arXiv subjects

Xiaodi Li

Publications and source records attributed to Xiaodi Li.

At least 19 recordsLinked to original sources

arXiv2026

Optimizing Clinical Trial Protocols Using EHR-Derived Heterogeneous Treatment Effects

Traditional randomized trials often obscure clinically meaningful heterogeneity in treatment response by focusing on average effects. Leveraging real-world data to emulate clinical trials and estimate heterogeneous treatment effects (HTEs) offers a promising path toward more precise and efficient trial design. In this study, we emulate the DAPA-HF trial using electronic health records from the Mayo Clinic Cloud (MCC) to investigate whether HTE-guided stratification can identify patient subgroups with distinct treatment responses to dapagliflozin versus placebo in patients with heart failure with reduced ejection fraction. All-cause mortality was evaluated using Cox proportional hazards models, with HTEs estimated using a Meta-S learner and subgroups defined using a decision tree-based thresholding approach. In the overall cohort of the emulation, no significant treatment difference was observed (HR, 1.681; 95% CI, 0.828-3.413; p = 0.1507). However, compared with the overall emulated cohort, in which dapagliflozin showed no statistically significant survival benefit, HTE-driven stratification identified subgroups with significant and directionally distinct treatment effects. The beneficial (low-HTE) subgroup showed a significant survival benefit from dapagliflozin (HR = 0.203, 95% CI, 0.087-0.476, p = 0.0002), whereas the harmful (high-HTE) subgroup showed a significant harmful association with markedly increased mortality risk (HR = 6.680, 95% CI, 2.759-16.171, p < 0.0001). These findings indicate that HTE-guided stratification can uncover clinically meaningful beneficial and harmful treatment-effect patterns that are masked in the full-cohort emulation.

stat.AP↗

arXiv2026

Lightweight Retrieval-Augmented Generation and Large Language Model-Based Modeling for Scalable Patient-Trial Matching

Patient-trial matching requires reasoning over long, heterogeneous electronic health records (EHRs) and complex eligibility criteria, posing significant challenges for scalability, generalization, and computational efficiency. Existing approaches either rely on full-document processing with large language models (LLMs), which is computationally expensive, or use traditional machine learning methods that struggle to capture unstructured clinical narratives. In this work, we propose a lightweight framework that combines retrieval-augmented generation and large language model-based modeling for scalable patient-trial matching. The framework explicitly separates two key components: retrieval-augmented generation is used to identify clinically relevant segments from long EHRs, reducing input complexity, while large language models are used to encode these selected segments into informative representations. These representations are further refined through dimensionality reduction and modeled using lightweight predictors, enabling efficient and scalable downstream classification. We evaluate the proposed approach on multiple public benchmarks (n2c2, SIGIR, TREC 2021/2022) and a real-world multimodal dataset from Mayo Clinic (MCPMD). Results show that retrieval-based information selection significantly reduces computational burden while preserving clinically meaningful signals. We further demonstrate that frozen LLMs provide strong representations for structured clinical data, whereas fine-tuning is essential for modeling unstructured clinical narratives. Importantly, the proposed lightweight pipeline achieves performance comparable to end-to-end LLM approaches with substantially lower computational cost.

cs.CL↗

arXiv2026

An Algorithm for Fast Assembling Large-Scale Defect-Free Atom Arrays

It is widely believed that tens of thousands of physical qubits are needed to build a practically useful quantum computer. Atom arrays formed by optical tweezers are among the most promising platforms for achieving this goal, owing to the excellent scalability and mobility of atomic qubits. However, assembling a defect-free atom array with ~ 10^4 qubits remains algorithmically challenging, alongside other hardware limitations. This is due to the computationally hard path-planning problems and the time-consuming generation of suffciently smooth trajectories for optical tweezer potentials by spatial light modulators (SLM). Here, we present a unified framework comprising two innovative components to fully address these algorithmic challenges: (1) a path-planning module that employs a supervised learning approach using a graph neural network combined with a modified auction decoder, and (2) a potential-generation module called the phase and profile-aware Weighted Gerchberg-Saxton algorithm. The inference time for the first module is nearly a size-independent constant overhead of ~ 5 ms, and the second module generates a potential frame with about 0.5 ms, a timescale shorter than the current commercial SLM refresh time. Altogether, our algorithm enables the assembly of an atom array with 10^4 qubits on a timescale much shorter than the typical vacuum lifetime of the trapped atoms.

