sayzardJan 6

Constellation Research (@constellationr)

DeepSeek(@DeepSeek_ai)가 발표한 기술 논문에 관한 언급으로, 논문은 대형 파운데이션 모델을 훈련하기 위한 새로운 아키텍처를 제안하며 'AI muscle head 시대'의 종언을 주장하는 증거를 제시한다고 설명합니다. 새로운 훈련 아키텍처 관련 연구 발표입니다.

https://x.com/constellationr/status/2008312586098667799

#research #architecture #foundationalmodels #deepseek #paper

Constellation Research (@constellationr) on X

.@DeepSeek_ai's paper latest evidence AI muscle head era coming to end https://t.co/2OyjsxfwjW DeepSeek published a technical paper that argues for a new architecture to train foundational models. DeepSeek's paper is just another development leading to the…

X (formerly Twitter)
Tero Keski-ValkamaDec 15

I have been impacted by layoffs today, so I am open for new opportunities!

If you are looking for a very experienced AI engineer full-remote from Spain, let's get in touch!

#AI #engineer #healthcare #supplychain #logistics #maps #robotics #RL #FoundationalModels #ML

Trending Stocks 📈Jul 11, 2025

'AI Business: Huge Growth?' - All-In Podcast

#revenue #foundationalmodels #aibusinesses #ai #revenuegrowth

Glen BersethJul 31, 2024

Myself and Yoshua Bengio are hiring a postdoctoral researcher @Mila_Quebec
! For this call, we are prioritizing candidates with experience in reinforcement learning, scientific discovery, or high-impact applications of ML. Apply here (https://docs.google.com/forms/d/e/1FAIpQLScqXiMClkgDBvrIZyxdtx60Pcbj3JzZeC-LFg3yiUOZlvgyLw/viewform?usp=sf_link)

We are looking for someone with skills in the areas of
#Reinforcementlearning
#ML4science
#Bayesianoptimization
#Foundationalmodels for #decisionmaking
#RealworldML

https://fracturedplane.notion.site/Open-PostDoc-Position-in-Machine-Learning-fcbcc0e8759441b2b6be12f8fe30080c

Google Forms: Sign-in

Access Google Forms with a personal Google account or Google Workspace account (for business use).

AIagent.at 🤖 AI NewsJun 13, 2024
»#Enterprises are preparing to build their own #LLMs - why that's a smart move: Enterprise LLMs can be built on #publiclyavailable, #foundationalmodels. That's a good start.« https://www.zdnet.com/article/enterprises-are-preparing-to-build-their-own-llms-why-thats-a-smart-move/?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI
Enterprises are preparing to build their own LLMs - why that's a smart move

Enterprise LLMs can be built on publicly available, foundational models. That's a good start.

ZDNET
Show threadEstelle PlatiniApr 6, 2024

The sources said that the approval to automatically adopt #Lavender’s kill lists, which had previously been used only as an auxiliary tool, was granted about two weeks into the war, after intelligence personnel “manually” checked the accuracy of a random sample of several hundred targets selected by the #AI system. When that sample found that Lavender’s results had reached 90 percent accuracy in identifying an individual’s affiliation with Hamas, the army authorized the sweeping use of the system. From that moment, if Lavender decided an individual was a militant in Hamas, the sources were essentially asked to treat that as an order.

“Still, I found them more ethical than the targets that we bombed just for ‘deterrence’ — highrises that are evacuated and toppled just to cause destruction.”

https://www.972mag.com/lavender-ai-israeli-army-gaza/ @israel @data

#metrics #probabilities #usability #ModelCalibration #MachineLearning #ML #OutputAudit #FoundationalModels

‘Lavender’: The AI machine directing Israel’s bombing spree in Gaza

The Israeli army has marked tens of thousands of Gazans as suspects for assassination, using an AI targeting system with little human oversight and a permissive policy for casualties, +972 and Local Call reveal.

+972 Magazine
Victoria Stuart 🇨🇦 🏳️‍⚧️Mar 1, 2024

A Phase Transition in Diffusion Models Reveals the Hierarchical Nature of Data
https://arxiv.org/abs/2402.16991

This caught my eye: "Natural data such as images are believed to be composed of features organised in a hierarchical and combinatorial manner, which neural networks capture during learning."

#NeuralNetworks #AI #MachineLearning #LLM #SyntheticData #DiffusionModels #text2vision #combinatorics #FoundationalModels #dimensionality #PatternRecognition #EmergentProperties

A Phase Transition in Diffusion Models Reveals the Hierarchical Nature of Data

Understanding the structure of real data is paramount in advancing modern deep-learning methodologies. Natural data such as images are believed to be composed of features organized in a hierarchical and combinatorial manner, which neural networks capture during learning. Recent advancements show that diffusion models can generate high-quality images, hinting at their ability to capture this underlying compositional structure. We study this phenomenon in a hierarchical generative model of data. We find that the backward diffusion process acting after a time $t$ is governed by a phase transition at some threshold time, where the probability of reconstructing high-level features, like the class of an image, suddenly drops. Instead, the reconstruction of low-level features, such as specific details of an image, evolves smoothly across the whole diffusion process. This result implies that at times beyond the transition, the class has changed, but the generated sample may still be composed of low-level elements of the initial image. We validate these theoretical insights through numerical experiments on class-unconditional ImageNet diffusion models. Our analysis characterizes the relationship between time and scale in diffusion models and puts forward generative models as powerful tools to model combinatorial data properties.

arXiv.org
Ric Harvey 🇪🇺🌍💚Nov 29, 2023

Hi folks, this time it’s time to look at the announcements from Adam Selipskys #Keynote at #AWS #reinvent 2023. I’ve broken this down into a couple of videos to make it easier to digest which I’ll release throughout the day. We are going to focus on #GenAI #Bedrock including it’s #FoundationalModels and #AmazonQ then I’ll follow up with videos on #Storage and general announcements.

https://youtu.be/LApZcLdHAow - Adam’s Keynote Gen AI announcements

Don't miss the announcements from Adam Selipsky's Keynote - AWS reinvent 2023

YouTube
Show threadJohannes LotzNov 17, 2023

🚀 UMedPT outperforms traditional ImageNet pretraining and provides a new base model for future research. It maintains performance with just 1% of the data for in-domain tasks, and 50% for out-of-domain tasks. This should be pretty handy for research with small/limited datasets.

🔜 Code coming soon! Thanks to the team and everyone who shares labeled data which only makes large-scale supervised training possible! 🩺💻👩‍💻 #OpenScience #FoundationalModels (3/3)

Show threadAmit Bahree 🌎💾Nov 19, 2022

@slatchison When I read contribution model - I was like wait, is that different from #FoundationalModels and #LLM? What did I miss? 😬

And it took a few seconds for it to grok. 🤪