arXiv subjects

Ce Li

Publications and source records attributed to Ce Li.

At least 19 recordsLinked to original sources

arXiv2026

A 0.35 mm Silicon Carbide Diffractive Waveguide with Dual Parameter Apodization for Full Color Augmented Reality

Augmented reality eyewear offers a transformative interface poised to reshape human information interaction. In this context, silicon carbide (SiC) offers unique advantages for diffractive waveguides with its high refractive index and excellent thermal conductivity. However, in single-layer full-color displays, existing SiC waveguides generally remain thicker than 0.5 mm to reduce the bounce count of total internal reflection, thereby avoiding severe spatial variations in luminance and color. Here, we demonstrate a 0.35 mm SiC diffractive waveguide with an ultra-lightweight of only 1.98 g. The challenge of spatial non-uniformity is addressed by dual-parameter apodized gratings with continuously varying depth and duty cycle, enabling fine spatial control over local diffraction efficiency. To realize high-throughput production, a parallel gradient transfer method compatible with nanoimprint lithography is introduced, enabling wafer-scale patterning of four lens pairs per 8-inch SiC wafer. Furthermore, magnesium fluoride planarization suppresses grating visibility and achieves a high see-through transmittance of 92%. Optical simulations confirm that this architecture achieves balanced full-color transmission across a 30-degree field of view. This strategy provides a scalable route toward ultra-light and visually unobtrusive glasses, opening new opportunities for practical consumer-grade wearable displays.

physics.optics↗

arXiv2026

Self-Reflective Multi-modal Reasoning for Short-Video Fake News Detection

Recent fake news detection pipelines increasingly leverage large language models and vision-language models for reasoning-based analysis. However, several challenges remain open: improving reasoning quality through self-reflection without ground-truth chain-of-thought supervision, using improved reasoning to benefit downstream model fine-tuning, and connecting single-sample fraudulent-pattern discovery with cross-sample verification. We propose SRM-FND, a self-reflective multimodal reasoning framework for short-video fake news detection. SRM-FND develops higher-quality reasoning through contrastive deliberation, iterative root-cause diagnosis, and corrective prompt refinement. A Blind Analyst, Counter-Conclusion Reasoner, and Self-Consistency Arbiter collaboratively identify and retain discriminative rationales. The framework also incorporates dual-phase, topic-adaptive vision-language model fine-tuning to improve multimodal grounding and enable lightweight topic specialization. For uncertain cases, it performs confidence-driven cross-sample review by retrieving credible and suspicious co-event examples. Experiments on FakeSV and FakeTT show that SRM-FND outperforms strong baselines, produces more reliable and interpretable predictions, and delivers noticeable improvements in cross-dataset performance.

cs.MM↗

arXiv2026

Possible Bound States in the $D^\ast\bar D^\ast$/$B^\ast\bar B^\ast$ and $D^\ast D^\ast$/$\bar B^\ast\bar B^\ast$ Systems within the Bethe-Salpeter Formalism

We investigate possible $S$-wave bound states in the $D^\ast\bar D^\ast$, $B^\ast\bar B^\ast$, $D^\ast D^\ast$, and $\bar B^\ast\bar B^\ast$ systems within the Bethe-Salpeter formalism using one-boson-exchange interactions. Bound state solutions are obtained in the isoscalar hidden-heavy systems with $J^{PC}=0^{++}$, $1^{+-}$, and $2^{++}$, whereas no isovector solutions are found within the parameter range considered. For the doubly heavy systems, solutions are obtained in the allowed $I(J^P)=0(1^+)$, $1(0^+)$, and $1(2^+)$ systems, although the $1(0^+)$ solution requires a comparatively large cutoff parameter. The bottom systems are bounded more favorably than their charmed counterparts because of their larger reduced masses.

