Supporting drug discovery, the determination of bioequivalent substances and the identification of new antibiotics are the objectives of the ABRoad project. This project is the result of a collaboration between InflamAlps, a pharmaceutical R&D company based in Monthey (canton of Valais), and the Idiap Research Institute.
The identification and selection of potential sources of antibiotic substances, which can then be validated experimentally, requires the interpretation of scientific literature on a large scale. Considering the size of the related scientific corpus, the task can be daunting. Thanks to recent advances in natural language processing, it is now possible to automate important parts of this sources of antibiotics selection process.
The ABRoad project’s main objective will be to design a natural language processing (NLP) software infrastructure. This infrastructure will then help biomedical data discovery by easing the analysis of large scientific textual databases, such as articles and patents, as well as the development of a model allowing to compare chemical formulas within their textual context. The project is expected to start in the fall of 2022.
A transposable technology for more cost-effective pharmaceutical research
By using cutting-edge methods developed by Idiap specifically in Natural Language Inference (NLI), the project aims to develop an advanced textual interpretation platform. To effectively support biomedical discovery, the project needs to take into account inferences in the contents it will analyze.
“In the past few years, these methods have dramatically evolved to support the interpretation of textual evidence at scale. With ABRoad, we will demonstrate their value in augmenting the antibiotic discovery process,” explained André Freitas, Head of the Reasoning & Explainable AI research group at Idiap.
The software infrastructure developed for the ABRoad project is a real proof-of-concept of the application of contemporary NLP methods and will give a strategic boost to biomedical companies in Valais, and beyond. The project also confirms the positioning of the canton of Valais as a national focal point in the area of Natural Language Processing.
