Writing
These are the essays on the site. Maps of Osaka is a history of the city in 14 sections, from the 7th century to the present, based on maps in the Osaka Municipal Library Digital Archive. The other two describe machine learning tests on patent documents. One fine-tunes DeBERTa to detect antecedent basis errors in patent claims, which reached 90% F1 on synthetic data and 8% recall on real data. The other runs 12 tests of DeepSeek-OCR, GLM-OCR and rotation heuristics for reading text and reference numbers on patent figure sheets.
OCR on Patent Figures with DeepSeek-OCR
12 tests extracting text and reference numbers from patent figure sheets using DeepSeek-OCR, GLM-OCR, and rotation heuristics.
Failing to Train DeBERTa to Detect Patent Antecedent Basis Errors
Fine-tuned DeBERTa to detect antecedent basis errors in patent claims. It reached 90% F1 on synthetic data and 8% recall on real data.
Maps of Osaka
Osaka in the maps its own library keeps, in order: a coastline nobody had seen, a castle buried under another castle, a map that took four years to cut after it was licensed, and the canals that are now an elevated road.