arXiv subjects

Jin Dai

Publications and source records attributed to Jin Dai.

At least 19 recordsLinked to original sources

arXiv2026

Angular Gaussian Supervised Contrastive Learning for Long-Tailed Electrocardiogram Arrhythmia Diagnosis

Long-tailed label distributions reduce the reliability of deep learning for electrocardiogram (ECG) arrhythmia diagnosis, particularly for clinically important but rare abnormalities. Existing rebalancing and logit adjustment methods mainly address class frequency while overlooking direction-dependent morphological variability across ECG classes. This study proposes Angular Gaussian Supervised Contrastive Learning (AG-SCL) for long-tailed multi-label ECG diagnosis. AG-SCL integrates three components into a unified framework: an Angular Gaussian contrastive branch that models full-covariance class uncertainty on unit-normalized embeddings, Adaptive Logit Adjustment that learns bounded label-state-specific prior corrections instead of fixed frequency-based margins, and tail-aware augmentation that generates morphology-preserving views while protecting the 7-25 Hz QRS-dominant band. The method was evaluated on the public PTB-XL benchmark and a nocturnal ECG dataset comprising 1317 hours of recordings from 141 subjects. AG-SCL achieved the best macro-level performance on both datasets. On PTB-XL, it obtained a balanced accuracy of 0.838, sensitivity of 0.709, specificity of 0.968, mean average precision of 0.495, and TPR at 5% FPR of 0.778. On Noc-ECG, the corresponding values were 0.918, 0.889, 0.947, 0.488, and 0.900. The largest gains occurred in rare or morphologically unstable rhythm classes, while ablation studies confirmed the contributions of full-covariance modelling, Adaptive Logit Adjustment, and tail-aware augmentation. AG-SCL improves long-tailed ECG diagnosis by combining prior calibration with anisotropic representation learning, enhancing sensitivity to rare arrhythmias while maintaining clinically relevant specificity. Our code is available at: https://github.com/Open-EXG/AG-SCL-for-Long-Tailed-ECG.

cs.LG↗

arXiv2026

A positive solution to the $L^p$ projection centroid conjecture

In a classical paper [20] in 2000, Lutwak-Yang-Zhang established the $L^p$ analog of the Petty projection inequality and the $L^p$ analog of the Busemann-Petty centroid inequality. In Section 7 of [20], Lutwak-Yang-Zhang proposed the important $L^p$ projection centroid conjecture. We give a positive solution to the $L^p$ projection centroid conjecture in this work.

math.MG↗

arXiv2026

Strain-Dependent Ionic Transport in Li3YCl6 Solid Electrolytes

Solid-state batteries require electrolytes that sustain high ionic conductivity under the mechanical environment of a functioning cell. Lattice strain, arising from stack pressure, thermal cycling, or lattice mismatch at interfaces, can either enhance or suppress Li+ transport in solid electrolytes, yet how it couples to the underlying diffusion mechanism remains poorly understood. Using Li3YCl6 halide superionic conductor, we address this with large-scale molecular dynamics simulations driven by an Atomic Cluster Expansion (ACE) machine learning interatomic potential trained on first-principles data. The ACE model faithfully reproduces experimental and \textit{ab initio} structural, mechanical, and transport properties of Li3YCl6. We find that Li+ diffusion in Li3YCl6 follows a two-regime Arrhenius behavior, crossing over at a critical temperature $T_c$ from one-dimensional hopping at low temperature to three-dimensional cooperative diffusion at high temperature. Strain substantially modulates diffusivity: tensile strain enhances it while compressive strain suppresses it, yet leaves $T_c$ invariant, indicating that strain tunes diffusion efficiency without reshaping the underlying transport framework. In each regime, the mechanistic origin differs: altered activation barriers dominate at low temperature, while modified pre-exponential factors become critical at high temperature. These results establish lattice strain as a design lever for ionic conductivity in Li3YCl6 solid-state electrolytes.

cond-mat.mtrl-sci↗

arXiv2026

GeMi: A Graph-based, Multimodal Recommendation System for Narrative Scroll Paintings

