Selected work

Representative Papers

A Browser-Based Gesture-Driven Avatar Interaction Framework for Metaverse Onboarding Environments

Aug 12, 2026

This work addresses the limitations of existing metaverse systems, which predominantly rely on handheld controllers or keyboards and thus struggle to deliver natural, controller-free immersive interaction within web browsers. The paper proposes a lightweight, purely web-deployable, controller-free interaction framework that, for the first time, integrates real-time hand and head gesture recognition based on MediaPipe with two controller-free locomotion techniques—arm-swinging navigation and in-place walking. This integration enables users to drive their avatars and interact naturally with multimedia content, documents, and quizzes through intuitive body movements. A structured evaluation with five participants demonstrates the system’s feasibility in standard browser environments and reveals critical design trade-offs between interaction accuracy and sense of immersion across different input modalities.

0 citationsRead paper

Reward-Density Heuristic for Dynamic Multi-Vehicle Routing: Performance and Computational Efficiency

Jul 07, 2026

This work addresses the challenge of dynamic multi-vehicle routing with the dual objectives of maximizing cumulative rewards within fixed time windows and responding promptly to newly arriving tasks. To this end, the authors propose an efficiency-driven greedy heuristic based on reward density. Designed for continuous replanning scenarios, the method achieves solution quality comparable to state-of-the-art metaheuristics—such as Adaptive Large Neighborhood Search (ALNS), genetic algorithms, and simulated annealing—while incurring negligible computational overhead. Empirical evaluations in unmanned aerial vehicle task assignment and taxi dispatching demonstrate that the proposed approach maintains near-optimal reward performance while reducing computation time by two to three orders of magnitude, thereby achieving a Pareto-efficient trade-off between solution quality and computational efficiency.

0 citationsRead paper

Efficient Perception in Automotive Detection and Tracking Using Neuromorphic Computing

Jul 06, 2026

This work addresses the high energy consumption and deployment challenges of conventional deep learning models in autonomous driving perception by proposing an efficient multi-object detection and tracking method based on spiking neural networks (SNNs) and neuromorphic computing. Leveraging transfer learning and a novel SpikeYOLO architecture, the approach achieves mAP scores of 0.937 and 0.771 and HOTA metrics of 0.701 and 0.445 on the KITTI and BDD100K MOT2020 datasets, respectively—performance comparable to traditional models while significantly reducing energy usage. This study presents the first comprehensive evaluation of SNNs for multi-object detection and tracking in real-world autonomous driving scenarios, demonstrating their potential for jointly optimizing accuracy and energy efficiency.

0 citationsRead paper

CALMem : Application-Layer Dual Memory for Conversational AI

May 20, 2026

Large language models are constrained by fixed context windows, leading to irreversible loss of dialogue history and degraded long-range coherence. This work proposes a purely application-layer dual-memory architecture that achieves near-infinite effective context without modifying the underlying model or introducing vendor lock-in. The architecture integrates episodic memory—based on sliding-window vector embeddings—with agent-writable structured semantic memory, and introduces, for the first time, an intra-conversational retrieval mechanism that enables efficient access to compressed dialogue history. A novel MOIM (Memory-Oriented Information Merging) adaptive injection mechanism dynamically incorporates relevant contextual information. The system supports zero-overhead fallback to native LLM behavior and significantly enhances contextual coherence in extended conversations.

0 citationsRead paper

Universal statistical laws governing culinary design

Apr 30, 2026

This study investigates whether global culinary recipes adhere to universal statistical regularities akin to those observed in natural languages. By constructing a large-scale corpus of traditional recipes spanning multiple cuisines and leveraging state-of-the-art named entity recognition to parse them into structured components—such as ingredients and cooking techniques—the authors apply methods from statistical linguistics to analyze their distributional properties. They reveal for the first time that recipe systems universally obey Zipf’s law, Heaps’ law, and the Menzerath–Altmann law, with macronutrient concentrations following a log-normal distribution. Furthermore, they propose a parsimonious generative model based on preference-driven reuse, constrained sampling, and incremental modification, which successfully reproduces these empirical patterns, offering a unified framework for understanding the combinatorial structure and evolutionary dynamics of recipes across cultures.

