OVERVIEW
CAS develops SciFinder®, the most trusted collection of chemical information used by leading pharmaceutical and chemical corporations to keep up to date on and leverage the latest advances in science. Chemisches Zentralblatt is the world’s oldest journal of German chemical abstracts from 1830 to 1969. Only available in German, the compendium was challenging to search. In 2016, CAS, a division of the American Chemical Society, introduced a new solution, ChemZentTM, developed in partnership with Iconic, unlocking 3 million abstracts of valuable, historical scientific information in English.
THE CHALLENGE
It was difficult to explore and identify foundational chemical information in the original German hardback version of Chemisches Zentralblatt. Organisations were also missing access to the complete digital repository of the earliest chemistry history. To solve this, CAS turned to Iconic, tasking us with translation of the mammoth compendium from German into English for indexing and search, in addition to translation and parsing of ongoing chemical information emanating from China, Japan, and Korea.
HOW ICONIC HELPED
Iconic developed innovative machine learning technology to extend its existing Machine Translation (MT) and Natural Language Processing (NLP) solutions for the large-scale digitisation, translation and indexing of 140 years’ worth of German journals and patents. Customising proprietary MT technology for historical chemical information with statistical and neural approaches, Iconic also developed NLP technologies for digitisation of 800,000 PDFs, and extraction and fielding of 3 million abstracts by author and title. Iconic’s expertise together with CAS industry-leading scientific information analysis made the launch of ChemZentTM possible within one year of idea inception. This foundational chemistry, as well as the most up to date information from Asia, is now searchable in English with summaries in English as well.