b^{-}\to Σ^{-} χχ$, $Ξ_b^{0}\to Σ^{0} χχ$, and $Ξ_b^{0}\to Λχχ$, which can be as large as $6.3,~9.2,~5.7,~5.8,~2.7$, and $1.0\times10^{-5}$ for $m_χ=2$ GeV, respectively. Some of these decays are accessible to the future experimental searches, such as Belle II.

hep-ph↗