cond-mat.quant-gas↗

arXiv2026

Certifying entanglement dimensionality by $k$-reduction moments

In this paper, we combine the k-reduction map, the moment method, and the classical shadow method into a practical protocol for certifying the entanglement dimensionality. Our approach is based on the observation that a state with entanglement dimensionality at most k must stay positive under the action of the k-reduction map. The core of our protocol utilizes the moment method to determine whether the k-reduced operator, i.e., the operator obtained after applying the k-reduction map on a quantum state, contains negative eigenvalues or not. Notably, we propose a systematic method for constructing reduction moment criteria, which apply to a much wider range of states than fidelity-based methods. The performance of our approach gets better and better with the moment order employed, which is corroborated by extensive numerical simulation. To apply our approach, it suffices to implement a unitary 3-design instead of a 4-design, which is more feasible in practice than the correlation matrix method. In the course of study, we show that the k-reduction negativity, the absolute sum of the negative eigenvalues of the k-reduced operator, is monotonic under local operations and classical communication for pure states.

quant-ph↗

arXiv2026

Stability analysis of nontrivial stationary solution and constant equilibrium point of reaction-diffusion neural networks with time delays under Dirichlet zero boundary value

In this paper, Lyapunov-Razumikhin technique, design of state-dependent switching laws, a fixed point theorem and variational methods are employed to derive the existence and the unique existence results of globally exponentially stable (positive) stationary solution of delayed reaction-diffusion cell neural networks under Dirichlet zero boundary value, including the global stability criteria in the classical meaning. Next, sufficient conditions are proposed to guarantee the global stability invariance of ordinary differential systems under the influence of diffusions. New theorems show that the diffusion is a double-edged sword in judging the stability of diffusion systems. Besides, an example is constructed to illuminate that any non-zero constant equilibrium point must be not in the phase plane of dynamic system under Dirichlet zero boundary value, or it must lead to a contradiction. Next, under Lipschitz assumptions on active function, another example is designed to prove that the small diffusion effect will cause the essential change of the phase plane structure of the dynamic behavior of the delayed neural networks via a Saddle point theorem. Finally, a numerical example illustrates the feasibility of the proposed methods. It is worth mentioning that some interesting mathematical problems are originally put forward in the last chapter.

math.DS↗

arXiv2025

PPSEBM: An Energy-Based Model with Progressive Parameter Selection for Continual Learning

Continual learning remains a fundamental challenge in machine learning, requiring models to learn from a stream of tasks without forgetting previously acquired knowledge. A major obstacle in this setting is catastrophic forgetting, where performance on earlier tasks degrades as new tasks are learned. In this paper, we introduce PPSEBM, a novel framework that integrates an Energy-Based Model (EBM) with Progressive Parameter Selection (PPS) to effectively address catastrophic forgetting in continual learning for natural language processing tasks. In PPSEBM, progressive parameter selection allocates distinct, task-specific parameters for each new task, while the EBM generates representative pseudo-samples from prior tasks. These generated samples actively inform and guide the parameter selection process, enhancing the model's ability to retain past knowledge while adapting to new tasks. Experimental results on diverse NLP benchmarks demonstrate that PPSEBM outperforms state-of-the-art continual learning methods, offering a promising and robust solution to mitigate catastrophic forgetting.

cs.CL↗

arXiv2025

Enhancing Lung Cancer Treatment Outcome Prediction through Semantic Feature Engineering Using Large Language Models

Accurate prediction of treatment outcomes in lung cancer remains challenging due to the sparsity, heterogeneity, and contextual overload of real-world electronic health data. Traditional models often fail to capture semantic information across multimodal streams, while large-scale fine-tuning approaches are impractical in clinical workflows. We introduce a framework that uses Large Language Models (LLMs) as Goal-oriented Knowledge Curators (GKC) to convert laboratory, genomic, and medication data into high-fidelity, task-aligned features. Unlike generic embeddings, GKC produces representations tailored to the prediction objective and operates as an offline preprocessing step that integrates naturally into hospital informatics pipelines. Using a lung cancer cohort (N=184), we benchmarked GKC against expert-engineered features, direct text embeddings, and an end-to-end transformer. Our approach achieved a mean AUROC of 0.803 (95% CI: 0.799-0.807) and outperformed all baselines. An ablation study further confirmed the complementary value of combining all three modalities. These results show that the quality of semantic representation is a key determinant of predictive accuracy in sparse clinical data settings. By reframing LLMs as knowledge curation engines rather than black-box predictors, this work demonstrates a scalable, interpretable, and workflow-compatible pathway for advancing AI-driven decision support in oncology.

