ABM-SCI: AI-tool for Training and Research

ABM-SCI is the first Artificial Boothmate designed for use in the Interpreting Classroom and for Research purposes.

An Artificial Boothmate is a computer-assisted tool based on Machine Learning that suggests interpreters in real-time terms, translations, and numbers in context.

The particularity of this tool is that it allows the trainer/researcher to create a session, both by uploading a curated glossary in EN+several target languages or by simply indicating the topic of the speech (full automation) and share the session with the students/probands. The trainer/researcher can read aloud a speech on her device, or connect an audio/video source, and all the students/probands can access in real-time the Artificial Boothmate from their own computers

The tool runs on the cloud, so students/probands do not need to install any tool. They simply receive a link from the trainer/researcher and open it in their Internet Browser.

The tool works with English as a source language (Spanish to be added soon). The tool gives the following suggestions:

  • from the glossary that the trainer/researcher has curated and uploded, or
  • the tool creates automatically an ontology on the topic indicated at the beginning of the session.
  • numbers/dates/etc. with (some) context
  • conversion of measurement (for example miles to km)
  • a consecutive pad integrating automatic speech recognition and machine translation

The easiest way to use this tool in your class or for your experiments is through InterpretBank. A standalone version is in preparation.

If you use ABM-SCI in your publications, please cite:

Fantinuoli C. "Speech Recognition in the Interpreter Workstation". Proceedings of the Translating and the Computer 39 (2017), 25-34.


Declaration of financial support: The author(s) received no financial support for the research, authorship, and/or publication of this project.