Recommendation Systems are effective in managing the ever-increasing amount of multimodal data available today and help users discover interesting new items. These systems can handle various media types such as images, text, audio, and video data, and this has made it possible to handle content-based recommendation utilizing features extracted from items while also incorporating user preferences. Graph Neural Network (GNN)-based recommendation systems are a special class of recommendation systems that can handle relationships between items and users, making them particularly attractive for content-based recommendations. Their popularity also stems from the fact that they use advanced machine learning techniques, such as deep learning on graph-structured data, to exploit user-to-item interactions. The nodes in the graph can access higher-order neighbor information along with state-of-the-art vision-language models for processing multimodal content, and there are well-designed algorithms for embedding, message passing, and propagation. In this work, we present the design of a GNN-based recommendation system on a novel data set collected from field research. Designed for an endangered performing art form, the recommendation system uses multimodal content (text and image data) to suggest similar paintings for viewing and purchase. To the best of our knowledge, there is no recommendation system designed for narrative scroll paintings -- our work therefore serves several purposes, including art conservation, a data storage system for endangered art objects, and a state-of-the-art recommendation system that leverages both the novel characteristics of the data and preferences of the user population interested in narrative scroll paintings.

cs.LG↗

arXiv2024

Cartesian atomic moment machine learning interatomic potentials

Machine learning interatomic potentials (MLIPs) have substantially advanced atomistic simulations in materials science and chemistry by balancing accuracy and computational efficiency. While leading MLIPs rely on representing atomic environments using spherical tensors, Cartesian representations offer potential advantages in simplicity and efficiency. Here, we introduce the Cartesian Atomic Moment Potential (CAMP), an approach to building MLIPs entirely in Cartesian space. CAMP constructs atomic moment tensors from neighboring atoms and employs tensor products to incorporate higher body-order interactions, providing a complete description of local atomic environments. Integrated into a graph neural network (GNN) framework, CAMP enables physically motivated, systematically improvable potentials. The model demonstrates excellent performance across diverse systems, including periodic structures, small organic molecules, and two-dimensional materials, achieving accuracy, efficiency, and stability in molecular dynamics simulations that rival or surpass current leading models. CAMP provides a powerful tool for atomistic simulations to accelerate materials understanding and discovery.

cond-mat.mtrl-sci↗

arXiv2024

Uncertainty Quantification and Propagation in Atomistic Machine Learning

Machine learning (ML) offers promising new approaches to tackle complex problems and has been increasingly adopted in chemical and materials sciences. Broadly speaking, ML models employ generic mathematical functions and attempt to learn essential physics and chemistry from a large amount of data. Consequently, because of the limited physical or chemical principles in the functional form, the reliability of the predictions is oftentimes not guaranteed, particularly for data far out of distribution. It is critical to quantify the uncertainty in model predictions and understand how the uncertainty propagates to downstream chemical and materials applications. Herein, we review existing uncertainty quantification (UQ) and uncertainty propagation (UP) methods for atomistic ML under a united framework of probabilistic modeling. We first categorize the UQ methods, with the aim to elucidate the similarities and differences between them. We also discuss performance metrics to evaluate the accuracy, precision, calibration, and efficiency of the UQ methods and techniques for model recalibration. With these metrics, we survey existing benchmark studies of the UQ methods using molecular and materials datasets. Furthermore, we discuss UP methods to propagate the uncertainty obtained from ML models in widely used materials and chemical simulation techniques, such as molecular dynamics and microkinetic modeling. We also provide remarks on the challenges and future opportunities of UQ and UP in atomistic ML.

