Hello, I'm JaeWoo Choi, majoring in AI Big Data & Management
(Convergence Double Major: Data Science) at Kookmin University.
My research interests focus on AI model design and adaptation for structured and unstructured data. I am particularly interested in integrating data analytics, large language models (LLMs), computer vision, and industrial AI to address real-world prediction and decision-support problems.
I aim to study how AI models can be adapted and optimized for diverse real-world tasks, with an emphasis on practicality, efficiency, and reliability.
- Efficient & Transferable AI: model compression, transfer learning, parameter-efficient adaptation, reliable model adaptation
- LLM & AI Agents: large language models, tool-using agents, workflow automation, decision-support systems
- Computer Vision: visual representation learning, image/video understanding, vision foundation model adaptation
- Industrial AI & Data Analytics: structured/time-series data analysis, anomaly detection, forecasting
-
๐ฅ ํ๋ ฅ๋ฐ์ ๋ ์์ธก AI ๊ฒฝ์ง๋ํ BARAM 2026 โ 1์(ํ๊ตญ๋์๋ฐ์ ์ฌ์ฅ์)
๐ฐ Article
LDAPSยทGFS ์์น์๋ณด์ ๋ฐ์ ๋ฐ์ดํฐ๋ฅผ ํ์ฉํ ์ต์ผ ํ๋ ฅ๋ฐ์ ๋ ์์ธก ๋ชจ๋ธ Windsight๋ฅผ ๊ฐ๋ฐํ์ต๋๋ค. MOS ๊ธฐ์ ๋ณด์ , Quantile LightGBMยทAnalog Ensemble, ํ๊ฐ ์งํ ๊ธฐ๋ฐ ์์ฌ๊ฒฐ์ ์ ๊ฒฐํฉํด ์์ธก ์ ํ๋์ ์ ์ฐ ์์ต์ ํจ๊ป ๊ณ ๋ คํ ๋ฐ์ ์ ์ด์ ์ต์ ํ๋ฅผ ๊ตฌํํ์ต๋๋ค. -
๐ฅ JB๊ธ์ต๊ทธ๋ฃน Fin:AI Challenge AI agent ์๋น์ค ๊ฐ๋ฐ โ 1์(JB๊ธ์ต๊ทธ๋ฃน ํ์ฅ์)
๐FinProofยท ๐ฐ Article