cs.LG↗

arXiv2025

A new general quantum state verification protocol by the classical shadow method

Verifying whether a quantum device produces a specific state is a fundamental task in many applications of modern quantum technologies. In the conventional framework of quantum state verification, designing an optimal or efficient protocol for each type of state often requires intricate customization. Recently, Hsin-Yuan Huang et al. introduced a novel approach called the shadow overlap protocol in arXiv:2404.07281v1, which leverages classical shadows to efficiently verify multiple types of states simultaneously. In this paper, we propose a new verification protocol that integrates key ideas from both the conventional framework and the shadow overlap protocol. To achieve this, we first reformulate the shadow overlap protocol using the language of hypothesis testing, which also underpins the conventional framework, and then analyze the similarities and differences between them. Our approach enhances the capabilities of the shadow overlap protocol while addressing some of its limitations, leading to improved sample complexity and a more natural treatment of states with specific structures. Finally, we demonstrate the effectiveness of our protocol by applying it to GHZ states and stabilizer states.

quant-ph↗

arXiv2025

InfinityHuman: Towards Long-Term Audio-Driven Human

Audio-driven human animation has attracted wide attention thanks to its practical applications. However, critical challenges remain in generating high-resolution, long-duration videos with consistent appearance and natural hand motions. Existing methods extend videos using overlapping motion frames but suffer from error accumulation, leading to identity drift, color shifts, and scene instability. Additionally, hand movements are poorly modeled, resulting in noticeable distortions and misalignment with the audio. In this work, we propose InfinityHuman, a coarse-to-fine framework that first generates audio-synchronized representations, then progressively refines them into high-resolution, long-duration videos using a pose-guided refiner. Since pose sequences are decoupled from appearance and resist temporal degradation, our pose-guided refiner employs stable poses and the initial frame as a visual anchor to reduce drift and improve lip synchronization. Moreover, to enhance semantic accuracy and gesture realism, we introduce a hand-specific reward mechanism trained with high-quality hand motion data. Experiments on the EMTD and HDTF datasets show that InfinityHuman achieves state-of-the-art performance in video quality, identity preservation, hand accuracy, and lip-sync. Ablation studies further confirm the effectiveness of each module. Code will be made public.

cs.CV↗

arXiv2025

A new recursion relation for tree-level NLSM amplitudes based on hidden zeros

In this note, we propose a novel BCFW-like recursion relation for tree-level non-linear sigma model (NLSM) amplitudes, which circumvents the computation of boundary terms by exploiting the recently discovered hidden zeros. Using this recursion, we reproduce three remarkable features of tree-level NLSM amplitudes: (i) the Adler zero, (ii) the $\delta$-shift construction, which generates NLSM amplitudes from ${\rm Tr}(\phi^3)$ amplitudes, and (iii) the universal expansion of NLSM amplitudes into bi-adjoint scalar amplitudes. Our results demonstrate that the hidden zeros, combined with standard factorization on physical poles, uniquely determine all tree-level NLSM amplitudes.

hep-th↗

arXiv2025

Random approximate quantum information masking

Masking information into quantum correlations is a cornerstone of many quantum information applications. While there exist the no-hiding and no-masking theorems, approximate quantum information masking (AQIM) offers a promising means of circumventing the constraints. Despite its potential, AQIM still remains underexplored, and constructing explicit approximate maskers remains a challenge. In this work, we investigate AQIM from multiple perspectives and propose using random isometries to construct approximate maskers. First, different notions of AQIM are introduced and we find there are profound intrinsic connections among them. These relationships are characterized by a set of figures of merit, which are introduced to quantify the deviation of AQIM from exact QIM. We then explore the possibility of realizing AQIM via random isometries in bipartite and multipartite systems. In bipartite systems, we identify a fundamental lower bound for a key figure of merit, implying that almost all random isometries fail to realize AQIM. This surprising result generalizes the original no-masking theorem to the no-random-AQIM theorem for bipartite systems. In contrast, in multipartite systems, we show almost all random isometries can realize AQIM. Remarkably, the number of physical qubits required to randomly mask a single logical qubit scales only linearly. We further explore the implications of these findings. In particular, we show that, under certain conditions, approximate quantum error correction is equivalent to AQIM. Consequently, AQIM naturally gives rise to approximate quantum error correction codes with constant code rates and exponentially small correction inaccuracies. Overall, our results establish quantum information masking as a central concept in quantum information theory, bridging diverse notions across multiple domains.

