This week is #MuseumWeek and as today’s focus is on Artificial Intelligence I would like to share some of the work we have been doing at Yad Vashem, where by using state-of-the-art NLP and generative AI we are bringing victim memories and their stories once lost, back to life. The primary project we’ve been working on is ensuring that the murdered who until now have been forgotten can at long last be remembered.
We are parsing through the tens of thousands of testimonies Yad Vashem has in its archives and then using advanced NER models to collect names, details, and places mentioned, often in passing, during survivor testimonies.
Despite the myriad of disparate languages and diverse media types of these Holocaust testimonies, we have extracted names and information about them with the goal of adding them to Yad Vashem’s Central Database of Holocaust Victims' Names and stopping them from being forgotten.
For those interested in the technical details:
Our approach was to utilize NER models to find the following entities: names, towns and cities, concentration camps, ghettos, etc.
Once we had the labelled entities, we used Relation Extraction models to group the names and places into clusters, finding not only names but actual stories. As an additional note, this project was recognized with the IT AWARD for its innovative approach to affirming our national commitment never to forget.
We are working on implementing generative AI to create automated summaries and improve the accessibility of stories, archival information, and lengthy academic texts for visitors and casual viewers.
I look forward to using the latest and greatest in NLP and AI to keep the memories, and the stories alive, and ensure that we will never forget the Shoah, and keep it alive for future generations.