pgvector ๊ธฐ๋ฐ RAG, ๋๋ฉ์ธ๋ณ Multi-Agent ๋ถ์, RAG + CoVe ๊ทผ๊ฑฐ ์ฌ๊ฒ์ฆ์ ๊ฒฐํฉํด ๊ธ์ต ๊ด๊ณ ์ ์๋ฐ ๊ฐ๋ฅ์ฑ, ํ๋จ ๊ทผ๊ฑฐ, ์ํ๋, ์์ ๊ถ๊ณ ์์ ์์ฑํ๋ Compliance AI Agent๋ฅผ ๊ตฌ์ถํ์ต๋๋ค. ์จํจ๋ก์ง ๊ธฐ๋ฐ ์ฌํ๋งฅ๋ฝ ์ง์๊ทธ๋ํ์ Cross-lingual NLI๋ฅผ ํ์ฉํด ๊ตญ๊ฐ๋ณ ์ฌํ๋งฅ๋ฝ ๋ฆฌ์คํฌ์ ๋ค๊ตญ์ด ์ค๋ฒ ์๋ฏธ ์๊ณก์ ํ์งํ๋ Human-in-the-loop ์ฌ์ ์ํฌํ๋ก์ฐ๋ฅผ ๊ตฌํํ์ต๋๋ค. -
๐ฅ Hecto AI Challenge: ๋ฅํ์ดํฌ ํ์ง โ 1์(๋์)
๐dinov3-vit-lora-deepfake-detectionยท ๐ฐ Article
DINOv3 ViT-L/16 ๊ธฐ๋ฐ์ visual foundation model์ LoRA adapter์ task-specific classification head๋ก ์ ์์์ผ ์ด๋ฏธ์งยท๋น๋์ค ๊ธฐ๋ฐ ๋ฅํ์ดํฌ ํ์ง ํ์ดํ๋ผ์ธ์ ๊ตฌ์ถํ์ต๋๋ค. -
๐ฅ ์ค๋งํธ ํด์ด๋ฌผ๋ฅ ร AI ๋ฏธ์ ์ฑ๋ฆฐ์ง: AIS ๋ฐ์ดํฐ๋ฅผ ํ์ฉํ ์ ๋ฐ ํญ๋ก ๋ณต์ ์์ธก โ 2์(์ธ์ฐํญ๋ง๊ณต์ฌ์ฌ์ฅ์)
๐ais-route-reconstruction
Kalman filtering, RTS smoothing, temporal modeling์ ๊ฒฐํฉํ์ฌ ๋๋ฝ๋ AIS ์๊ณ์ด ๊ตฌ๊ฐ์ ๋ณต์ํ๊ณ , ์ฐ์ ๋ฐ์ดํฐ ๊ธฐ๋ฐ trajectory reconstruction ํ์ดํ๋ผ์ธ์ ๊ตฌ์ถํ์ต๋๋ค. -
๐ฅ ์ค๋งํธ ํด์ด๋ฌผ๋ฅ ร AI ๋ฏธ์ ์ฑ๋ฆฐ์ง: ์ด์์ ํธ ๊ฐ์ง ๊ธฐ๋ฐ ๋น์ ์ ์๋ ์ง๋จ โ 2์
๐maritime-anomaly-signal-diagnosis
๋น์๋ณ ์ผ์ ๋ฐ์ดํฐ๋ฅผ ๋์์ผ๋ก feature engineering, distance-based representation, LDA/KMeans-derived features๋ฅผ ๊ฒฐํฉํ์ฌ ์ฐ์ ๋ฐ์ดํฐ ๊ธฐ๋ฐ ๋ค์ค ํด๋์ค ์ด์์ง๋จ ๋ชจ๋ธ์ ์ค๊ณํ์ต๋๋ค. -
๐ฅ LG Aimers 7th AI Hackathon: ์์์ ์ฅ ๋ฉ๋ด ์์์์ธก โ 3์(ํ๊ฒฝ๋ท์ปด ์ฌ์ฅ์)
๐lg-aimers-7thยท ๐ฐ Article
๋ฆฌ์กฐํธ ๋ด ์์์ ์ฅ๋ณ ์์น, ์ด์ฉ ๋์ , ๋ฐฉ๋ฌธ๊ฐ ํ๋ฆ, ์์ฝยท๊ฐ์คยท๋ ์จ ๋ฑ ์ด์ ๋ณ์๋ฅผ ๋ฐ์ํ์ฌ ์ ํยท์๊ณ์ด ๋ฐ์ดํฐ ๊ธฐ๋ฐ ์์์์ธก ํ์ดํ๋ผ์ธ์ ๊ตฌ์ถํ์ต๋๋ค. -
๐ฅ ๊ตญ๋ฏผ๋ํ๊ต ๋ฐ์ดํฐ ๋ถ์ ์์ด๋์ด ์ฝ๋ผํค โ 1์(์ต์ฐ์์)
๐solpick-solar-roi-bidding-platform
ํ์๊ด ๋ฐ์ ๋ ์์ธก, ROI ๋ถ์, ์ง์๊ธ ๋งค์นญ, ์๊ณต์ฌ ์ถ์ฒ์ ๊ฒฐํฉํ ์์ฌ๊ฒฐ์ ์ง์ํ AI ์๋น์ค ํ๋กํ ํ์ ์ ๊ธฐํยท๊ตฌํํ์ต๋๋ค.