Selected work

Representative Papers

Spherical Geometrical Bases of Spherical Origami

May 01, 2026

This study addresses the absence of a rigorous geometric framework for spherical origami, particularly concerning folding rules on the unit sphere and their realization in three-dimensional space. The work proposes the first complete axiomatic system for spherical origami by extending the classical Huzita–Justin axioms from Euclidean to spherical geometry, providing explicit equations for all seven axioms. Innovatively, it replaces geodesics with isometric curves as three-dimensional creases, thereby expanding the repertoire of realizable folded forms. By integrating spherical and differential geometry with computational origami theory and computer graphics, the authors successfully generate a spherical origami bird model, demonstrating both the theoretical completeness and practical feasibility of the proposed framework.

0 citationsRead paper

Apple Peel Unfolding of Archimedean and Catalan Solids

Apr 17, 2026

This study investigates whether Archimedean solids and their duals, the Catalan solids, can be unfolded into non-overlapping planar nets via a continuous “apple-peeling” path—traversing adjacent faces without gaps or overlaps. We formalize this unfolding paradigm for the first time, introducing rigorous face-selection rules and developing an automated verification algorithm grounded in geometric and topological analysis. Applying this framework, we systematically classify the unfoldability of all Archimedean and Catalan solids: three Archimedean and six Catalan solids admit perfect apple-peeling unfoldings; three Archimedean and three Catalan solids are unfoldable only under restricted conditions; and the remaining solids cannot be unfolded in this manner.

0 citationsRead paper

D-COT: Disciplined Chain-of-Thought Learning for Efficient Reasoning in Small Language Models

Feb 25, 2026

This work addresses the performance degradation and increased computational overhead in small language models (SLMs) caused by “overthinking” during chain-of-thought (CoT) distillation from large language models. To mitigate reasoning drift, the authors propose a structured reasoning framework that introduces control tokens—such as <TEMP_LOW> and <TEMP_HIGH>—to guide SLMs along efficient, ordered reasoning pathways. The approach integrates control-token-driven CoT distillation, structured trajectory optimization, and few-shot fine-tuning. Evaluated on Qwen3-8B with only 5,000 training samples, the method achieves a 9.9% absolute accuracy gain on GPQA-diamond and a 9.1% improvement on zero-shot MMLU-Pro, while substantially reducing token consumption. These results demonstrate the model’s enhanced capacity to internalize structured reasoning patterns.

0 citationsRead paper

Digital Nature Revisited: A Ten-Year Synthesis of Art, Technology, and the Evolution of"Nature": Reimagining Post-Truth Ecologies Through Art, Algorithm, and Animism

Nov 11, 2025

This paper addresses the persistent conceptual ambiguity and interdisciplinary divergence surrounding the term “digital nature” over the past decade. Methodologically, it integrates media art, bio-art, and generative art practices with algorithmic creation, large language models, and philosophical inquiry to systematically map the concept’s genealogy and expose the ethical–political risks embedded in techno-uncanny phenomena. Its contributions are threefold: first, it proposes a novel dual-axis analytical model—spanning virtual/real and anthropocentric/object-oriented dimensions; second, it repositions digital nature as a platform for intercivilizational dialogue; and third, it introduces the concept of “supernatural reality” to extend critical engagement with post-truth ecologies. Collectively, these advances advance art–technology collaboration toward a more culturally grounded and critically reflexive practice paradigm.

0 citationsRead paper

Investigation of Frame Differences as Motion Cues for Video Object Segmentation

Mar 12, 2025

For automatic video object segmentation (AVOS) without first-frame annotations, existing methods rely on computationally expensive optical flow to extract motion cues, hindering real-time deployment on edge devices. This paper proposes replacing optical flow with lightweight frame differencing as a motion prior and designs an extended U-Net architecture for end-to-end segmentation. We systematically demonstrate— for the first time—that frame differencing provides motion cues comparable in effectiveness to optical flow under static-camera conditions, while drastically reducing computational overhead. Experiments on standard AVOS benchmarks show our method achieves a 0.8% improvement in mean J&F over optical-flow-based baselines, with 3.2× faster inference speed and substantially reduced memory consumption and latency—enabling real-time edge deployment. Our core contribution is establishing frame differencing as a viable, efficient motion prior and introducing a new lightweight AVOS paradigm that jointly optimizes accuracy and efficiency.

