🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Shout out to Bodhisattwa Majumder, who was the main driver behind our recently accepted ACL 2025'20 paper!

Check out Google AI's blog post covering our upcoming ACL paper on extracting structured data from template documents. Shout out to Sandeep Tata, Marc Najork, Bodhisattwa Majumder, Nav, and Qi!

Check out Lending Doc AI, one of the products powered by the kind of ML research we described in ai.googleblog.com/2020/06/extrac…

Applications are open for our CS Research Mentorship Program — CSRMP. Students from underrepresented groups in computing are paired with Google mentors to support their pursuit of research pathways. Learn more & apply by Nov 18 ➡️ research.google/outreach/csrmp/

If you're working on interesting problems in understanding documents, consider submitting your work to the Document Intelligence Workshop at KDD aka.ms/di-2021 4-page abstracts due by May 10th. #KDD #AI

Vision paper w/ @ytay017 Dara Bahri Marc Najork on one (of many) possible paths towards building domain search experts by bringing the latest and greatest from NLP, ML, and IR to bear on the problem.

Andrei Andrei Broder does it again! AFAIK, he's the first person to receive the Kanellakis Award TWICE. The Power of Two Choices is also such an elegant idea. Congrats to the winners!

If you're interested in document understanding, check out the Document Intelligence workshop (document-intelligence.github.io/DI-2021/) at #KDD2021. We have an exciting line-up of invited talks! 1/7

Congrats to my colleague Marian R Croak on her election to the National Inventors Hall of Fame! Really wonderful to see her strong contributions & inventive spirit recognized in this way! If you've used a phone, you've probably used her inventions on VoIP. blog.google/technology/ai/…

# RLHF is just barely RL Reinforcement Learning from Human Feedback (RLHF) is the third (and last) major stage of training an LLM, after pretraining and supervised finetuning (SFT). My rant on RLHF is that it is just barely RL, in a way that I think is not too widely

this is a tweet

We are happy to announce Google Cloud Machine Learning, based on #TensorFlow goo.gl/Gi6yB7 #GCPNEXT2016