Selected work

Representative Papers

Beyond Perturbation Magnitude: Direction-Dependent Responses in Multimodal Geometric Representations

Oct 06, 2026

This study addresses the challenge that responses of multimodal geometric alignment scores to modality degradation cannot be adequately explained by perturbation magnitude alone, which accounts for only a small fraction of variance. To overcome this limitation, we propose Directional Geometric Response (DGR) theory, departing from conventional scalar perspectives. By leveraging Gramian volume gradient projections and first-order Taylor expansions, DGR integrates operating points, magnitudes, and directions to precisely model geometric volume variations. The framework is validated through experiments employing frozen embeddings with controlled audio-visual noise injection. Our findings demonstrate that directional dependence constitutes the primary driver of multimodal geometric responses. DGR achieves out-of-sample R² values ranging from 0.838 to 0.969 and ranking accuracy exceeding 0.864, significantly outperforming direction-agnostic baselines.

0 citationsRead paper

More Sensors Only One Field: Rethinking Continual Spatio-Temporal Forecasting

Sep 25, 2026

This study addresses the failure of conventional graph-based spatial representations in sensor networks undergoing expansion, which hinders continuous spatiotemporal forecasting. Motivated by the insight that sensor addition alters observational evidence rather than underlying dynamics, this work proposes the STFO operator and introduces a unified field evolution representation framework. The method parameterizes knowledge as a shared operator, decoupling layout variations through distinct observation and query interfaces. By integrating normalized coordinate aggregation, spectral descriptors, Fourier propagation, and attention mechanisms, it enables spatial mapping reuse while adaptively handling process drift. Experimental results demonstrate that STFO achieves state-of-the-art performance across multiple datasets, reducing the mean absolute error by 8.4% and 4.7% on the PEMS-Stream and CA-Stream benchmarks, respectively.

0 citationsRead paper

ReliableRAG: Combating Misinformation in Retrieval-Augmented Generation via Reliability-Guided Reasoning Chains

Aug 26, 2026

针对检索增强生成中虚假信息问题，提出ReliableRAG框架，通过细粒度评估和构建可靠的推理链来提高答案的准确性和鲁棒性。

0 citationsRead paper

Protein Structure Prediction: From Evolutionary Constraints to Generative Modeling

Aug 17, 2026

This study addresses the ambiguity surrounding the methodological evolution of protein structure prediction by proposing a "four-stage, three-transition" analytical framework. By systematically examining technical paradigm shifts across representation, architecture, and evaluation dimensions, this research integrates deep learning and generative modeling to elucidate the intrinsic trajectory from evolutionary constraints to generative design. Consequently, this work constructs a comprehensive methodological evolution map that clarifies historical transitions in model capabilities and application roles. Ultimately, it provides a systematic theoretical foundation for understanding the developmental logic and future trends within the field, offering critical insights into how predictive methodologies have matured over time.

0 citationsRead paper

Color Fundus Photography Analysis: Co-evolution of Data, Preprocessing, and Modeling toward Multimodal AI

Jul 26, 2026

Current AI research on color fundus photography (CFP) lacks a systematic perspective on the co-evolution of datasets, preprocessing, and modeling. This work proposes the first unified framework that synergistically optimizes data curation, preprocessing, and multimodal modeling, integrating neural data engineering, hardware-aware annotation, self-supervised electronic health record imputation, vision foundation models, state space models, and multimodal mixture-of-experts architectures. The study delineates a clear evolutionary trajectory for CFP analysis—from single-task convolutional neural networks toward multimodal, longitudinally integrated clinical systems—and provides a methodological roadmap toward robust ophthalmic AI capable of clinical deployment, cross-domain generalization, and edge intelligence.

0 citationsRead paper

Recent publications

Latest Papers

Beyond Perturbation Magnitude: Direction-Dependent Responses in Multimodal Geometric Representations

Oct 06, 2026

This study addresses the challenge that responses of multimodal geometric alignment scores to modality degradation cannot be adequately explained by perturbation magnitude alone, which accounts for only a small fraction of variance. To overcome this limitation, we propose Directional Geometric Response (DGR) theory, departing from conventional scalar perspectives. By leveraging Gramian volume gradient projections and first-order Taylor expansions, DGR integrates operating points, magnitudes, and directions to precisely model geometric volume variations. The framework is validated through experiments employing frozen embeddings with controlled audio-visual noise injection. Our findings demonstrate that directional dependence constitutes the primary driver of multimodal geometric responses. DGR achieves out-of-sample R² values ranging from 0.838 to 0.969 and ranking accuracy exceeding 0.864, significantly outperforming direction-agnostic baselines.

0 citationsRead paper

More Sensors Only One Field: Rethinking Continual Spatio-Temporal Forecasting

Sep 25, 2026

This study addresses the failure of conventional graph-based spatial representations in sensor networks undergoing expansion, which hinders continuous spatiotemporal forecasting. Motivated by the insight that sensor addition alters observational evidence rather than underlying dynamics, this work proposes the STFO operator and introduces a unified field evolution representation framework. The method parameterizes knowledge as a shared operator, decoupling layout variations through distinct observation and query interfaces. By integrating normalized coordinate aggregation, spectral descriptors, Fourier propagation, and attention mechanisms, it enables spatial mapping reuse while adaptively handling process drift. Experimental results demonstrate that STFO achieves state-of-the-art performance across multiple datasets, reducing the mean absolute error by 8.4% and 4.7% on the PEMS-Stream and CA-Stream benchmarks, respectively.

0 citationsRead paper

ReliableRAG: Combating Misinformation in Retrieval-Augmented Generation via Reliability-Guided Reasoning Chains

Aug 26, 2026

针对检索增强生成中虚假信息问题，提出ReliableRAG框架，通过细粒度评估和构建可靠的推理链来提高答案的准确性和鲁棒性。

0 citationsRead paper

Protein Structure Prediction: From Evolutionary Constraints to Generative Modeling

Aug 17, 2026

This study addresses the ambiguity surrounding the methodological evolution of protein structure prediction by proposing a "four-stage, three-transition" analytical framework. By systematically examining technical paradigm shifts across representation, architecture, and evaluation dimensions, this research integrates deep learning and generative modeling to elucidate the intrinsic trajectory from evolutionary constraints to generative design. Consequently, this work constructs a comprehensive methodological evolution map that clarifies historical transitions in model capabilities and application roles. Ultimately, it provides a systematic theoretical foundation for understanding the developmental logic and future trends within the field, offering critical insights into how predictive methodologies have matured over time.

0 citationsRead paper

Color Fundus Photography Analysis: Co-evolution of Data, Preprocessing, and Modeling toward Multimodal AI

Jul 26, 2026

Current AI research on color fundus photography (CFP) lacks a systematic perspective on the co-evolution of datasets, preprocessing, and modeling. This work proposes the first unified framework that synergistically optimizes data curation, preprocessing, and multimodal modeling, integrating neural data engineering, hardware-aware annotation, self-supervised electronic health record imputation, vision foundation models, state space models, and multimodal mixture-of-experts architectures. The study delineates a clear evolutionary trajectory for CFP analysis—from single-task convolutional neural networks toward multimodal, longitudinally integrated clinical systems—and provides a methodological roadmap toward robust ophthalmic AI capable of clinical deployment, cross-domain generalization, and edge intelligence.

0 citationsRead paper