quant-ph↗

arXiv2025

LLM-Match: An Open-Sourced Patient Matching Model Based on Large Language Models and Retrieval-Augmented Generation

Patient matching is the process of linking patients to appropriate clinical trials by accurately identifying and matching their medical records with trial eligibility criteria. We propose LLM-Match, a novel framework for patient matching leveraging fine-tuned open-source large language models. Our approach consists of four key components. First, a retrieval-augmented generation (RAG) module extracts relevant patient context from a vast pool of electronic health records (EHRs). Second, a prompt generation module constructs input prompts by integrating trial eligibility criteria (both inclusion and exclusion criteria), patient context, and system instructions. Third, a fine-tuning module with a classification head optimizes the model parameters using structured prompts and ground-truth labels. Fourth, an evaluation module assesses the fine-tuned model's performance on the testing datasets. We evaluated LLM-Match on four open datasets - n2c2, SIGIR, TREC 2021, and TREC 2022 - using open-source models, comparing it against TrialGPT, Zero-Shot, and GPT-4-based closed models. LLM-Match outperformed all baselines.

cs.CL↗

arXiv2025

Launching Insights: A Pilot Study on Leveraging Real-World Observational Data from the Mayo Clinic Platform to Advance Clinical Research

Backgrounds: Artificial intelligence (AI) is transforming healthcare, yet translating AI models from theoretical frameworks to real-world clinical applications remains challenging. The Mayo Clinic Platform (MCP) was established to address these challenges by providing a scalable ecosystem that integrates real-world multiple modalities data from multiple institutions, advanced analytical tools, and secure computing environments to support clinical research and AI development. Methods: In this study, we conducted four research projects leveraging MCP's data infrastructure and analytical capabilities to demonstrate its potential in facilitating real-world evidence generation and AI-driven clinical insights. Utilizing MCP's tools and environment, we facilitated efficient cohort identification, data extraction, and subsequent statistical or AI-powered analyses. Results: The results underscore MCP's role in accelerating translational research by offering de-identified, standardized real-world data and facilitating AI model validation across diverse healthcare settings. Compared to Mayo's internal Electronic Health Record (EHR) data, MCP provides broader accessibility, enhanced data standardization, and multi-institutional integration, making it a valuable resource for both internal and external researchers. Conclusion: Looking ahead, MCP is well-positioned to transform clinical research through its scalable ecosystem, effectively bridging the divide between AI innovation and clinical deployment. Future investigations will build upon this foundation, further exploring MCP's capacity to advance precision medicine and enhance patient outcomes.

cs.CY↗

arXiv2025

DoraemonGPT: Toward Understanding Dynamic Scenes with Large Language Models (Exemplified as A Video Agent)

Recent LLM-driven visual agents mainly focus on solving image-based tasks, which limits their ability to understand dynamic scenes, making it far from real-life applications like guiding students in laboratory experiments and identifying their mistakes. Hence, this paper explores DoraemonGPT, a comprehensive and conceptually elegant system driven by LLMs to understand dynamic scenes. Considering the video modality better reflects the ever-changing nature of real-world scenarios, we exemplify DoraemonGPT as a video agent. Given a video with a question/task, DoraemonGPT begins by converting the input video into a symbolic memory that stores task-related attributes. This structured representation allows for spatial-temporal querying and reasoning by well-designed sub-task tools, resulting in concise intermediate results. Recognizing that LLMs have limited internal knowledge when it comes to specialized domains (e.g., analyzing the scientific principles underlying experiments), we incorporate plug-and-play tools to assess external knowledge and address tasks across different domains. Moreover, a novel LLM-driven planner based on Monte Carlo Tree Search is introduced to explore the large planning space for scheduling various tools. The planner iteratively finds feasible solutions by backpropagating the result's reward, and multiple solutions can be summarized into an improved final answer. We extensively evaluate DoraemonGPT's effectiveness on three benchmarks and several in-the-wild scenarios. The code will be released at https://github.com/z-x-yang/DoraemonGPT.

