Speaker: Jérôme Massot, Total R&D (working with Google and Stanford AI Lab)
Date: February 16th, 2021
Time: 12pm-1pm Pacific
Location: Zoom - This is a free webinar, registration required.
Abstract: The number of academic and institutional articles published each year on the subject of Carbon Capture, Storage and Utilization (CCUS) is growing exponentially. If one searches on the words “Carbon storage” in Science-Direct, the Elsevier’s database Search Engine, more than 15,000 references are returned, with 500 of them published so far in 2021. It would be impossible for a scientist or researcher to read everything. This large corpus of available papers needs to be analyzed, curated and classified by sub-topic. The shared knowledge has to be extracted and summarized. And finally, only a fraction of articles need to be read to deepen technical understanding. To achieve these objectives, Neural Language Processing (NLP) can help. However, NLP requires a lot of data to be trained efficiently.
This webinar will show how the most recent NLP techniques associated to Knowledge Graph technology can bring benefits to the CCUS research community. The internal engine of each technique will be explained concisely and, as there is not good AI without good data, the data needed to obtain the best performances will be detailed for each. To conclude the webinar, several demos of existing knowledge extraction platforms will be shared in order to trigger an open discussion about which technology seems the more interesting for serving the CCUS researcher community.