Searcharxiv⌕ Search

arXiv subjects

Jinsong Liu

Publications and source records attributed to Jinsong Liu.

At least 19 recordsLinked to original sources
arXiv2026

Harmonic Maps from Punctured Riemann Surfaces to the Hyperbolic Plane with Prescribed Scherk Asymptotics

Let $X$ be a genus $g\geq 0$ compact Riemann surface that admits an antiholomorphic involution $ι$ with non empty fixed-point set, and let $D = \{p_1,\dots,p_k\}\subset \text{Fix}(ι)$. At each puncture, we prescribe a Scherk map associated to a given real polynomial quadratic differential with even degree and negative leading coefficient. Here a Scherk map is the harmonic diffeomorphism from $\mathbb{C}$ to the interior of an ideal polygon in $\mathbb{H}^2$, whose Hopf differential is the given polynomial. We construct a harmonic map \[ h:X\backslash D \to \mathbb{H}^2 \] whose asymptotic behavior at each puncture matches the prescribed Scherk map. In particular, the image of $h$ tends to an ideal polygon near each end. The resulting $h$ covers harmonic maps obtained by taking the $\mathbb{H}^2$ factors of horizontal catenoids in $\mathbb{H}^2 \times \mathbb{R}$.

math.DG↗
arXiv2026

Rethinking the Design of Reinforcement Learning-Based Deep Research Agents

Large language models (LLMs) augmented with external tools are increasingly deployed as deep research agents that gather, reason over, and synthesize web information to answer complex queries. Although recent open-source systems achieve strong empirical performance via reinforcement learning from web interactions, the impact of key design choices remains under-explored. We formalize deep research as reinforcement learning in an episodic finite Markov decision process and construct a competitive baseline agent grounded in this formulation. Building on this foundation, we systematically examine critical design decisions at both training and inference time and identify four factors that substantially improve performance: replacing rule-based rewards with AI feedback from an LLM judge, fine-tuning with the on-policy RLOO algorithm instead of the off-policy GRPO algorithm, filtering low-quality training samples, and employing an error-tolerant test-time rollout strategy. Together, these design choices yield a deep research agent that establishes state-of-the-art performance among 7B-scale agents when evaluated across ten widely used benchmarks.

cs.AI↗
arXiv2026

Closing Reasoning Gaps in Clinical Agents with Differential Reasoning Learning

Clinical decision support requires not only correct answers but also clinically valid reasoning. We propose Differential Reasoning Learning (DRL), a framework that improves clinical agents by learning from reasoning discrepancies. From reference reasoning rationales (e.g., physician-authored clinical rationale, clinical guidelines, or outputs from more capable models) and the agent's free-form chain-of-thought (CoT), DRL extracts reasoning graphs as directed acyclic graphs (DAGs) and performs a clinically weighted graph edit distance (GED)-based discrepancy analysis. An LLM-as-a-judge aligns semantically equivalent nodes and diagnoses discrepancies between graphs. These graph-level discrepancy diagnostics are converted into natural-language instructions and stored in a Differential Reasoning Knowledge Base (DR-KB). At inference, we retrieve top-$k$ instructions via Retrieval-Augmented Generation (RAG) to augment the agent prompt and patch likely logic gaps. Evaluation on open medical question answering (QA) benchmarks and a Return Visit Admissions (RVA) prediction task from internal clinical data demonstrates gains over baselines, improving both final-answer accuracy and reasoning fidelity. Ablation studies confirm gains from infusing reference reasoning rationales and the top-$k$ retrieval strategy. Clinicians' review of the output provides further assurance of the approach. Together, results suggest that DRL supports more reliable clinical decision-making in complex reasoning scenarios and offers a practical mechanism for deployment under limited token budgets.

cs.AI↗
arXiv2026

Systematic Study of Amorphous ABC Heterostructures at the Atomic Scale as a Second-Order Nonlinear Optical Metamaterial

