Initial results with process supervision — training an AI by rewarding it for coming up with a correct thought process, not just outputting the right answer. Great results in mathematical reasoning:
MACHINE LEARNING
-
OpenAI Improves Math Reasoning with RLHF and Chain-of-Thought
By
–
research from @OpenAI on improving math reasoning by RLHF with a reward model trained on 800k human-generated chain-of-thought data (which @scale_AI partnered w/
@OpenAI on!) RLHF seems to be a scalable technique for making LLMs smarter in many ways -
How AI Can Transform The Software Engineering Process
By
–
How AI Can Transform The Software Engineering Process
#AI #AIio #BigData #ML #NLU #Futureofwork @briansolis @Xbond49 @VitalikButerin @DataScienceCtrl @Benioff @johnnosta @IrmaRaste @Tiffani_Bova http://
ow.ly/U09430svtq3 -
Diffusion Models: Technical Deep Dive and Sprite Generation
By
–
4/How Diffusion Models Work: Learn the technical details of how diffusion models – which power Midjourney, DALL·E 2, and Stable Diffusion – work. You’ll also end up with working code to generate your own video game sprites in Jupyter! https://
learn.deeplearning.ai/diffusion-mode
ls
… -
OpenAI Releases Process Supervision Method for Mathematical Reasoning
By
–
Open AI releases paper + dataset Let’s Verify Step by Step trained a model to achieve a new state-of-the-art in mathematical problem solving by rewarding each correct step of reasoning (“process supervision”) instead of simply rewarding the correct final answer (“outcome
-
Tab-CoT: Zero-shot Tabular Chain of Thought Framework
By
–
Tab-CoT: Zero-shot Tabular Chain of Thought propose a new Chain-of-Thought framework Tab-CoT, which use a tabular format to conduct complex reasoning process in a highly structured manner. Despite its simplicity, we show that our approach is capable of performing reasoning
-
New ML Methodology for Software Development Assistance Tools
By
–
Today we describe a new methodology for training large #ML models on the historical process of code development to build models that provide software development assistance, like comment resolution, build repair, and tip prediction. See how it’s done at https://t.co/IrReKKjyRx pic.twitter.com/V13NBtEAXY
— Google AI (@GoogleAI) 31 mai 2023Today we describe a new methodology for training large #ML models on the historical process of code development to build models that provide software development assistance, like comment resolution, build repair, and tip prediction. See how it’s done at https://
goo.gle/3oECaPO -
Process Supervision Achieves State-of-Art Mathematical Reasoning
By
–
We trained an AI using process supervision — rewarding the thought process rather than the outcome — to achieve new state-of-art in mathematical reasoning. Encouraging sign for alignment of advanced AIs: …
https://
openai.com/research/impro
ving-mathematical-reasoning-with-process-supervision
… -
HarmoneyNZ Improves Risk Model Efficiency With DataRobot
By
–
📈 @HarmoneyNZ, the Australian online marketing place-lending company, chose DataRobot to increase efficiencies, where a 1% improvement to a risk model brings $1M in annual savings.
— DataRobot (@DataRobot) 31 mai 2023
Learn more ⬇️https://t.co/M5tnyRAkgl pic.twitter.com/WYRVesT0Zn@HarmoneyNZ
, the Australian online marketing place-lending company, chose DataRobot to increase efficiencies, where a 1% improvement to a risk model brings $1M in annual savings. Learn more https://
bit.ly/42eYRYn -
Sparsity’s Impact on Neural Network Training Research
By
–
Our latest podcast is live! In this episode, @vithursant19 and @draecomino explore cutting-edge research on sparsity and its impact on training neural networks.
— Cerebras (@cerebras) 31 mai 2023
Click here to listen: https://t.co/SxmKlz2DtW pic.twitter.com/O7yYyJkbvMOur latest podcast is live! In this episode, @vithursant19 and @draecomino explore cutting-edge research on sparsity and its impact on training neural networks. Click here to listen: https://
hubs.li/Q01RPx_C0