cond-mat.mtrl-sci↗

arXiv2024

Path Planning and Motion Control for Accurate Positioning of Car-like Robots

This paper investigates the planning and control for accurate positioning of car-like robots. We propose a solution that integrates two modules: a motion planner, facilitated by the rapidly-exploring random tree algorithm and continuous-curvature (CC) steering technique, generates a CC trajectory as a reference; and a nonlinear model predictive controller (NMPC) regulates the robot to accurately track the reference trajectory. Based on the $μ$-tangency conditions in prior art, we derive explicit existence conditions and develop associated computation methods for a special class of CC paths which not only admit the same driving patterns as Reeds-Shepp paths but also consist of cusp-free clothoid turns. Afterwards, we create an autonomous vehicle parking scenario where the NMPC endeavors to follow the reference trajectory. Feasibility and computational efficiency of the CC steering are validated by numerical simulation. CarSim-Simulink joint simulations statistically verify that with exactly same NMPC, the closed-loop system with CC trajectories as references substantially outperforms the case where Reeds-Shepp trajectories are used as references.

cs.RO↗

arXiv2023

Ultra-low threshold pH sensor based on a whispering gallery mode microbubble resonator

Laser sensing has a wide range of applications. In this paper, we propose a pH sensing laser with an ultra-low threshold and low sample consumption based on a whispering-gallery-mode microbubble resonator. Rhodamine 6G aqueous solutions with different pH values are used as the lasing gain media, which are injected through the microfluidic channel and interact with the high-quality-factor microbubble resonator to achieve lasing. Subtle pH changes of the aqueous solution lead to changes in lasing intensity in real time and the threshold reaches a minimum of 0.091 uJ/mm2. The low pump energy density effectively avoids the self-aggregation and photobleaching effects of dye molecules present in high-concentration rhodamine 6G solutions. The lasing characteristics under different pH conditions were determined experimentally and theoretically, and the results are in good agreement. Due to the deprotonation of amino groups in highly alkaline environments, the lasing threshold is highly dependent on the pH of rhodamine 6G aqueous solutions. In the pH range of 10.16-13.14, the lasing intensity changes considerably with the increasing pH. The proposed pH-sensing laser exhibits a fast response time, low toxicity, and a high signal-to-noise ratio, making it promising for highly sensitive alkaline detection in biological applications.

physics.optics↗

arXiv2023

Graviton-Photon Conversion in Atoms and the Detection of Gravitons

We study graviton-photon conversion in potential ground-based experiments. From graviton to photon transition, we calculate the cross section of graviton-atom interaction in the presence of spherical atomic electric fields; the obtained results hold for graviton energy around 100 keV to 1 GeV, and would be enhanced along the coherent length in extremely high frequencies; thus it gives a chance to catch MeV level gravitons from the universe with current neutrino facilities. From photon to graviton transition, we propose an experiment using entangled photon pairs to count missing photons passing through transverse magnetic tunnel, which could be used to verify the energy quantization of gravitational field.

gr-qc↗

arXiv2023

From Feynman's ratchet to timecrystalline molecular motors

Cats use the connection governing parallel transport in the space of shapes to land safely on their feet. Here we argue that this connection also explains the impressive performance of molecular motors by enabling molecules to evade conclusions of Feynman's ratchet-and-pawl analysis. We first demonstrate, using simple molecular models, how directed rotational motion can emerge from shape changes even without angular momentum. We then computationally design knotted polyalanine molecules and show how their shape space connection organizes individual atom thermal vibrations into collective rotational motion, independently of angular momentum. Our simulations show that rotational motion arises effortlessly even in ambient water, making the molecule an effective theory time crystal. Our findings have potential for practical molecular motor design and engineering and can be verified through high-precision nuclear magnetic resonance measurements.

cond-mat.soft↗

arXiv2023

Research on the transition dynamics and linear (nonlinear) optical properties of mCherry