Systematic exploration of amorphous ABC heterostructures revealed that nanoscale morphological modifications markedly improved their artificial bulk second-order susceptibility. These amorphous birefringent heterostructures were fabricated through plasma-enhanced atomic layer deposition of three oxides, effectively breaking the centrosymmetry on the nanoscale. We observe a dependence of the optical nonlinearity on the thickness variation of three constituent materials, SiO$_2$, TiO$_2$, and Al$_2$O$_3$, ranging from tens of nanometers to the atomic scale, and these thin films exhibit second-order susceptibility at their interfaces. Our findings reveal that the enhancement of nonlinear optical properties is strongly correlated with a high density of layers and superior interface quality, where the interface second-order nonlinearity transitions to bulk-like second-harmonic generation. An effective bulk second-order susceptibility of $χ_{zzz} = 2.0 \pm 0.2$~pm/V at the wavelength of 1032~nm is achieved, comparable to typical values for conventional monocrystalline nonlinear materials.

physics.app-ph↗
arXiv2025

Linear and Nonlinear Optical Properties of SiO$_2$/TiO$_2$ Heterostructures Grown by Plasma Enhanced Atomic Layer Deposition

Second harmonic (SH) radiation can only be generated in non-centrosymmetric bulk crystals under the electric-dipole approximation. Nonlinear thin films made from bulk crystals are technologically challenging because of complex and high temperature fabrication processes. In this work, heterostructures made of amorphous materials SiO$_2$ and TiO$_2$ were prepared by a CMOS-compatible technique named plasma enhanced atomic layer deposition (PEALD) with deposition temperature at 100 °C. By using the uniaxial dispersion model, we characterized the form-birefringence properties, which can enable the phase matching condition in waveguides or other nonlinear optical applications. By applying a fringe-based technique, we determined the largest diagonal component of the effective second-order bulk susceptibility $χ_{zzz}^{(2)}$ = 1.30$\pm$0.13 pm/V at a wavelength of 1032 nm. Noteworthy, we observed strong SH signals from two-component nanolaminates, which are several orders of magnitude larger than from single layers. The SH signals from our samples only require the broken inversion symmetry at the interface. Here optical properties of nanocomposites can be precisely tuned by the promising PEALD technology.

cond-mat.mtrl-sci↗
arXiv2025

Spatiotemporally Localized Optical Links and Knots

Optical links and knots have attracted growing attention owing to their exotic topologic features and promising applications in next-generation information transfer and storage. However, current protocols for optical topology realization rely on paraxial propagation of spatial modes, which inherently limits their three-dimensional topological structures to longitudinal space-filling. In this work we propose and experimentally demonstrate a scheme for creating optical knots and links that are localized in space within a transverse plane of a paraxial field, as well as in time. These spatiotemporal topological structures arise from polychromatic wave fields with tightly coupled spatial and temporal degrees of freedom that can be realized in the form of superpositions of toroidal light vortices of opposite topological charges. The (2+1)-dimensional nature of a toroidal light vortex imparts spatiotemporally localized wave fields with nontrivial topological textures, encompassing both individual and nested links or knots configurations. Moreover, the resulting topological textures are localized on an ultrashort timescale propagate at the group velocity of the wave packets and exhibit remarkable topological robustness during propagation as optical carriers. The nascent connection between spatiotemporally localized fields and topology offers exciting prospects for advancing space-time photonic topologies and exploring their potential applications in high-capacity informatics and communications.

physics.optics↗
arXiv2025

The growth rate on the volume of $\mathcal{M}_g^{<L(g)}$

Let $\mathcal{M}_g$ be the moduli space of hyperbolic surfaces of genus g endowed with the Weil-Petersson metric. In this paper, we introduce a function $L(g)$ of genus $g$ and call the geodesics whose length less than $L(g)$ short geodesics. We compute the growth rate on the volume of the subset of hyperbolic surfaces with short geodesics. In particular, when $g$ approaches infinity, if $L(g)$ also approaches infinity, then the volume of surfaces characterized by short geodesics is equal to $V_g$ almost surely.

