Skip to content

Instantly share code, notes, and snippets.

View sjchoi86's full-sized avatar
🇰🇷
Making robots smarter

Sungjoon sjchoi86

🇰🇷
Making robots smarter
View GitHub Profile
@sjchoi86
sjchoi86 / presentation guideline.md
Created August 24, 2026 02:34
Presentation guideline

Presentation guideline

이 문서는 학회 키노트, 회사 기술 발표, 논문 발표와 연구 소개 자료를 만들 때 사용하는 canonical guideline이다. 목표는 발표의 주제만 전달받아도 지금까지 사용한 발표 문법, 정보 밀도, 시각적 스타일과 기술적 엄밀성을 일관되게 재현하는 것이다.

이 문서의 원칙은 다음 세 실제 자료에서 확인한 발표 방식을 반영한다.

  • keynote@ur.pdf: 큰 질문에서 시작해 연구 사례를 거쳐 system-level thesis로 수렴하는 학회 키노트
  • papers@robotis.pdf: 외부 논문의 capability progression을 설명한 뒤 우리 연구로 연결하는 회사 기술 발표
  • gate review (final).pdf: 원래 과업 시나리오와 KPI에서 출발해 낮은 abstraction의 구현 detail, subsystem evidence, 핵심 모델 비교와 hardware-conditioned inference evidence를 거친 뒤 system-level 판단과 다음 Gate로 돌아오는 과제 중간 발표
@sjchoi86
sjchoi86 / PAPER_WORKFLOW.md
Last active August 1, 2026 06:04
Paper workflow

Paper Collection and Optional Slack Publishing Workflow

Purpose

This document is the source of truth for collecting robotics and AI research materials and, only when explicitly requested, publishing them to Slack. A Codex agent with access to this file should be able to complete the workflow from paper links or titles without relying on prior conversation history.

The default deliverables are local artifacts only:

  1. A dated local archive containing PDFs, presentation-compatible MP4 videos, cropped key figures, and one consolidated Markdown review.
  2. A completion report containing the local paths, item counts, and unavailable artifacts.
@sjchoi86
sjchoi86 / 02_gaussian_random_path.ipynb
Created July 28, 2026 11:41
Gaussian random path notebook
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@sjchoi86
sjchoi86 / SOUL.md
Created May 28, 2026 02:49
ri_motion_v5 coding style and notebook philosophy

SOUL.md

This document defines the coding style and notebook philosophy for ri_motion_v5, a robotics research codebase maintained by Sungjoon Choi at the Robot Intelligence Lab, Korea University.

The workspace supports robot-intelligence experiments that combine MuJoCo simulation, robot and object assets, kinematics, inverse kinematics, forward dynamics, contact-rich manipulation, motion capture and retargeting, trajectory sampling and optimization, probabilistic algorithms, camera and

Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@sjchoi86
sjchoi86 / 00_check_cpu_cuda_mps.ipynb
Created February 26, 2026 16:52
foundation_model_notebooks
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
class SitePositionIKSolver(object):
"""
Inverse Kinematics (IK) solver for site position targets using Damped Least Squares (DLS) method.
This class provides methods to set up IK targets, configure solver parameters, and compute joint updates to achieve desired site positions.
Usage:
------
# Create IK solver (use all rev joints by default)
ik_solver = SitePositionIKSolver(
@sjchoi86
sjchoi86 / g1_contact_handling_explained.ipynb
Created October 20, 2025 10:49
Uploaded from jupyterlab
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
@sjchoi86
sjchoi86 / g1_contact_model_explained.ipynb
Created October 20, 2025 10:49
Uploaded from jupyterlab
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.