hep-ph↗

arXiv2026

Monolithic Multifocal Diamond Metalens for High-Power Laser Systems

High-power laser systems increasingly rely on multi-beam processing to enhance manufacturing throughput. However, conventional multifocal systems remain constrained by bulky architectures, stringent alignment requirements, and susceptibility to laser-induced degradation under intense irradiation. Here, we demonstrate a monolithic multifocal diamond metalens with a 7.2 mm aperture that maintains exceptional thermal stability and power tolerance. The device employs high-aspect-ratio truncated-cone diamond nanopillars to generate two focal spots separated by 200 {\mu}m at a focal length of 4 mm. Under sustained 25 W pulsed-laser irradiation for 1 h, the diamond metalens exhibits a focal shift of only 25.5 {\mu}m, resulting in a maximum processing-depth variation of 33.2 {\mu}m during 4H silicon carbide (SiC) laser scribing, far below the 319.1 {\mu}m deviation observed for a commercial objective lens combined with a beam-splitting diffractive optical element (DOE). Even under extreme optical loading, the metalens withstands continuous-wave laser irradiation up to 8.25 kW for 30 s without structural degradation, while complementary pulsed testing yields a laser-induced damage threshold (LIDT) of 2.45 J/(cm^2) for diamond. This work broadens the operating envelope of transmissive meta-optics to extreme optical loads, opening new opportunities across high-power photonic systems.

physics.optics↗

arXiv2026

Reward Shaping for (Inference-Time) Alignment: A Stackelberg Game Perspective

Existing alignment methods directly use the reward model learned from user preference data to optimize an LLM policy, subject to KL regularization with respect to the base policy. This practice is suboptimal for maximizing user's utility because the KL regularization may cause the LLM to inherit the bias in the base policy that conflicts with user preferences. While amplifying rewards for preferred outputs can mitigate this bias, it also increases the risk of reward hacking. This tradeoff motivates the problem of optimally designing reward models under KL regularization. We formalize this reward model optimization problem as a Stackelberg game, and show that a simple reward shaping scheme can effectively approximate the optimal reward model. We empirically evaluate our method in inference-time alignment settings and demonstrate that it integrates seamlessly into existing alignment methods with minimal overhead. Our method consistently improves average reward and achieves win-tie rates exceeding 66% against all baselines, averaged across evaluation settings.

cs.LG↗

arXiv2026

Are LLMs Smarter Than Chimpanzees? An Evaluation on Perspective Taking and Knowledge State Estimation

Cognitive anthropology suggests that the distinction of human intelligence lies in the ability to infer other individuals' knowledge states and understand their intentions. In comparison, our closest animal relative, chimpanzees, lack the capacity to do so. With this paper, we aim to evaluate LLM performance in estimating other individuals' knowledge states and their potential actions. We design two tasks to test (1) if LLMs can predict story characters' next actions based on their own knowledge vs. improperly using information unavailable from their perspective, and (2) if LLMs can detect when story characters, through their actions, demonstrate knowledge they should not possess. Results reveal that most current state-of-the-art LLMs achieve near-random performance on both tasks, and are substantially inferior to humans. We argue future LLM research should place more weight on the abilities of knowledge estimation and intention understanding.

cs.AI↗

arXiv2025

JT-DA: Enhancing Data Analysis with Tool-Integrated Table Reasoning Large Language Models

In this work, we present JT-DA-8B (JiuTian Data Analyst 8B), a specialized large language model designed for complex table reasoning tasks across diverse real-world scenarios. To address the lack of high-quality supervision in tabular reasoning scenarios, we construct a comprehensive and diverse training corpus with 34 well-defined table reasoning tasks, by aggregating 29 public table QA datasets and 3 million tables. An automatic pipeline is proposed to generate realistic multi-step analytical tasks involving reasoning patterns. The model is trained upon open-source JT-Coder-8B model, an 8B-parameter decoder-only foundation model trained from scratch. In the training stage, we leverage LLM-based scoring and workflow-aligned filtering to distill high-quality, table-centric data. Both supervised fine-tuning (SFT) and Reinforcement learning (RL) are adopted to optimize our model. Afterwards, a four-stage table reasoning workflow is proposed, including table preprocessing, table sensing, tool-integrated reasoning, and prompt engineering, to improve model interpretability and execution accuracy. Experimental results show that JT-DA-8B achieves strong performance in various table reasoning tasks, demonstrating the effectiveness of data-centric generation and workflow-driven optimization.

