Automatically Tasks Extraction From Written Communication

Azure & DevOps Jerusalem is partnering with Reactor Tel Aviv & JerusML community for a special meetup - “Automatically Tasks Extraction From Written Communication”.

Schedule:
18:55-19:00: Signing in
19:00-19:10: Opening words & announcements
19:10-20:10: “Automatically Tasks Extraction From Written Communication" by Tom Braude, Microsoft
20:10-20:30: Q&A from viewers

About this session:
Our daily communication can be overwhelming.
On average, we receive over 120 emails and conduct multiple chat conversations with our colleagues on a daily basis.
Modern NLP techniques can alleviate mental overload.
Though potentially extremely useful, bridging the gap between SOTA techniques, trained and evaluated on curated datasets, and real-world, natural data is not trivial.
In this talk, Tom will present some of his team's recent developments to automatically identify tasks from written communication.

About the speaker:
Tom Braude.
Tom is a Researcher at Microsoft, working on making users' life easier by adding machine learning capabilities to Office365 products, and specifically products based on written communication such as Emails and Chats.
His research focuses on natural language processing and includes topics such as semantic understanding, large scale pretraining and model compression techniques.

*** This event will be in Hebrew ***