0 citationsRead paper

Recent publications

Latest Papers

Spherical Geometrical Bases of Spherical Origami

May 01, 2026

This study addresses the absence of a rigorous geometric framework for spherical origami, particularly concerning folding rules on the unit sphere and their realization in three-dimensional space. The work proposes the first complete axiomatic system for spherical origami by extending the classical Huzita–Justin axioms from Euclidean to spherical geometry, providing explicit equations for all seven axioms. Innovatively, it replaces geodesics with isometric curves as three-dimensional creases, thereby expanding the repertoire of realizable folded forms. By integrating spherical and differential geometry with computational origami theory and computer graphics, the authors successfully generate a spherical origami bird model, demonstrating both the theoretical completeness and practical feasibility of the proposed framework.

0 citationsRead paper

Apple Peel Unfolding of Archimedean and Catalan Solids

Apr 17, 2026

This study investigates whether Archimedean solids and their duals, the Catalan solids, can be unfolded into non-overlapping planar nets via a continuous “apple-peeling” path—traversing adjacent faces without gaps or overlaps. We formalize this unfolding paradigm for the first time, introducing rigorous face-selection rules and developing an automated verification algorithm grounded in geometric and topological analysis. Applying this framework, we systematically classify the unfoldability of all Archimedean and Catalan solids: three Archimedean and six Catalan solids admit perfect apple-peeling unfoldings; three Archimedean and three Catalan solids are unfoldable only under restricted conditions; and the remaining solids cannot be unfolded in this manner.

0 citationsRead paper

D-COT: Disciplined Chain-of-Thought Learning for Efficient Reasoning in Small Language Models

Feb 25, 2026

This work addresses the performance degradation and increased computational overhead in small language models (SLMs) caused by “overthinking” during chain-of-thought (CoT) distillation from large language models. To mitigate reasoning drift, the authors propose a structured reasoning framework that introduces control tokens—such as <TEMP_LOW> and <TEMP_HIGH>—to guide SLMs along efficient, ordered reasoning pathways. The approach integrates control-token-driven CoT distillation, structured trajectory optimization, and few-shot fine-tuning. Evaluated on Qwen3-8B with only 5,000 training samples, the method achieves a 9.9% absolute accuracy gain on GPQA-diamond and a 9.1% improvement on zero-shot MMLU-Pro, while substantially reducing token consumption. These results demonstrate the model’s enhanced capacity to internalize structured reasoning patterns.

0 citationsRead paper

Digital Nature Revisited: A Ten-Year Synthesis of Art, Technology, and the Evolution of"Nature": Reimagining Post-Truth Ecologies Through Art, Algorithm, and Animism

Nov 11, 2025

This paper addresses the persistent conceptual ambiguity and interdisciplinary divergence surrounding the term “digital nature” over the past decade. Methodologically, it integrates media art, bio-art, and generative art practices with algorithmic creation, large language models, and philosophical inquiry to systematically map the concept’s genealogy and expose the ethical–political risks embedded in techno-uncanny phenomena. Its contributions are threefold: first, it proposes a novel dual-axis analytical model—spanning virtual/real and anthropocentric/object-oriented dimensions; second, it repositions digital nature as a platform for intercivilizational dialogue; and third, it introduces the concept of “supernatural reality” to extend critical engagement with post-truth ecologies. Collectively, these advances advance art–technology collaboration toward a more culturally grounded and critically reflexive practice paradigm.

0 citationsRead paper

Investigation of Frame Differences as Motion Cues for Video Object Segmentation

Mar 12, 2025

For automatic video object segmentation (AVOS) without first-frame annotations, existing methods rely on computationally expensive optical flow to extract motion cues, hindering real-time deployment on edge devices. This paper proposes replacing optical flow with lightweight frame differencing as a motion prior and designs an extended U-Net architecture for end-to-end segmentation. We systematically demonstrate— for the first time—that frame differencing provides motion cues comparable in effectiveness to optical flow under static-camera conditions, while drastically reducing computational overhead. Experiments on standard AVOS benchmarks show our method achieves a 0.8% improvement in mean J&F over optical-flow-based baselines, with 3.2× faster inference speed and substantially reduced memory consumption and latency—enabling real-time edge deployment. Our core contribution is establishing frame differencing as a viable, efficient motion prior and introducing a new lightweight AVOS paradigm that jointly optimizes accuracy and efficiency.

0 citationsRead paper