🧠 Daily Study Log [2025-07-11]
Felt more energized today and wrapped up a key milestone!
Closed the SCU_Competition project, dove into CV Chapter 3, studied EfficientNet Day 3, reviewed AI research tools, and organized initial research notes on Kakao’s platform dynamics.

📊 SCU_Competition — Final Submission Wrap-Up

Focus:

• Completed final round of tuning and documentation
• Reviewed feature importance and ensemble structure one last time
• Archived experiment logs and updated leaderboard results

Result:

• ✅ Best submission remains strong (AUC 0.9002)
• ✅ Project officially closed — ready for final summary post

Reflection:

• Learned the power of feature engineering + ensemble
• Reproducibility and logging helped a lot during iteration

🧠 CV Study — Chapter 3 Summary

Topics Covered:

• Implemented MLP vs. CNN from scratch
• Understood how CNN learns spatial hierarchies
• Compared performance and parameter count

Reflection:

• CNN outperformed MLP significantly
• This chapter gave great hands-on intuition before jumping into more complex models

📘 Paper Study — EfficientNet (Day 3)

Focus:

• Studied EfficientNet-B0 architecture in depth
• Reviewed how the model was designed via Neural Architecture Search (NAS)
• Understood the use of compound scaling coefficients (ϕ, α, β, γ) in building B1–B7 variants

Reflection:

• NAS-generated B0 architecture is surprisingly elegant
• Scaling strategy is both principled and effective
• Looking forward to examining performance benchmarks and ablation results next

🧪 AI Research Tools — FastCampus Course

Focus:

• Followed lessons on using generative AI tools to assist academic writing
• Explored prompt engineering for summarization, paraphrasing, and citation organization

Reflection:

• Tools can speed up literature review and draft generation
• Planning to apply this for my upcoming research paper

🧠 Research Journal — Kakao Platform Study

Topics Organized:

• Summarized key ideas around platform economy and user lock-in
• Explored Kakao’s influence on digital accessibility, market fairness, and user behavior patterns

Reflection:

• If well-developed, this could be a formal research topic
• Might pursue funding or academic support if aligned with program goals

❌ Skipped: SQL Study

Didn’t get around to SQL today — not a priority, but will return to it soon when needed for research or modeling.

✅ TL;DR

📍 SCU: Finalized and wrapped up competition work — ready to move on
📍 CV: Studied Chapter 3 — MLP vs. CNN complete
📍 Paper: EfficientNet Day 3 — B0 architecture & scaling logic deep dive
📍 AI Tools: FastCampus course on academic writing with AI — good prep
📍 Research: Outlined Kakao platform study — may develop into funded project
📍 Tomorrow: Write, scale, and study — a well-rounded day ahead