Skip to content
View RezaSangari's full-sized avatar

Block or report RezaSangari

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
RezaSangari/README.md

Reza Sangari Abiz

AI Researcher working on Scientific Machine Learning, Physics-Informed Neural Networks (PINNs), Soft Computing, and Metaheuristic Optimization.

M.Sc. in Soft Computing & Artificial Intelligence — Damghan University (2026, Excellent, GPA 17.49/20).

B.Sc. in Computer Science — Birjand University of Technology.

Research Interests

  • Scientific Machine Learning and Physics-Informed Neural Networks (PINNs)
  • Numerical methods for nonlinear partial differential equations
  • Metaheuristic optimization: Moth-Flame Optimization (MFO), Butterfly Optimization Algorithm (BOA), and hybrid optimization
  • AI applications in computer vision and interdisciplinary research

Featured Research

1. Hybrid MFO–BOA for Nonlinear PDEs

M.Sc. Thesis — Damghan University (2026)

Solving Nonlinear Partial Differential Equations Using the Hybrid Moth-Flame and Butterfly Metaheuristic Algorithm and Its Applications in Engineering Problems.

The hybrid optimizer outperformed standalone MFO by 62% and BOA by 69% in Burgers' equation RMSE. Wilcoxon tests: p = 0.003 versus MFO and p = 0.001 versus BOA. Mean RMSE (Burgers' equation, 30 runs): Hybrid MFO–BOA 2.8e-4; standalone MFO 7.3e-4; standalone BOA 9.1e-4.

Reproducibility note: These values are reported from the M.Sc. thesis and have not been reproduced by the current public implementation — its Burgers least-squares floor is 6.801e-02 with the default 5x5 basis (about 243× the reported hybrid RMSE of 2.8e-4), and 1.52e-02 at the largest tested 144-coefficient basis. Full details: Reproducibility and experimental provenance.

Code: MFO-BOA-PDE-Solver

2. PINNs for Stiff Reaction–Diffusion Systems

Current work on physics-informed neural networks tuned using a hybrid MFO–BOA optimizer.

Manuscript: In preparation for MedLife 2027, First International Interdisciplinary Sciences Congress, Tehran.

Code: repository in preparation — PINN implementation to follow.

3. Persian Lip-Reading

Dataset: LRW-Persian — 743 words and 414,278 clips.

Best run so far: 35.31% TEST Top-1 (paper record: 41.09%). The paper's model was fine-tuned from a pretrained model, while this run trains from random initialization.

Dataset link: LRW-Persian on Hugging Face

4. AI and Nanotechnology

Artificial Intelligence and Machine Learning in Nanotechnology.

NCWNN7 review article — Submitted October 2026.

5. Generative AI and Political Cognition

Generative AI and Political Cognition of Iranian Users: Political Trust, Critical Thinking, and News Resharing.

Manuscript in preparation.

GitHub Projects

Professional Membership

Member of IAENG — Societies of Artificial Intelligence and Computer Science.

Member ID: 588946.

Contact

For research collaborations, research assistant opportunities, and PhD inquiries, please contact me through my professional email or LinkedIn.

Links

Pinned Loading

  1. MFO-BOA-PDE-Solver MFO-BOA-PDE-Solver Public

    Hybrid Moth-Flame Optimization (MFO) + Butterfly Optimization Algorithm (BOA) with B-spline collocation for nonlinear PDEs. M.Sc. thesis implementation, Damghan University (2026).

    Python

  2. RezaSangari.github.io RezaSangari.github.io Public

    Personal profile website of Reza Sangari Abiz — AI Researcher

    HTML