In this study, we explore the electron transition mechanism and optical properties of the popular red fluorescent protein mCherry. By examining the charge transfer spectrum and combining it with the mCherry hole-electron distribution, we identify that the charge transfer between the phenolate and imidazolinone loops significantly contributes to the absorption spectrum. Quantitative analysis of charge transfer shows that, overall, the electrons are transferred to the C16 atom in the middle of phenolate and the imidazolinone loops during absorption. We speculate that C16 may also absorb protons to enable the photoconversion of mCherry in the excited state, similar to the blinking mechanism of IrisFP. In addition, we further investigated the optical properties of mcherry in the external field by polarizability (hyperpolarizability), showing the anisotropy of the polarization, the first hyperpolarization and the second hyperpolarization by unit spherical representation. Our results suggest that significant polarization and second hyperpolarizability occur when the field direction and electron transfer direction are aligned. We also analyzed the polarizability and first hyperpolarizabilities for different external fields. The polarizability mutated when the external field satisfies the S_0,min-> S_1 transition. Finally, the study of the first hyperpolarizability shows that adjusting the appropriate field can lead to a linear photoelectric effect or second harmonic generation of mCherry. These studies have certain reference values for various red fluorescent protein correlation simulations and experiments because of the similarity of the red fluorescent protein.

physics.chem-ph↗

arXiv2022

The SU(2) Lie-Poisson Algebra and its Descendants

In this paper, a novel discrete algebra is presented which follows by combining the SU(2) Lie-Poisson bracket with the discrete Frenet equation. Physically, the construction describes a discrete piecewise linear string in R3. The starting point of our derivation is the discrete Frenet frame assigned at each vertix of the string. Then the link vector that connect the neighbouring vertices assigns the SU(2) Lie-Poisson bracket. Moreover, the same bracket defines the transfer matrices of the discrete Frenet equation which relates two neighbouring frames along the string. The procedure extends in a self-similar manner to an infinite hierarchy of Poisson structures. As an example, the first descendant of the SU(2) Lie-Poisson structure is presented in detail. For this, the spinor representation of the discrete Frenet equation is employed, as it converts the brackets into a computationally more manageable form. The final result is a nonlinear, nontrivial and novel Poisson structure that engages four neighbouring vertices.

hep-th↗

arXiv2022

GPPF: A General Perception Pre-training Framework via Sparsely Activated Multi-Task Learning

Pre-training over mixtured multi-task, multi-domain, and multi-modal data remains an open challenge in vision perception pre-training. In this paper, we propose GPPF, a General Perception Pre-training Framework, that pre-trains a task-level dynamic network, which is composed by knowledge "legos" in each layers, on labeled multi-task and multi-domain datasets. By inspecting humans' innate ability to learn in complex environment, we recognize and transfer three critical elements to deep networks: (1) simultaneous exposure to diverse cross-task and cross-domain information in each batch. (2) partitioned knowledge storage in separate lego units driven by knowledge sharing. (3) sparse activation of a subset of lego units for both pre-training and downstream tasks. Noteworthy, the joint training of disparate vision tasks is non-trivial due to their differences in input shapes, loss functions, output formats, data distributions, etc. Therefore, we innovatively develop a plug-and-play multi-task training algorithm, which supports Single Iteration Multiple Tasks (SIMT) concurrently training. SIMT lays the foundation of pre-training with large-scale multi-task multi-domain datasets and is proved essential for stable training in our GPPF experiments. Excitingly, the exhaustive experiments show that, our GPPF-R50 model achieves significant improvements of 2.5-5.8 over a strong baseline of the 8 pre-training tasks in GPPF-15M and harvests a range of SOTAs over the 22 downstream tasks with similar computation budgets. We also validate the generalization ability of GPPF to SOTA vision transformers with consistent improvements. These solid experimental results fully prove the effective knowledge learning, storing, sharing, and transfer provided by our novel GPPF framework.

cs.CV↗

arXiv2021

Design of a HOM-Damped 166.6 MHz Compact Quarter-Wave beta=1 Superconducting Cavity for High Energy Photon Source

