5 projects
Work
Two research projects, two tools, one hackathon.
Lead project
Read this one first
CNM-BERT
ResearchUnder review2026
BERT treats 樹 as an arbitrary ID and throws away the components a reader sees instantly. CNM-BERT feeds that structure back in and lifts out-of-vocabulary accuracy by 9.8 F1 for about 5% more training time.
PyTorchBERTTree-MLPIDSCLUE
Read the write-up5 in total
The full index
Everything we have shipped, including the one above.


