Read all about recent advances in using ML to improve efficiency of ML workloads — including TpuGraphs, a performance prediction dataset on large tensor computational graphs, and Graph Segment Training, a novel method to scale graph neural networks →
https://
goo.gle/3RrnQUZ
@googleai
-
ML Advances: TpuGraphs Dataset and Graph Segment Training
By
–
-
Google Magic Editor AI-Powered Photo Editing Tool
By
–
Google Research helped develop Magic Editor, an AI-powered photo editor that allows users to easily make complex edits to photos using generative AI. We fine-tuned models for inpainting, image transformation, super-resolution, and more. Learn more → https://t.co/9HB9NeIHOn pic.twitter.com/8LM512r5Gn
— Google AI (@GoogleAI) 15 décembre 2023Google Research helped develop Magic Editor, an AI-powered photo editor that allows users to easily make complex edits to photos using generative AI. We fine-tuned models for inpainting, image transformation, super-resolution, and more. Learn more → https://
goo.gle/3t4lcfU -
StyleDrop: AI Model for Stylized Text-to-Image Synthesis
By
–
Introducing StyleDrop, a model that allows a significantly higher level of stylized text-to-image synthesis by using a few style reference images that describe the style for text-to-image generation, bypassing the burden of text prompt engineering. More→ https://t.co/F3Rw3QlbtP pic.twitter.com/2J4wljmFwF
— Google AI (@GoogleAI) 15 décembre 2023Introducing StyleDrop, a model that allows a significantly higher level of stylized text-to-image synthesis by using a few style reference images that describe the style for text-to-image generation, bypassing the burden of text prompt engineering. More→ https://
goo.gle/3v0PsZw -
NeurIPS 2023 Best Paper: Privacy Auditing with One Training Run
By
–
Congratulations to Thomas Steinke, Milad Nasr, & Matthew Jagielski for their paper "Privacy Auditing with One (1) Training Run", which won the #NeurIPS2023 Award for Best Paper! https://
arxiv.org/abs/2305.08846 -
Google Quantum Hardware Engineering Manager Ani Nersisyan
By
–
Meet Ani Nersisyan, Quantum Hardware Engineering Manager on the Google Quantum AI team.
— Google AI (@GoogleAI) 14 décembre 2023
Her team navigates a diverse set of problems — from understanding processor performance and underlying physics, to communicating with foundries on fabrication stack details. Learn more ↓ pic.twitter.com/S9B1mtJUzJMeet Ani Nersisyan, Quantum Hardware Engineering Manager on the Google Quantum AI team. Her team navigates a diverse set of problems — from understanding processor performance and underlying physics, to communicating with foundries on fabrication stack details. Learn more ↓
-
Google Research Drives Sustainability and Algorithm Innovation
By
–
Whether we're shaping the future of sustainability or optimizing algorithms, Google Research has never been more driven to improve the lives of billions of people. Learn how our researchers are asking the big questions to create meaningful change → https://t.co/K7VF1B0Qt2 pic.twitter.com/aS3zjh8XuN
— Google AI (@GoogleAI) 14 décembre 2023Whether we're shaping the future of sustainability or optimizing algorithms, Google Research has never been more driven to improve the lives of billions of people. Learn how our researchers are asking the big questions to create meaningful change → https://
goo.gle/47TvEGg -
Google Research Enhances Clear Calling with Full-Band Audio ML
By
–
We’re proud to highlight Google Research’s contributions to improving Clear Calling, the background noise reduction feature on Pixel, which can now handle full-band audio & is powered by an audio-to-audio ML model that was optimized to run at low latency on Google Tensor. pic.twitter.com/adCJzfQUrM
— Google AI (@GoogleAI) 14 décembre 2023We’re proud to highlight Google Research’s contributions to improving Clear Calling, the background noise reduction feature on Pixel, which can now handle full-band audio & is powered by an audio-to-audio ML model that was optimized to run at low latency on Google Tensor.
-
Google Demonstrates Multimodal RAG Med-PaLM M at NeurIPS 2023
By
–
Stop by the #NeurIPS2023 Google booth today at 4:15pm when Ryan Knuffman & Milica Cvetkovic will demostrate Multimodal RAG, a question answering system powered by Med-PaLM M and grounded in multimodal retrieval-augmented generation that is geared towards patient inquiries.
-
Sharpness-Aware Minimization and Hessian Structure at NeurIPS
By
–
At 4:15pm today @TheGradient will be at the #NeurIPS2023 Google booth to talk about the differences between Sharpness-Aware Minimization (that improves generalization) and similar methods (that don't), which can be explained by the structure of the Hessian of the loss function.
-
Google Resonator: Game Interface Explores Music AI Models
By
–
Join us at the #NeurIPS2023 Google booth today at 12:45pm when Erin Drake Kajioka & Michal Todorovic will host a demo of resonator, an experimental game-based interface for exploring & discovering music through 3D visualizations of a music understanding model embedding. pic.twitter.com/D7izRwqexq
— Google AI (@GoogleAI) 13 décembre 2023Join us at the #NeurIPS2023 Google booth today at 12:45pm when Erin Drake Kajioka & Michal Todorovic will host a demo of resonator, an experimental game-based interface for exploring & discovering music through 3D visualizations of a music understanding model embedding.