math.GT↗
arXiv2025

Deep Reinforcement Learning Xiangqi Player with Monte Carlo Tree Search

This paper presents a Deep Reinforcement Learning (DRL) system for Xiangqi (Chinese Chess) that integrates neural networks with Monte Carlo Tree Search (MCTS) to enable strategic self-play and self-improvement. Addressing the underexplored complexity of Xiangqi, including its unique board layout, piece movement constraints, and victory conditions, our approach combines policy-value networks with MCTS to simulate move consequences and refine decision-making. By overcoming challenges such as Xiangqi's high branching factor and asymmetrical piece dynamics, our work advances AI capabilities in culturally significant strategy games while providing insights for adapting DRL-MCTS frameworks to domain-specific rule systems.

cs.AI↗
arXiv2025

Propagation Dynamics of Photonic Toroidal Vortices Mediated by Orbital Angular Momenta

The dynamics of vortex rings in fluids have long captivated researchers due to the intriguing complexity of their behavior, despite the apparent simplicity of their structure. In optics, photonic toroidal vortices constitute a novel class of three-dimensional, space-time nonseparable structured light fields that carry transverse orbital angular momentum. However, as solutions to the dispersive form of Maxwell's equations, these wavepackets do not survive upon nondispersive propagation, and their dynamics remain elusive. In this article, the dynamics of photonic toroidal vortices under various dispersion regimes, mediated by both transverse and longitudinal orbital angular momentum, are investigated through simulations and experiments. The results reveal that the motion of a toroidal vortex is strongly affected by the presence of longitudinal orbital angular momentum. The swirling flow destabilizes the toroidal structure under dispersion conditions and induces topological transformations in the vortex line characterized by its annihilation and subsequent reformation in vacuum. Remarkably, the renascent toroidal vortex exhibits robust propagation in vacuum while maintaining its toroidal structure. These findings are supported by experimental validation and highlight the potential of photonic toroidal vortices as controllable channels for directional energy and information transfer.

physics.optics↗
arXiv2025

A Deep Learning Framework Integrating CNN and BiLSTM for Financial Systemic Risk Analysis and Prediction

This study proposes a deep learning model based on the combination of convolutional neural network (CNN) and bidirectional long short-term memory network (BiLSTM) for discriminant analysis of financial systemic risk. The model first uses CNN to extract local patterns of multidimensional features of financial markets, and then models the bidirectional dependency of time series through BiLSTM, to comprehensively characterize the changing laws of systemic risk in spatial features and temporal dynamics. The experiment is based on real financial data sets. The results show that the model is significantly superior to traditional single models (such as BiLSTM, CNN, Transformer, and TCN) in terms of accuracy, recall, and F1 score. The F1-score reaches 0.88, showing extremely high discriminant ability. This shows that the joint strategy of combining CNN and BiLSTM can not only fully capture the complex patterns of market data but also effectively deal with the long-term dependency problem in time series data. In addition, this study also explores the robustness of the model in dealing with data noise and processing high-dimensional data, providing strong support for intelligent financial risk management. In the future, the research will further optimize the model structure, introduce methods such as reinforcement learning and multimodal data analysis, and improve the efficiency and generalization ability of the model to cope with a more complex financial environment.

cs.LG↗
arXiv2024

Leveraging Convolutional Neural Network-Transformer Synergy for Predictive Modeling in Risk-Based Applications

With the development of the financial industry, credit default prediction, as an important task in financial risk management, has received increasing attention. Traditional credit default prediction methods mostly rely on machine learning models, such as decision trees and random forests, but these methods have certain limitations in processing complex data and capturing potential risk patterns. To this end, this paper proposes a deep learning model based on the combination of convolutional neural networks (CNN) and Transformer for credit user default prediction. The model combines the advantages of CNN in local feature extraction with the ability of Transformer in global dependency modeling, effectively improving the accuracy and robustness of credit default prediction. Through experiments on public credit default datasets, the results show that the CNN+Transformer model outperforms traditional machine learning models, such as random forests and XGBoost, in multiple evaluation indicators such as accuracy, AUC, and KS value, demonstrating its powerful ability in complex financial data modeling. Further experimental analysis shows that appropriate optimizer selection and learning rate adjustment play a vital role in improving model performance. In addition, the ablation experiment of the model verifies the advantages of the combination of CNN and Transformer and proves the complementarity of the two in credit default prediction. This study provides a new idea for credit default prediction and provides strong support for risk assessment and intelligent decision-making in the financial field. Future research can further improve the prediction effect and generalization ability by introducing more unstructured data and improving the model architecture.

