I couldn't agree more. And they could have a powerful role in the conversation about how/if to archive influential datasets which is a woeful gap at the moment
DATA
-
Three AI challenges for Orange: training, security, talent
By
–
Retour au talk de @Cheydema pour la ceo de @orange , il faut adresser 3 grand défi avec l’#IA :
— Defend Intelligence (Anis Ayari) (@DFintelligence) 14 juin 2023
– des formations internes pour rester up to date avec les avancés
– se focaliser sur la sécurité des donnés (qui sont un véritable asset pour les entreprises)
– attirer les meilleurs… pic.twitter.com/wn8TNRw9lIBack to the talk by @Cheydema for the CEO of @orange, we need to address 3 major challenges with #AI: – internal training to stay up to date with advances – focus on data security (which is a real asset for companies) – attract the best
-
Yann LeCun: AI intrinsically good amplifies our behavior
By
–
Ce à quoi @ylecun lui a répondu que l’IA est intrinsequement bon car ce domaine nous rend plus intelligent, et que l’#IA ne fait que amplifier notre propre comportement (par nos données) #vivatech @orange vous en pensez quoi ? pic.twitter.com/7mo3JhNFsD
— Defend Intelligence (Anis Ayari) (@DFintelligence) 14 juin 2023To which @ylecun replied that AI is intrinsically good because this field makes us smarter, and that #AI only amplifies our own behavior (through our data) #vivatech @orange what do you think?
-
New AI Algorithms for Uncertainty-Calibrated Data Explanations
By
–
Researchers have found a new set of AI algorithms that construct & judge data explanations using sequential Monte Carlo techniques. The algorithms provide uncertainty-calibrated insights & track the likelihood of explanations. Learn more: https://
bit.ly/42HajfA @nasscom -
Socially Aware Temporally Casual Decoder Recommender Systems Study
By
–
STUDY: Socially Aware Temporally Casual Decoder Recommender Systems paper page: https://
huggingface.co/papers/2306.07
946
… With the overwhelming amount of data available both on and offline today, recommender systems have become much needed to help users find items tailored to their interests. -
AI Model Collapse: Risks of Training on AI-Generated Content
By
–
Good one on #UGC vs #AIgeneratedcontent #GenerativeAI The AI feedback loop: Researchers warn of ‘model collapse’ as AI trains on AI-generated content https://
venturebeat.com/ai/the-ai-feed
back-loop-researchers-warn-of-model-collapse-as-ai-trains-on-ai-generated-content/
… via @VentureBeat -
H2O AI Moderates Panel on Organizational Transformation with AI at CDAO México 2023
By
–
[English] We're at #CDAOMexico 2023! It was an honor to moderate today's panel "Creating Data-Driven Companies: The AI-Driven Organizational Transformation" alongside outstanding leaders from @BancoAzteca, @amazon, and @rappicardmx. Thanks to everyone who participated!
-
Annual State Of Data Quality Survey 2023
By
–
The Annual State Of Data Quality Survey, 2023 https://
bit.ly/3Mt9dhM #AI #MachineLearning #DeepLearning #LLMs #DataScience -
SlimPajama Reduces AI Training Compute Costs by 50 Percent
By
–
Why we built SlimPajama – it's all about training efficiency. Without de-duplication, a model would have to go through 1.2T tokens before seeing ~600B unique tokens. SlimPajama sees 600B tokens in half the time. That saves you 50% on compute costs! https://
cerebras.net/blog/slimpajam
a-a-627b-token-cleaned-and-deduplicated-version-of-redpajama
… -
Domino Data Lab Welcomes Thomas Been as New CMO
By
–
Thrilled to welcome Thomas Been as our new CMO at @DominoDataLab
! Discover why he's #ExcAItedToJoinDomino and the transformative potential of AI in his first blog post. https://
domino.buzz/43UdqBT #datascience #MLOps