Skip to content
View jseongnam's full-sized avatar
1 follower · 0 following

Achievements

Achievement: Pull Sharkx2

Achievements

Achievement: Pull Sharkx2

Block or report jseongnam

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.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. 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
jseongnam/README.md

Seokjun Jeong

Hybrid Neural-Numerical AI Systems

**AI Research Engineer | Scientific Machine Learning | Neural Numerical Methods**

I design hybrid neural-numerical AI systems that combine mathematical representations, neural prediction/correction models, and classical numerical validation to solve nonlinear scientific and engineering problems reliably.

Research Identity

My research focuses on hybrid neural-numerical AI systems.

Rather than replacing classical numerical solvers with neural networks, I use neural models as:

  • interval localization modules
  • root/correction predictors
  • warm-start generators
  • residual-validated decision components

The final goal is to improve prediction, initialization, correction, and computational efficiency while preserving mathematical reliability.

Selected Publications

A Root Prediction System for Single-Variable Equations with Existing Taylor Polynomials

IEEE Access, Early Access, 2026
DOI: 10.1109/ACCESS.2026.3697368
Code: taylor-root-prediction

Hybrid Deep Learning and Newton Refinement: A Baseline-Aware Correction Framework for Nonlinear Pipe-Flow Equations

SCI manuscript under review

Representative Project

Taylor Root Prediction

Official implementation of the IEEE Access paper on Taylor coefficient-based neural root prediction.

Core components:

  • Transformer-based interval localization
  • 25th-order local Taylor representation
  • coefficient-based neural root regression
  • multi-candidate prediction
  • residual/domain/stability validation
  • baseline comparison and failure analysis

Repository: github.com/jseongnam/taylor-root-prediction

Technical Stack

Research Keywords

Scientific ML · Neural Numerical Methods · Root Finding · Taylor Series · Residual Validation · Newton Refinement · Computer Vision · Action Recognition

Contact

Email: wjdtjrwns1109@gmail.com
GitHub: jseongnam

Pinned Loading

  1. taylor-root-prediction taylor-root-prediction Public

    Official IEEE Access implementation for Taylor coefficient-based neural root prediction with Transformer interval localization and residual validation.

    Jupyter Notebook

  2. cai-Colebrook cai-Colebrook Public

    Hybrid neural-numerical warm-start correction framework for Colebrook-White pipe-flow equations with Newton refinement.

    Python

  3. single-variable-root-prediction single-variable-root-prediction Public

    Neural root prediction pipeline for single-variable synthetic equations with branch-aware learning and residual validation.

    Python