q-fin.RM↗
arXiv2024

Reward Learning From Preference With Ties

Reward learning plays a pivotal role in Reinforcement Learning from Human Feedback (RLHF), ensuring the alignment of language models. The Bradley-Terry (BT) model stands as the prevalent choice for capturing human preferences from datasets containing pairs of chosen and rejected responses. In preference modeling, the focus is not on absolute values but rather on the reward difference between chosen and rejected responses, referred to as preference strength. Thus, precise evaluation of preference strength holds paramount importance in preference modeling. However, an easily overlooked factor significantly affecting preference strength measurement is that human attitudes towards two responses may not solely indicate a preference for one over the other and ties are also a common occurrence. To address this, we propose the adoption of the generalized Bradley-Terry model -- the Bradley-Terry model with ties (BTT) -- to accommodate tied preferences, thus leveraging additional information. We prove that even with the access to the true distributions of prompt and response, disregarding ties can lead to a notable bias in preference strength measurement. Comprehensive experiments further validate the advantages of incorporating ties in preference modeling. Notably, fine-tuning with BTT significantly outperforms fine-tuning with BT on synthetic preference datasets with ties, labeled by state-of-the-art open-source LLMs.

cs.LG↗
arXiv2024

Rapid hyperspectral photothermal mid-infrared spectroscopic imaging from sparse data for gynecologic cancer tissue subtyping

Ovarian cancer detection has traditionally relied on a multi-step process that includes biopsy, tissue staining, and morphological analysis by experienced pathologists. While widely practiced, this conventional approach suffers from several drawbacks: it is qualitative, time-intensive, and heavily dependent on the quality of staining. Mid-infrared (MIR) hyperspectral photothermal imaging is a label-free, biochemically quantitative technology that, when combined with machine learning algorithms, can eliminate the need for staining and provide quantitative results comparable to traditional histology. However, this technology is slow. This work presents a novel approach to MIR photothermal imaging that enhances its speed by an order of magnitude. Our method significantly accelerates data collection by capturing a combination of high-resolution and interleaved, lower-resolution infrared band images and applying computational techniques for data interpolation. We effectively minimize data collection requirements by leveraging sparse data acquisition and employing curvelet-based reconstruction algorithms. This method enables the reconstruction of high-quality, high-resolution images from undersampled datasets and achieving a 10X improvement in data acquisition time. We assessed the performance of our sparse imaging methodology using a variety of quantitative metrics, including mean squared error (MSE), structural similarity index (SSIM), and tissue subtype classification accuracies, employing both random forest and convolutional neural network (CNN) models, accompanied by ROC curves. Our statistically robust analysis, based on data from 100 ovarian cancer patient samples and over 65 million data points, demonstrates the method's capability to produce superior image quality and accurately distinguish between different gynecological tissue types with segmentation accuracy exceeding 95%.

cs.CV↗
arXiv2024

cuPDLP-C: A Strengthened Implementation of cuPDLP for Linear Programming by C language

A recent GPU implementation of the Restarted Primal-Dual Hybrid Gradient Method for Linear Programming was proposed in Lu and Yang (2023). Its computational results demonstrate the significant computational advantages of the GPU-based first-order algorithm on certain large-scale problems. The average performance also achieves a level close to commercial solvers for the first time in history. However, due to limitations in experimental hardware and the disadvantage of implementing the algorithm in Julia compared to C language, neither the commercial solver nor cuPDLP reached their maximum efficiency. Therefore, in this report, we have re-implemented and optimized cuPDLP in C language. Utilizing state-of-the-art CPU and GPU hardware, we extensively compare cuPDLP with the best commercial solvers. The experiments further highlight its substantial computational advantages and potential for solving large-scale linear programming problems. We also discuss the profound impact this breakthrough may have on mathematical programming research and the entire operations research community.

