Selected work

Representative Papers

Unified Interactive Multimodal Moment Retrieval via Cascaded Embedding-Reranking and Temporal-Aware Score Fusion

Dec 14, 2025

To address three key challenges in video moment retrieval—poor cross-modal noise robustness, weak modeling of event temporal coherence, and heavy reliance on manual modality selection—this paper proposes an end-to-end interactive multimodal moment retrieval framework. Methodologically: (1) a cascaded dual-embedding re-ranking architecture enhances cross-modal alignment; (2) a temporal-aware exponentially decaying scoring mechanism explicitly models event continuity while suppressing implausible time intervals; (3) GPT-4o enables automatic query decomposition and adaptive score fusion, eliminating manual modality specification. The system integrates BEIT-3, SigLIP, BLIP-2, and ASR/OCR features, with beam search for temporally constrained optimization. Experiments demonstrate substantial improvements in recall-precision trade-off under ambiguous queries, generation of highly coherent event sequences, and enhanced practicality of video retrieval systems.

0 citationsRead paper

Wariness and Poverty Traps

Oct 16, 2025

This paper examines how individual prudence—the concern for the minimum utility over time—affects poverty traps and equilibrium multiplicity in an overlapping-generations (OLG) framework. Employing a dynamic general equilibrium model with an endogenous utility floor constraint, the analysis combines comparative statics and transitional path analysis. Results show that prudence does not unilaterally exacerbate poverty: when productivity is low or factor substitution elasticity is small, heightened prudence deepens poverty traps and generates multiple steady states—including a low-growth equilibrium; conversely, under higher productivity or greater substitutability, it may facilitate convergence. The key contribution lies in endogenizing the “minimum utility concern” as an intertemporal decision constraint—thereby identifying a novel interaction between prudence and technological structure. This extension enriches the OLG paradigm’s capacity to explain persistent inequality and growth stagnation.

0 citationsRead paper

Formal Reasoning for Intelligent QA Systems: A Case Study in the Educational Domain

Sep 15, 2025

To address the lack of causal reasoning and verifiability in large language models (LLMs) for high-stakes question answering in closed domains (e.g., education), this paper proposes MCFR—a novel framework that tightly integrates LLMs with formal model checking for the first time. MCFR employs a neuro-symbolic architecture to automatically translate natural-language questions into formal specifications expressed as state-transition systems, and then performs property verification over the constructed transition system. It supports verifiable multi-step dynamic reasoning, including conditional transitions and procedural progress. Evaluated on EduMC-QA, a custom educational benchmark, MCFR significantly outperforms baseline LLMs—including ChatGPT, DeepSeek, and Claude—in reasoning accuracy and logical consistency. Moreover, it ensures factual correctness, interpretability, and regulatory compliance through formally grounded inference.

0 citationsRead paper

Recent publications

Latest Papers

Unified Interactive Multimodal Moment Retrieval via Cascaded Embedding-Reranking and Temporal-Aware Score Fusion

Dec 14, 2025

To address three key challenges in video moment retrieval—poor cross-modal noise robustness, weak modeling of event temporal coherence, and heavy reliance on manual modality selection—this paper proposes an end-to-end interactive multimodal moment retrieval framework. Methodologically: (1) a cascaded dual-embedding re-ranking architecture enhances cross-modal alignment; (2) a temporal-aware exponentially decaying scoring mechanism explicitly models event continuity while suppressing implausible time intervals; (3) GPT-4o enables automatic query decomposition and adaptive score fusion, eliminating manual modality specification. The system integrates BEIT-3, SigLIP, BLIP-2, and ASR/OCR features, with beam search for temporally constrained optimization. Experiments demonstrate substantial improvements in recall-precision trade-off under ambiguous queries, generation of highly coherent event sequences, and enhanced practicality of video retrieval systems.

0 citationsRead paper

Wariness and Poverty Traps

Oct 16, 2025

This paper examines how individual prudence—the concern for the minimum utility over time—affects poverty traps and equilibrium multiplicity in an overlapping-generations (OLG) framework. Employing a dynamic general equilibrium model with an endogenous utility floor constraint, the analysis combines comparative statics and transitional path analysis. Results show that prudence does not unilaterally exacerbate poverty: when productivity is low or factor substitution elasticity is small, heightened prudence deepens poverty traps and generates multiple steady states—including a low-growth equilibrium; conversely, under higher productivity or greater substitutability, it may facilitate convergence. The key contribution lies in endogenizing the “minimum utility concern” as an intertemporal decision constraint—thereby identifying a novel interaction between prudence and technological structure. This extension enriches the OLG paradigm’s capacity to explain persistent inequality and growth stagnation.

0 citationsRead paper

Formal Reasoning for Intelligent QA Systems: A Case Study in the Educational Domain

Sep 15, 2025

To address the lack of causal reasoning and verifiability in large language models (LLMs) for high-stakes question answering in closed domains (e.g., education), this paper proposes MCFR—a novel framework that tightly integrates LLMs with formal model checking for the first time. MCFR employs a neuro-symbolic architecture to automatically translate natural-language questions into formal specifications expressed as state-transition systems, and then performs property verification over the constructed transition system. It supports verifiable multi-step dynamic reasoning, including conditional transitions and procedural progress. Evaluated on EduMC-QA, a custom educational benchmark, MCFR significantly outperforms baseline LLMs—including ChatGPT, DeepSeek, and Claude—in reasoning accuracy and logical consistency. Moreover, it ensures factual correctness, interpretability, and regulatory compliance through formally grounded inference.

0 citationsRead paper