cs.AI↗

arXiv2025

From Best Responses to Learning: Investment Efficiency in Dynamic Environment

We study the welfare of a mechanism in a dynamic environment where a learning investor can make a costly investment to change her value. In many real-world problems, the common assumption that the investor always makes the best responses, i.e., choosing her utility-maximizing investment option, is unrealistic due to incomplete information in a dynamically evolving environment. To address this, we consider an investor who uses a no-regret online learning algorithm to adaptively select investments through repeated interactions with the environment. We analyze how the welfare guarantees of approximation allocation algorithms extend from static to dynamic settings when the investor learns rather than best-responds, by studying the approximation ratio for optimal welfare as a measurement of an algorithm's performance against different benchmarks in the dynamic learning environment. First, we show that the approximation ratio in the static environment remains unchanged in the dynamic environment against the best-in-hindsight benchmark. Second, we provide tight characterizations of the approximation upper and lower bounds relative to a stronger time-varying benchmark. Bridging mechanism design with online learning theory, our work shows how robust welfare guarantees can be maintained even when an agent cannot make best responses but learns their investment strategies in complex, uncertain environments.

cs.GT↗

arXiv2025

TopoMAS: Large Language Model Driven Topological Materials Multiagent System

Topological materials occupy a frontier in condensed-matter physics thanks to their remarkable electronic and quantum properties, yet their cross-scale design remains bottlenecked by inefficient discovery workflows. Here, we introduce TopoMAS (Topological materials Multi-Agent System), an interactive human-AI framework that seamlessly orchestrates the entire materials-discovery pipeline: from user-defined queries and multi-source data retrieval, through theoretical inference and crystal-structure generation, to first-principles validation. Crucially, TopoMAS closes the loop by autonomously integrating computational outcomes into a dynamic knowledge graph, enabling continuous knowledge refinement. In collaboration with human experts, it has already guided the identification of novel topological phases SrSbO3, confirmed by first-principles calculations. Comprehensive benchmarks demonstrate robust adaptability across base Large Language Model, with the lightweight Qwen2.5-72B model achieving 94.55% accuracy while consuming only 74.3-78.4% of tokens required by Qwen3-235B and 83.0% of DeepSeek-V3's usage--delivering responses twice as fast as Qwen3-235B. This efficiency establishes TopoMAS as an accelerator for computation-driven discovery pipelines. By harmonizing rational agent orchestration with a self-evolving knowledge graph, our framework not only delivers immediate advances in topological materials but also establishes a transferable, extensible paradigm for materials-science domain.

cond-mat.mtrl-sci↗

arXiv2025

TReB: A Comprehensive Benchmark for Evaluating Table Reasoning Capabilities of Large Language Models

The majority of data in businesses and industries is stored in tables, databases, and data warehouses. Reasoning with table-structured data poses significant challenges for large language models (LLMs) due to its hidden semantics, inherent complexity, and structured nature. One of these challenges is lacking an effective evaluation benchmark fairly reflecting the performances of LLMs on broad table reasoning abilities. In this paper, we fill in this gap by presenting a comprehensive table reasoning benchmark, TReB. Firstly, we propose a taxonomy to systematically measure both shallow table understanding abilities and deep table reasoning abilities, covering a total of 26 sub-tasks. We then construct a high quality dataset through a dedicated data processing and synthesis procedure. Based on these well-constructed samples, we design an evaluation framework to robustly measure table reasoning capabilities with three distinct inference modes. Experimental results with our data and framework reveal that existing LLMs still have significant room for improvement in addressing the complex and real world table related tasks. Both the dataset and evaluation framework are publicly available, with the dataset hosted on https://huggingface.co/datasets/JT-LM/JIUTIAN-TReB, and the framework on https://github.com/JT-LM/jiutian-treb.

cs.CL↗

arXiv2024

Information Design with Unknown Prior

Information designers, such as online platforms, often do not know the beliefs of their receivers. We design learning algorithms so that the information designer can learn the receivers' prior belief from their actions through repeated interactions. Our learning algorithms achieve no regret relative to the optimality for the known prior at a fast speed, achieving a tight regret bound $\Theta(\log T)$ in general and a tight regret bound $\Theta(\log \log T)$ in the important special case of binary actions.