cs.CV↗

arXiv2025

LSEBMCL: A Latent Space Energy-Based Model for Continual Learning

Continual learning has become essential in many practical applications such as online news summaries and product classification. The primary challenge is known as catastrophic forgetting, a phenomenon where a model inadvertently discards previously learned knowledge when it is trained on new tasks. Existing solutions involve storing exemplars from previous classes, regularizing parameters during the fine-tuning process, or assigning different model parameters to each task. The proposed solution LSEBMCL (Latent Space Energy-Based Model for Continual Learning) in this work is to use energy-based models (EBMs) to prevent catastrophic forgetting by sampling data points from previous tasks when training on new ones. The EBM is a machine learning model that associates an energy value with each input data point. The proposed method uses an EBM layer as an outer-generator in the continual learning framework for NLP tasks. The study demonstrates the efficacy of EBM in NLP tasks, achieving state-of-the-art results in all experiments.

cs.LG↗

arXiv2024

Towards Scalable Automated Grading: Leveraging Large Language Models for Conceptual Question Evaluation in Engineering

This study explores the feasibility of using large language models (LLMs), specifically GPT-4o (ChatGPT), for automated grading of conceptual questions in an undergraduate Mechanical Engineering course. We compared the grading performance of GPT-4o with that of human teaching assistants (TAs) on ten quiz problems from the MEEN 361 course at Texas A&M University, each answered by approximately 225 students. Both the LLM and TAs followed the same instructor-provided rubric to ensure grading consistency. We evaluated performance using Spearman's rank correlation coefficient and Root Mean Square Error (RMSE) to assess the alignment between rankings and the accuracy of scores assigned by GPT-4o and TAs under zero- and few-shot grading settings. In the zero-shot setting, GPT-4o demonstrated a strong correlation with TA grading, with Spearman's rank correlation coefficient exceeding 0.6 in seven out of ten datasets and reaching a high of 0.9387. Our analysis reveals that GPT-4o performs well when grading criteria are straightforward but struggles with nuanced answers, particularly those involving synonyms not present in the rubric. The model also tends to grade more stringently in ambiguous cases compared to human TAs. Overall, ChatGPT shows promise as a tool for grading conceptual questions, offering scalability and consistency.

cs.CY↗

arXiv2024

Controllable Edge-Type-Specific Interpretation in Multi-Relational Graph Neural Networks for Drug Response Prediction

Graph Neural Networks have been widely applied in critical decision-making areas that demand interpretable predictions, leading to the flourishing development of interpretability algorithms. However, current graph interpretability algorithms tend to emphasize generality and often overlook biological significance, thereby limiting their applicability in predicting cancer drug responses. In this paper, we propose a novel post-hoc interpretability algorithm for cancer drug response prediction, CETExplainer, which incorporates a controllable edge-type-specific weighting mechanism. It considers the mutual information between subgraphs and predictions, proposing a structural scoring approach to provide fine-grained, biologically meaningful explanations for predictive models. We also introduce a method for constructing ground truth based on real-world datasets to quantitatively evaluate the proposed interpretability algorithm. Empirical analysis on the real-world dataset demonstrates that CETExplainer achieves superior stability and improves explanation quality compared to leading algorithms, thereby offering a robust and insightful tool for cancer drug prediction.

cs.LG↗

arXiv2024

DRExplainer: Quantifiable Interpretability in Drug Response Prediction with Directed Graph Convolutional Network

Predicting the response of a cancer cell line to a therapeutic drug is pivotal for personalized medicine. Despite numerous deep learning methods that have been developed for drug response prediction, integrating diverse information about biological entities and predicting the directional response remain major challenges. Here, we propose a novel interpretable predictive model, DRExplainer, which leverages a directed graph convolutional network to enhance the prediction in a directed bipartite network framework. DRExplainer constructs a directed bipartite network integrating multi-omics profiles of cell lines, the chemical structure of drugs and known drug response to achieve directed prediction. Then, DRExplainer identifies the most relevant subgraph to each prediction in this directed bipartite network by learning a mask, facilitating critical medical decision-making. Additionally, we introduce a quantifiable method for model interpretability that leverages a ground truth benchmark dataset curated from biological features. In computational experiments, DRExplainer outperforms state-of-the-art predictive methods and another graph-based explanation method under the same experimental setting. Finally, the case studies further validate the interpretability and the effectiveness of DRExplainer in predictive novel drug response. Our code is available at: https://github.com/vshy-dream/DRExplainer.

cs.LG↗