Jinju Lee

ARPES reveals subtle surface defects invisible to LEED and XPS, demonstrating that a $(1\times1)$ LEED pattern doesn't guarantee momentum-resolved electronic structure.

Chord-symbol adaptation can boost genre-specific harmonic prediction, but it falls short of capturing complete genre identity, revealing a critical gap in music modeling.

Fine-tuning a chord generation model on a new genre requires only a surprisingly small amount of old-genre data to prevent catastrophic forgetting, but objective metrics don't always capture subjective stylistic preferences.