We’re proud to be a Diamond Sponsor of the 40th annual @ICMLConf
, held this week in Honolulu, Hawaii. Registered for #ICML2023? We hope you’ll visit the Google booth to learn more about our ML research! Learn more about our involvement → https://
goo.gle/44Rxc29
@googleai
-
Google Diamond Sponsor of 40th Annual ICML Conference
By
–
-
AI Societal Context Understanding Tools Solutions
By
–
Understanding the broader societal context in which AI products are built and used is essential to making them robust & responsible. Today on the blog, learn about the Societal Context Understanding Tools and Solutions (SCOUTS) team’s work in this space → https://
goo.gle/3DiPq0h -
SimPer: Self-Supervised Contrastive Learning for Periodic Data
By
–
SimPer is a self-supervised contrastive framework for learning periodic information in data, while improving data efficiency and generalization to distribution shifts. Check it out and copy the code → https://t.co/P7eEpf1AKq pic.twitter.com/40G18nzKeu
— Google AI (@GoogleAI) 18 juillet 2023SimPer is a self-supervised contrastive framework for learning periodic information in data, while improving data efficiency and generalization to distribution shifts. Check it out and copy the code → https://
goo.gle/3XZcFWC -
Ravi Kumar’s Lattice Algorithm Wins ACM STOC Award
By
–
Congratulations to Ravi Kumar and co-authors on winning ACM's STOC 20 year Test of Time award for their 2003 paper titled “A sieve algorithm for the shortest lattice vector problem”! Learn more → https://
goo.gle/46Ms6p7 -
Symbol Tuning: Improving In-Context Learning with Arbitrary Symbols
By
–
Today on the blog, read all about symbol tuning, a method that fine-tunes models on tasks where natural language labels are replaced with arbitrary symbols. Learn how symbol tuning can improve in-context learning for benefits across various settings →
https://
goo.gle/3rt3r8U -
Universal Dependency Annotation Wins 2023 ACL Test of Time Award
By
–
Congratulations to the authors of "Universal dependency annotation for multilingual parsing", which won one of this year's 2023 ACL Test of Time Awards for most influential papers released 10 years ago. Check it out at https://
aclanthology.org/P13-2017.pdf -
Google Distillation Method Trains Smaller LLMs Efficiently
By
–
Deploying LLMs is memory inefficient and compute-intensive. At the #ACL2023 Google booth at 3pm today, @chunliang_tw will describe a new method, distilling step-by-step, that trains successful small models using fewer training data for fine-tuning and distillation.
-
PropSegmEnt: Expert-Annotated Dataset for Textual Attribution
By
–
PropSegmEnt is a large expert-annotated dataset for for realistic textual attribution, claim provenance, and more. Drop by the @aclmeeting Google booth at 12:30 today to learn about it from authors Alex Fabrikant & @soshsihao
. -
Interactive Plan-Based Conditional Generation Platform at ACL2023
By
–
At 10:30am today at the #ACL2023 Google booth, we'll present an interactive platform demonstration for plan-based conditional generation, a new conceptualization of text plans as a sequence of question-answer (QA) pairs. Drop by to learn all about it!
-
Wordcraft: LaMDA-Powered Writing Assistant for Creative Writers
By
–
Need help kickstarting your writing? Today at the @aclmeeting Google booth (3:45pm), we'll demonstrate Wordcraft, an open-source LaMDA-augmented text editor that provides creative writers with writing suggestions and brainstorming tools. Come learn all about it!