used to be the same for ConvNets in 2013. will get democratized in due time.
RESEARCH
-
Meta’s MTIA v1 AI Inference Accelerator Architecture Overview
By
–
A closer look at the architecture of Meta's first-generation AI inference accelerator. MTIA v1 is a custom-designed ASIC built with next-generation recommendation model requirements in mind.
— AI at Meta (@AIatMeta) 21 juin 2023
Read more ➡️ https://t.co/ho0QqWdiw9
We'll be sharing more at #ISCA2023. pic.twitter.com/COxAAgV1tSA closer look at the architecture of Meta's first-generation AI inference accelerator. MTIA v1 is a custom-designed ASIC built with next-generation recommendation model requirements in mind. Read more https://
bit.ly/46hV3sU We'll be sharing more at #ISCA2023. -
Public sector moonshot AI investment national research infrastructure
By
–
Addressing this imbalance will require a significant public sector investment in AI on the scale of a moonshot, which will include the National AI Research Resource and national labs similar to CERN inside the U.S. 5/
-
Private compute dominance threatens fundamental AI research access
By
–
The vast amounts of compute and data required to train these systems is accessible only to private industry leaders, creating a significant imbalance in the AI development ecosystem putting at risk the critical basic science research of the tech needed for our moonshot. 4/
-
AI’s Role in Cancer Cure and Healthcare System Improvement
By
–
AI has the opportunity to help with so many unsolved problems from curing cancer and life-saving drug discovery to improving the healthcare delivery system. 3/
-
Understanding People and Society for Better Product Design
By
–
To design and create useful and high quality products, you should understand people and society.
-
MAGVIT: Single Transformer for Ten Video Generation Tasks
By
–
Introducing MAGVIT, a single transformer designed to address ten video generation tasks, including future image animation, video editing and video outpainting. Stop by the #CVPR2023 Google booth at 10am today to see it in action! More details at: https://t.co/nd7Gew2ozv pic.twitter.com/8J35j5fnIG
— Google AI (@GoogleAI) 21 juin 2023Introducing MAGVIT, a single transformer designed to address ten video generation tasks, including future image animation, video editing and video outpainting. Stop by the #CVPR2023 Google booth at 10am today to see it in action! More details at: https://
magvit.cs.cmu.edu -
Lupe Fiasco, Fox Harrell & Nick Montfort on AI and Computational Creativity
By
–
Lupe Fiasco, Fox Harrell & Nick Montfort discuss AI, computational creativity, and teaching: http://
bit.ly/46j4mZE -
Starting New LLM Project: Which Technologies to Choose?
By
–
If you started a new LLM project today would you use:
-
Semi-DETR: Transformer-Based Semi-Supervised Object Detection
By
–
We proposed Semi-DETR, the first Transformer-based end-to-end semi-supervised object detector. Results outperform all SOTA methods on COCO & Pascal VOC by clear margins. Read more: https://
openaccess.thecvf.com/content/CVPR20
23/papers/Zhang_Semi-DETR_Semi-Supervised_Object_Detection_With_Detection_Transformers_CVPR_2023_paper.pdf
… #CVPR2023