cs.GT↗

arXiv2024

Ultra-Thin, Ultra-Light, Rainbow-Free AR Glasses Based on Single-Layer Full-Color SiC Diffrcative Waveguide

As information interaction technology advances, the efficiency, dimensionality, and user experience of information transmission have significantly improved. Communication has evolved from letters to telegraphs, markedly increasing transmission speed; from telephones to video calls, enhancing communication dimensions; and from smartphones to augmented reality (AR) displays, which provide increasingly immersive user experiences. Surface relief grating (SRG) diffractive waveguides have attracted considerable attention for their optimal balance between weight, size, optical performance, and mass production capabilities, positioning them as a leading solution for AR displays. However, as consumer expectations for higher display quality and better device integration rise, traditional high-refractive-index glass-based diffractive waveguides face limitations, including bulkiness, heavy weight, and conspicuous rainbow artifacts in full-color displays. To overcome these challenges, a novel solution: ultra-thin, lightweight silicon carbide (SiC) AR prescription glasses was proposed. This solution achieves full-color displays without rainbow artifacts, with total weight of just 2.685 g and thickness of only 0.55 mm. Moreover, these glasses are compatible with prescription Fresnel lenses and are well-suited for scalable mass production. This innovation provides a robust platform for the seamless integration of augmented reality into daily life, offering significant potential to enhance user interaction.

physics.optics↗

arXiv2024

PromptSAM+: Malware Detection based on Prompt Segment Anything Model

Machine learning and deep learning (ML/DL) have been extensively applied in malware detection, and some existing methods demonstrate robust performance. However, several issues persist in the field of malware detection: (1) Existing work often overemphasizes accuracy at the expense of practicality, rarely considering false positive and false negative rates as important metrics. (2) Considering the evolution of malware, the performance of classifiers significantly declines over time, greatly reducing the practicality of malware detectors. (3) Prior ML/DL-based efforts heavily rely on ample labeled data for model training, largely dependent on feature engineering or domain knowledge to build feature databases, making them vulnerable if correct labels are scarce. With the development of computer vision, vision-based malware detection technology has also rapidly evolved. In this paper, we propose a visual malware general enhancement classification framework, `PromptSAM+', based on a large visual network segmentation model, the Prompt Segment Anything Model(named PromptSAM+). Our experimental results indicate that 'PromptSAM+' is effective and efficient in malware detection and classification, achieving high accuracy and low rates of false positives and negatives. The proposed method outperforms the most advanced image-based malware detection technologies on several datasets. 'PromptSAM+' can mitigate aging in existing image-based malware classifiers, reducing the considerable manpower needed for labeling new malware samples through active learning. We conducted experiments on datasets for both Windows and Android platforms, achieving favorable outcomes. Additionally, our ablation experiments on several datasets demonstrate that our model identifies effective modules within the large visual network.

cs.CR↗

arXiv2024

Mitigating the Impact of Malware Evolution on API Sequence-based Windows Malware Detector

In dynamic Windows malware detection, deep learning models are extensively deployed to analyze API sequences. Methods based on API sequences play a crucial role in malware prevention. However, due to the continuous updates of APIs and the changes in API sequence calls leading to the constant evolution of malware variants, the detection capability of API sequence-based malware detection models significantly diminishes over time. We observe that the API sequences of malware samples before and after evolution usually have similar malicious semantics. Specifically, compared to the original samples, evolved malware samples often use the API sequences of the pre-evolution samples to achieve similar malicious behaviors. For instance, they access similar sensitive system resources and extend new malicious functions based on the original functionalities. In this paper, we propose a framework MME(Mitigating the impact of Malware Evolution), a framework that can enhance existing API sequence-based malware detectors and mitigate the adverse effects of malware evolution. To help detection models capture the similar semantics of these post-evolution API sequences, our framework represents API sequences using API knowledge graphs and system resource encodings and applies contrastive learning to enhance the model's encoder. Results indicate that, compared to regular Text-CNN, our framework can significantly reduce the false positive rate by 13.10% and improve the F1-Score by 8.47% on five years of data, achieving the best experimental results. Additionally, evaluations show that our framework can save on the human costs required for model maintenance. We only need 1% of the budget per month to reduce the false positive rate by 11.16% and improve the F1-Score by 6.44%.

