Cassyni, the platform for organising, running and publishing research seminars, is launching a next-generation AI video product to enrich and extract knowledge from thousands of hours of research videos. Cassyni’s technology automatically transforms unstructured research videos into structured searchable content with navigable chapters, high-quality transcripts, extracted slides from speaker presentations and resolved references.
Andrew Preston, Cassyni Co-founder said:
“Researcher’s don’t just read PDFs from start to finish, they jump through sections, search for keywords, follow references, and examine figures. Right now, the only way to consume research video is to watch it linearly from start to finish or seek through randomly to find the right slide.
“Imagine being able to instantly search full-text through every slide and every word of hundreds of thousands of hours of high-quality research video and recorded academic discussions. To be able to go back and immediately find the right slide from the 90 minute seminar you attended. To be able to click on a reference embedded within the video and be linked directly to the publication. Our technology enables all of this, fully integrating video into the academic knowledge ecosystem.”
As well as enriching the seminars contained within Cassyni’s library, the technology can be applied more widely.
Ben Kaube, Cassyni Co-founder said:
“Many research organisations, institutions and publishers are sitting on a trove of high-quality research videos including seminars and other recorded meetings of tremendous value to the research community. Cassyni’s technology makes this content discoverable and automatically links it to the wider literature.”
As part of a pilot with Springer Nature Mathematics, Cassyni will be indexing and enriching video content from a range of meetings including B Waves, Logica Universalis and the International Conference on Optimization and Decision Science.
Marc Strauss, Publishing Director at Springer Nature said:
“Cassyni’s technology helps us unlock the value contained in existing video content recorded at conferences and other virtual events. Cassyni’s indexing, reference extraction and linking makes it much easier and more engaging for researchers to find our content and dive deeper into cutting-edge research topics.”
