Microsoft’s Structured Prompting Breaks In-Context Learning Length Limits, Scales to Thousands of Examples https://
syncedreview.com/2022/12/20/mic
rosofts-structured-prompting-breaks-in-context-learning-length-limits-scales-to-thousands-of-examples/
…
RESEARCH
-
Microsoft’s Structured Prompting Scales In-Context Learning to Thousands
By
–
-
BCG’s Most Innovative Companies Ranking for 2022
By
–
The Most Innovative Companies of 2022 >>> as ranked by @BCG via @MikeQuindazzi >>> Full Report http://
shorturl.at/efqGS -
Integrating datasets with Hugging Face model evaluator tools
By
–
cool work @EthanJPerez & team! Would be awesome to add the datasets to http://
hf.co/datasets or some sort of integration with https://
huggingface.co/spaces/autoeva
luate/model-evaluator?dataset=acronym_identification
… if that makes sense! -
OpenAI Open-Sources Point-E: Fast 3D Point Cloud Generation
By
–
Just open-sourced Point·E — DALL·E but for 3D point clouds. Extremely fast to generate: https://t.co/fXrdnxi7Qw pic.twitter.com/PUKRZO9fum
— Greg Brockman (@gdb) 20 décembre 2022Just open-sourced Point·E — DALL·E but for 3D point clouds. Extremely fast to generate: https://
github.com/openai/point-e -
Instruction Models Now Available in Open Source
By
–
Instruction models making their way into the open-source!
-
Flamingo AI Advances Computer Vision Beyond Past Predictions
By
–
10 yrs ago @karpathy wrote a blog post on the outlook of AI: https://
karpathy.github.io/2012/10/22/sta
te-of-computer-vision/
… in which he describes how difficult it would be for an AI to understand a given photo, concluding "we are very, very far and this depresses me."
Today, our Flamingo steps up to the challenge. -
AI Inference Hardware Optimization for Machine Learning Workloads
By
–
Thank you for your continued support as we work to help companies run AI inference workloads faster, cooler, and more cost-effectively using #atmemorycomputation. We look forward to the innovations and advancements that 2023 will hold! #ML #deeplearning #neuralnetworks #AI
-
Fine-tuning vs. In-context Learning: Exploring Alternative Approaches
By
–
this is great question! i think its an unexplored space. finetuning is much more heavy weight (error prone, hard to do, more data required, etc), so it will be interesting to see if you can achieve similar performance just by using examples
-
AlphaCode and AlphaTensor Named Top 10 Science Breakthroughs
By
–
Awesome to see #AlphaCode (AI for competitive programming) and #AlphaTensor (AI for efficient matrix multiplication) both feature in Science's Top 10 Breakthroughs of the Year! Congrats to the teams!!
-
Connect-the-Dots: New Algorithm for Differential Privacy Loss Analysis
By
–
Introducing Connect-the-Dots, a new accounting algorithm that uses an indirect approach to accurately discretize privacy loss distributions, yielding a useful tool for understanding the privacy cost of combinations of differential privacy algorithms. https://
goo.gle/3FIoFmD