Institution profile

Michigan Technological University

Academic institutionnorthamerica · us
Official website
Research library97linked papers
Opportunities0open roles
Selected work

Representative Papers

Boxplots and quartile plots for grouped and periodic angular data

Feb 05, 2026

This study addresses the lack of effective visualization methods for grouped periodic angular data in fields such as psychology, genomics, and meteorology. The authors propose concentric circular boxplots and circular quartile plots to characterize grouped angular distributions, and further extend the approach to a three-dimensional toroidal visualization for multiple groups to reveal periodic patterns. A novel scaling strategy is introduced, wherein box width is inversely proportional to the square root of the distance from the center, enhancing visual perception. This work is the first to integrate concentric circular boxplots with toroidal 3D visualization specifically for angular data. The effectiveness and practical utility of the proposed methods are demonstrated through applications to real-world datasets, including motor resonance phases, circadian clock gene expression phases, and periodic wind direction patterns.

1 citationsRead paper

Latent characterisation of the complete BATSE gamma ray bursts catalogue using Gaussian mixture of factor analysers and model-estimated overlap-based syncytial clustering

Nov 12, 2024Monthly notices of the Royal Astronomical Society

This study addresses the ongoing debate regarding the optimal number of gamma-ray burst (GRB) classes—ranging from two to five—by analyzing 1,150 GRBs from the BATSE catalog using nine observational parameters. For the first time, it integrates a mixture of Gaussian factor analyzers with the Model-based Overlapping Synchronous Clustering (MOBSynC) method to identify five ellipsoidal clusters, which are subsequently hierarchically merged via MOBSynC into three and then two classes. The proposed framework reveals GRB intrinsic structure through a small set of latent factors, clearly characterizing distinct subpopulations such as “short–faint–hard” and “long–bright–intermediate.” This approach provides a unified explanation for the validity of classifications at multiple granularities and introduces the first latent-variable-driven, hierarchical model for GRB structural analysis.

1 citationsRead paper

DART: A DAG-Based Reputation and Incentive Framework via Blockchain-Enabled Governance for Trustworthy LLM Multi-Agent Collaboration

Sep 01, 2026

Large language model (LLM)-based multi-agent systems (MAS) predominantly rely on centralized orchestration and lack formal verification mechanisms for agent reliability, participation, and system-level behavioral alignment. These shortcomings leave open environments severely vulnerable to uncooperative or malicious agents. This work proposes DART, a Directed Acyclic Graph (DAG)-based reputation and incentive regulation framework for trustworthy multi-agent collaboration, combining centralized operational orchestration with blockchain-enabled decentralized governance and accountability. DART unifies DAG workflow orchestration, capability and reputation-aware task allocation, dynamic behavior updates, multi-factor incentives, and smart contract accountability paired with IPFS storage. Under this paradigm, agent selection dynamically balances task alignment, historical reputation, and workload, while post-execution behavioral evidence continuously calibrates agent trust and the probability of future participation. Evaluated across four axes, DART achieves 93.6% Pass@1 on GSM8K and builds a full-stack application in 142 s using two agents, outperforming centralized baselines. Across five independent 150-round longitudinal trials, Full DART achieves a mean task success rate of 93.33 +/- 2.26%, output quality of 0.9357 +/- 0.0117, retry rate of 0.2307 +/- 0.0816, and allocation delay of 1.1153 +/- 0.0408 s, consistently outperforming its ablated configurations DART isolates persistent and intermittent malicious agents, obtaining a 99.3% output containment rate and restoring system success to 99.8%. These results demonstrate the potential of coupling reputation, incentives, DAG-based coordination, and verifiable blockchain-enabled governance to support adaptive and accountable multi-agent collaboration.

0 citationsRead paper
Recent publications

Latest Papers

DART: A DAG-Based Reputation and Incentive Framework via Blockchain-Enabled Governance for Trustworthy LLM Multi-Agent Collaboration

Sep 01, 2026

Large language model (LLM)-based multi-agent systems (MAS) predominantly rely on centralized orchestration and lack formal verification mechanisms for agent reliability, participation, and system-level behavioral alignment. These shortcomings leave open environments severely vulnerable to uncooperative or malicious agents. This work proposes DART, a Directed Acyclic Graph (DAG)-based reputation and incentive regulation framework for trustworthy multi-agent collaboration, combining centralized operational orchestration with blockchain-enabled decentralized governance and accountability. DART unifies DAG workflow orchestration, capability and reputation-aware task allocation, dynamic behavior updates, multi-factor incentives, and smart contract accountability paired with IPFS storage. Under this paradigm, agent selection dynamically balances task alignment, historical reputation, and workload, while post-execution behavioral evidence continuously calibrates agent trust and the probability of future participation. Evaluated across four axes, DART achieves 93.6% Pass@1 on GSM8K and builds a full-stack application in 142 s using two agents, outperforming centralized baselines. Across five independent 150-round longitudinal trials, Full DART achieves a mean task success rate of 93.33 +/- 2.26%, output quality of 0.9357 +/- 0.0117, retry rate of 0.2307 +/- 0.0816, and allocation delay of 1.1153 +/- 0.0408 s, consistently outperforming its ablated configurations DART isolates persistent and intermittent malicious agents, obtaining a 99.3% output containment rate and restoring system success to 99.8%. These results demonstrate the potential of coupling reputation, incentives, DAG-based coordination, and verifiable blockchain-enabled governance to support adaptive and accountable multi-agent collaboration.

0 citationsRead paper