cs.CR↗

arXiv2023

Domain Adaptive Semantic Segmentation by Optimal Transport

Scene segmentation is widely used in the field of autonomous driving for environment perception, and semantic scene segmentation (3S) has received a great deal of attention due to the richness of the semantic information it contains. It aims to assign labels to pixels in an image, thus enabling automatic image labeling. Current approaches are mainly based on convolutional neural networks (CNN), but they rely on a large number of labels. Therefore, how to use a small size of labeled data to achieve semantic segmentation becomes more and more important. In this paper, we propose a domain adaptation (DA) framework based on optimal transport (OT) and attention mechanism to address this issue. Concretely, first we generate the output space via CNN due to its superiority of feature representation. Second, we utilize OT to achieve a more robust alignment of source and target domains in output space, where the OT plan defines a well attention mechanism to improve the adaptation of the model. In particular, with OT, the number of network parameters has been reduced and the network has been better interpretable. Third, to better describe the multi-scale property of features, we construct a multi-scale segmentation network to perform domain adaptation. Finally, in order to verify the performance of our proposed method, we conduct experimental comparison with three benchmark and four SOTA methods on three scene datasets, and the mean intersection-over-union (mIOU) has been significant improved, and visualization results under multiple domain adaptation scenarios also show that our proposed method has better performance than compared semantic segmentation methods.

cs.CV↗

arXiv2020

A Heterogeneous Dynamical Graph Neural Networks Approach to Quantify Scientific Impact

Quantifying and predicting the long-term impact of scientific writings or individual scholars has important implications for many policy decisions, such as funding proposal evaluation and identifying emerging research fields. In this work, we propose an approach based on Heterogeneous Dynamical Graph Neural Network (HDGNN) to explicitly model and predict the cumulative impact of papers and authors. HDGNN extends heterogeneous GNNs by incorporating temporally evolving characteristics and capturing both structural properties of attributed graph and the growing sequence of citation behavior. HDGNN is significantly different from previous models in its capability of modeling the node impact in a dynamic manner while taking into account the complex relations among nodes. Experiments conducted on a real citation dataset demonstrate its superior performance of predicting the impact of both papers and authors.

cs.SI↗

arXiv2018

Projection Convolutional Neural Networks for 1-bit CNNs via Discrete Back Propagation

The advancement of deep convolutional neural networks (DCNNs) has driven significant improvement in the accuracy of recognition systems for many computer vision tasks. However, their practical applications are often restricted in resource-constrained environments. In this paper, we introduce projection convolutional neural networks (PCNNs) with a discrete back propagation via projection (DBPP) to improve the performance of binarized neural networks (BNNs). The contributions of our paper include: 1) for the first time, the projection function is exploited to efficiently solve the discrete back propagation problem, which leads to a new highly compressed CNNs (termed PCNNs); 2) by exploiting multiple projections, we learn a set of diverse quantized kernels that compress the full-precision kernels in a more efficient way than those proposed previously; 3) PCNNs achieve the best classification performance compared to other state-of-the-art BNNs on the ImageNet and CIFAR datasets.

cs.CV↗

arXiv2018

Memory Attention Networks for Skeleton-based Action Recognition

Skeleton-based action recognition task is entangled with complex spatio-temporal variations of skeleton joints, and remains challenging for Recurrent Neural Networks (RNNs). In this work, we propose a temporal-then-spatial recalibration scheme to alleviate such complex variations, resulting in an end-to-end Memory Attention Networks (MANs) which consist of a Temporal Attention Recalibration Module (TARM) and a Spatio-Temporal Convolution Module (STCM). Specifically, the TARM is deployed in a residual learning module that employs a novel attention learning network to recalibrate the temporal attention of frames in a skeleton sequence. The STCM treats the attention calibrated skeleton joint sequences as images and leverages the Convolution Neural Networks (CNNs) to further model the spatial and temporal information of skeleton data. These two modules (TARM and STCM) seamlessly form a single network architecture that can be trained in an end-to-end fashion. MANs significantly boost the performance of skeleton-based action recognition and achieve the best results on four challenging benchmark datasets: NTU RGB+D, HDM05, SYSU-3D and UT-Kinect.

cs.CV↗