Retrieval of chemical information embedded in a variety of documents is a challenging task due to the complexity of the Japanese language. Japan is one of the world’s leading countries in several research fields, including pharmaceuticals, where a large number of patents and research articles are produced each year. While machine translation is a widely used tool by English speaking professionals, it does have limitations with more specialized fields. In the chemistry and the pharmaceutical fields compounds can be described in Japanese language documents with the combination of katakana, kanji, roman characters and numbers, and consequently machine translation is often insufficient to provide the reliable and accurate retrieval of chemical data. We investigate the bidirectional conversion of chemical and biochemical structures and names in various formats, in addition to IUPAC and traditional name support, also recognizing common names, and CAS registry numbers. We enable indexing documents, facilitating retrieval of the exact location of the chemical data extracted for each document.

Challenges of chemical name recognition in Japanese documents

ACS Fall 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.