Superconducting cavities with low RF frequencies and heavy damping of higher order modes (HOM) are desired for the main accelerator of High Energy Photon Source (HEPS), a 6 GeV synchrotron light source promising ultralow emittance currently under construction in Beijing. A compact 166.6 MHz superconducting cavity was proposed adopting a quarter-wave beta=1 geometry. Based on the successful development of a proof-of-principle cavity, a HOM-damped 166.6 MHz compact superconducting cavity was subsequently designed. A ferrite damper was installed on the beam pipe to reduce HOM impedance below the stringent threshold of coupled-bunch instabilities. Being compact, RF field heating on the cavity vacuum seal was carefully examined against quenching the NbTi flange. The cavity was later dressed with a helium vessel and the tuning mechanism was also realized. Excellent RF and mechanical properties were eventually achieved. Finally, the two-cavity string was designed to ensure smooth transitions among components and proper shielding of synchrotron light. This paper presents a complete design of a fully dressed HOM-damped low-frequency beta=1 superconducting cavity for HEPS.

physics.acc-ph↗

arXiv2021

Vortex precession and exchange in a Bose-Einstein condensate

Vortices in a Bose-Einstein condensate are modelled as spontaneously symmetry breaking minimum energy solutions of the time dependent Gross-Pitaevskii equation, using the method of constrained optimization. In a non-rotating axially symmetric trap, the core of a single vortex precesses around the trap center and, at the same time, the phase of its wave function shifts at a constant rate. The precession velocity, the speed of phase shift, and the distance between the vortex core and the trap center, depend continuously on the value of the conserved angular momentum that is carried by the entire condensate. In the case of a symmetric pair of identical vortices, the precession engages an emergent gauge field in their relative coordinate, with a flux that is equal to the ratio between the precession and shift velocities.

cond-mat.quant-gas↗

arXiv2020

Rotation by shape change, autonomous molecular motors and effective timecrystalline dynamics

A deformable body can rotate even with no angular momentum, simply by changing its shape. A good example is a falling cat, how it maneuvers in air to land on its feet. Here a first principles molecular level example of the phenomenon is presented. For this the thermal vibrations of individual atoms in an isolated cyclopropane molecule are simulated in vacuum and at ultralow internal temperature values, and the ensuing molecular motion is followed stroboscopically. It is observed that in the limit of long stroboscopic time steps the vibrations combine into an apparent uniform rotation of the entire molecule even in the absence of angular momentum. This large time scale rotational motion is then modeled in an effective theory approach, in terms of timecrystalline Hamiltonian dynamics. The phenomenon is a temperature sensitive measurable. As such it has potential applications that range from models of autonomous molecular motors to development of molecular level detector, sensor and control technologies.

physics.comp-ph↗

arXiv2020

Classical Hamiltonian Time Crystals -- General Theory And Simple Examples

We focus on a Hamiltonian system with a continuous symmetry, and dynamics that takes place on a presymplectic manifold. We explain how the symmetry can become spontaneously broken by a time crystal, that we define as the minimum of the available mechanical free energy that is simultaneously a time dependent solution of Hamilton's equation. The mathematical description of such a timecrystalline spontaneous symmetry breaking builds on concepts of equivariant Morse theory in the space of Hamiltonian flows. As an example we analyze a general family of timecrystalline Hamiltonians that is designed to model polygonal, piecewise linear closed strings. The vertices correspond to the locations of pointlike interaction centers; the string is akin a chain of atoms, that are joined together by covalent bonds, modeled by the links of the string. We argue that the timecrystalline character of the string can be affected by its topology. For this we show that a knotty string is usually more timecrystalline than a string with no self-entanglement. We also reveal a relation between phase space topology and the occurrence of timecrystalline dynamics. For this we show that in the case of three point particles, the presence of a time crystal can relate to a Dirac monopole that resides in the phase space. Our results propose that physical examples of Hamiltonian time crystals can be realized in terms of closed, knotted molecular rings.

nlin.PS↗

arXiv2019

Topologically enhanced time crystals and molecular knots

A time crystal is a time dependent physical system that does not reach a standstill, even in state of minimum energy. Here we show that the stability of a time crystal can be enhanced by its topology. For this we simulate time crystals made of chainlike ensembles of mutually interacting point particles. When we tie the chain into a knot we find that its timecrystalline qualities improve. The theoretical models we consider are widely used in coarse grained descriptions of linear polymers. Thus we expect that physical realizations of time crystals can be found in terms of knotted molecules.

nlin.PS↗