0 citationsRead paper

Recent publications

Latest Papers

A Browser-Based Gesture-Driven Avatar Interaction Framework for Metaverse Onboarding Environments

Aug 12, 2026

This work addresses the limitations of existing metaverse systems, which predominantly rely on handheld controllers or keyboards and thus struggle to deliver natural, controller-free immersive interaction within web browsers. The paper proposes a lightweight, purely web-deployable, controller-free interaction framework that, for the first time, integrates real-time hand and head gesture recognition based on MediaPipe with two controller-free locomotion techniques—arm-swinging navigation and in-place walking. This integration enables users to drive their avatars and interact naturally with multimedia content, documents, and quizzes through intuitive body movements. A structured evaluation with five participants demonstrates the system’s feasibility in standard browser environments and reveals critical design trade-offs between interaction accuracy and sense of immersion across different input modalities.

0 citationsRead paper

Reward-Density Heuristic for Dynamic Multi-Vehicle Routing: Performance and Computational Efficiency

Jul 07, 2026

This work addresses the challenge of dynamic multi-vehicle routing with the dual objectives of maximizing cumulative rewards within fixed time windows and responding promptly to newly arriving tasks. To this end, the authors propose an efficiency-driven greedy heuristic based on reward density. Designed for continuous replanning scenarios, the method achieves solution quality comparable to state-of-the-art metaheuristics—such as Adaptive Large Neighborhood Search (ALNS), genetic algorithms, and simulated annealing—while incurring negligible computational overhead. Empirical evaluations in unmanned aerial vehicle task assignment and taxi dispatching demonstrate that the proposed approach maintains near-optimal reward performance while reducing computation time by two to three orders of magnitude, thereby achieving a Pareto-efficient trade-off between solution quality and computational efficiency.

0 citationsRead paper

Efficient Perception in Automotive Detection and Tracking Using Neuromorphic Computing

Jul 06, 2026

This work addresses the high energy consumption and deployment challenges of conventional deep learning models in autonomous driving perception by proposing an efficient multi-object detection and tracking method based on spiking neural networks (SNNs) and neuromorphic computing. Leveraging transfer learning and a novel SpikeYOLO architecture, the approach achieves mAP scores of 0.937 and 0.771 and HOTA metrics of 0.701 and 0.445 on the KITTI and BDD100K MOT2020 datasets, respectively—performance comparable to traditional models while significantly reducing energy usage. This study presents the first comprehensive evaluation of SNNs for multi-object detection and tracking in real-world autonomous driving scenarios, demonstrating their potential for jointly optimizing accuracy and energy efficiency.

0 citationsRead paper

CALMem : Application-Layer Dual Memory for Conversational AI

May 20, 2026

Large language models are constrained by fixed context windows, leading to irreversible loss of dialogue history and degraded long-range coherence. This work proposes a purely application-layer dual-memory architecture that achieves near-infinite effective context without modifying the underlying model or introducing vendor lock-in. The architecture integrates episodic memory—based on sliding-window vector embeddings—with agent-writable structured semantic memory, and introduces, for the first time, an intra-conversational retrieval mechanism that enables efficient access to compressed dialogue history. A novel MOIM (Memory-Oriented Information Merging) adaptive injection mechanism dynamically incorporates relevant contextual information. The system supports zero-overhead fallback to native LLM behavior and significantly enhances contextual coherence in extended conversations.

0 citationsRead paper

Universal statistical laws governing culinary design

Apr 30, 2026

This study investigates whether global culinary recipes adhere to universal statistical regularities akin to those observed in natural languages. By constructing a large-scale corpus of traditional recipes spanning multiple cuisines and leveraging state-of-the-art named entity recognition to parse them into structured components—such as ingredients and cooking techniques—the authors apply methods from statistical linguistics to analyze their distributional properties. They reveal for the first time that recipe systems universally obey Zipf’s law, Heaps’ law, and the Menzerath–Altmann law, with macronutrient concentrations following a log-normal distribution. Furthermore, they propose a parsimonious generative model based on preference-driven reuse, constrained sampling, and incremental modification, which successfully reproduces these empirical patterns, offering a unified framework for understanding the combinatorial structure and evolutionary dynamics of recipes across cultures.

0 citationsRead paper