math.OC↗
arXiv2023

Six-degrees-of-freedom test mass readout via optical phase-locking heterodyne interferometry

Accurate position and posture measurements of the freely-falling test mass are crucial for the success of spaceborne gravitational wave detection missions. This paper presents a novel laboratory-developed test mass motion readout that utilizes quadrant photodetectors to measure the translation and tilt of a test mass. Departing from conventional methods like Zeeman effect or AOM frequency shift modulation, the readout system employs the phase locking of two lasers to generate the dual-frequency heterodyne source. Notably, the out-of-loop sensitivity of the phase locking reaches below 30 pm/Hz1/2 within the frequency band of 1 mHz and 10 Hz. The system comprises three measurement interferometers and one reference interferometer, featuring a symmetric design that enables measurements of up to six degrees of freedom based on polarization-multiplexing and differential wavefront sensing. Ground-simulated experimental results demonstrate that the proposed system has achieved a measurement sensitivity of 4 pm/Hz1/2 and 2 nrad/Hz1/2 at 1 Hz, a resolution of 5 nm and 0.1 urad, a range of 200 um and 600 urad, respectively. These findings showcase the system's potential as an alternative method for precisely monitoring the motion of test masses in spaceborne gravitational wave detection missions and other applications requiring accurate positioning and multi-degrees-of-freedom sensing.

astro-ph.IM↗
arXiv2023

$L^p$ norm of truncated Riesz transform and an improved dimension-free $L^p$ estimate for maximal Riesz transform

In this paper, we prove that the $L^p(\mathbb{R}^d)$ norm of the maximal truncated Riesz transform in terms of the $L^p(\mathbb{R}^d)$ norm of Riesz transform is dimension-free for any $2\leq p<\infty$, using integration by parts formula for radial Fourier multipliers. Moreover, we show that $$\|R_j^*f\|_{L^p}\leq \left({2+\frac{1}{\sqrt{2}}}\right)^{\frac{2}{p}}\|R_jf\|_{L^p},\ \ \mbox{for}\ \ p\geq2,\ \ d\geq2.$$ As by products of our calculations, we infer the $L^p$ norm contractivity of the truncated Riesz transforms $R^t_j$ in terms of $R_j$, and their accurate $L^p$ norms. More precisely, we prove: $$\|R^t_jf\|_{L^p}\leq\|R_jf\|_{L^p}$$ and $$\|R^t_j\|_{L^p}=\|R_j\|_{L^p},$$ for all $1 0.$

math.CA↗
arXiv2023

The spectrality of Cantor-Moran measure and Fuglede's Conjecture

Let $\{(p_n, \mathcal{D}_n, L_n)\}$ be a sequence of Hadamard triples on $\mathbb{R}$. Suppose that the associated Cantor-Moran measure $$ μ_{\{p_n,\mathcal{D}_n\}}=δ_{p_1^{-1}\mathcal{D}_1}\astδ_{(p_2p_1)^{-1}\mathcal{D}_2}\ast\cdots, $$ where $\sup_n\{|p_n^{-1}d|:d\in \mathcal{D}_n\}<\infty$ and $\sup\#\mathcal{D}_n<\infty$. It has been observed that the spectrality of $μ_{\{p_n,D_n\}}$ is determined by equi-positivity. A significant problem is what kind of Moran measures can satisfy this property. In this paper, we introduce the conception of \textit{Double Points Condition Set} (\textit{DPCS}) to characterize the equi-positivity equivalently. As applications of our characterization, we show that all singularly continuous Cantor-Moran measures are spectral. For the absolutely continuous case, we study Fuglede's Conjecture on Cantor-Moran set. We show that the equi-positivity of $μ_{\{p_n,D_n\}}$ implies the tiling of its support, and the reverse direction holds under certain conditions.

math.FA↗