Hubert Wassner (@hwassner) header
Hubert Wassner (@hwassner) profile picture

Chief #AI #datascientist @abtasty . Previously teaching #machineLearning #algorithmic at @GroupeESIEA. Worked on #DNA, #biometry, now on #marketing.

Paris
Joined March 2010
Hubert Wassner retweeted
Il y a quelques semaines, j’ai voulu tenter une expérience. Et si on laissait réellement à des IA les rênes d’un site ? J’ai donc créé un network de sites gérés à 100% par des IA (#chatGPT notamment), totalement autonomes… Le network Generation ! Thread explicatif ⤵️⤵️
88
531
177
1,333
1,090,083
Hubert Wassner retweeted
My interpretation of prompt engineering is this: 1. A LLM is a repository of many (millions) of vector programs mined from human-generated data, learned implicitly as a by-product of language compression. A "vector program" is just a very non-linear function that maps part of the latent space unto itself. 2. When you're prompting, you're fetching one of these programs and running it on an input -- part of your prompt serves as a kind of "program key" (as in database key) and part serves as program argument(s). Like, in "write this paragraph in the style of Shakespeare: {my paragraph}", the part "write this paragraph in the stye of X: Y" is a program key, with arguments X=Shakespeare and Y={my paragraph}. 3. The program fetched by your key may or may not work well for the task at hand. There's no reason why it should be optimal. There are lots of related programs to choose from. 4. Prompt engineering represents a search over many keys in order a find a program that is empirically more accurate for what you're trying to do. It's no different than trying different keywords when searching for a Python library. 5. Everything else is unnecessary anthropomorphism on the part of the prompter. You're not talking to a human who understands language the way you do. Stop pretending you are.
115
505
100
2,854
711,151
Hubert Wassner retweeted
Happy to announce DreamFusion, our new method for Text-to-3D! dreamfusion3d.github.io We optimize a NeRF from scratch using a pretrained text-to-image diffusion model. No 3D data needed! Joint work w/ the incredible team of @BenMildenhall @ajayj_ @jon_barron #dreamfusion
126
1,378
411
5,422
Hubert Wassner retweeted
Stable Diffusion is only 30 days old… a MEGA THREAD 🧵 on its rapid rise.
261
3,009
812
12,587
New notebook: "Grokking Stable Diffusion" Hopefully useful for anyone looking for content that goes deeper than just how to run the model! colab.research.google.com/dr… Covering: - The diffusion loop - Messing with text embeddings - img2img - arbitrary guidance More deets in thread 🧵
30
290
26
1,816
Hubert Wassner retweeted
Combining @StabilityAI #StableDiffusion generative powers + Human guidance and graphic skills* with tools like @Photoshop in a coherent workflow. * Of which I've got about 0 as can be seen below.
72
842
260
3,098
Diffusion models like #DALLE and #StableDiffusion are state of the art for image generation, yet our understanding of them is in its infancy. This thread introduces the basics of how diffusion models work, how we understand them, and why I think this understanding is broken.🧵
72
1,179
243
6,311
Dans le n6 d'#ActuIA je parle de mes travaux sur les interfaces cerveau-ordinateur au service des personnes en situation de handicap en France @ActuIAFr #braini
1
4
15
Nous avons le plaisir de vous annoncer que nous sommes partenaires de la #ParisAIWeek !
#parisaiweek Nous sommes fiers d'être partenaires d'@ActuIAFr IA pour la Paris AI Week💪 📢 Rendez-vous sur parisaiweek.com pour proposer votre événement et découvrir la programmation cc @ActuIAFr @LaFrenchTech @TechSaclay @FRdigitale @namr_france @Lararouyres
4
9
En kiosque, #ActuIA le magazine de l'#IntelligenceArtificielle Avec @JiliJeanlouis (@OVHcloud), pour une utilisation raisonnée des ressources en #IntelligenceArtificielle A retrouver proche de chez vous sur 👉web2store.mlp.fr/produit.asp… #IA #intelligenceartificielle #AI #cloud
6
9
This is a day I’ve dreamed of my whole life, this is the reason @DeepMind was founded, to build AI and use it accomplish extraordinary scientific breakthroughs like #AlphaFold 2, to advance science and benefit humanity. I could not be more proud of the incredible team!
Today with @emblebi, we're launching the #AlphaFold Protein Structure Database, which offers the most complete and accurate picture of the human proteome, doubling humanity’s accumulated knowledge of high-accuracy human protein structures - for free: dpmd.ai/alphafolddb 1/
135
1,058
127
5,871
Hubert Wassner retweeted
Drone photographer Lior Patel followed a herd of sheep for several months, as the herd was shepherded to its summer pasture. Entrancing and relaxing.
739
16,748
4,362
62,847
This is not a motherboard - it's the Gwangyang Steel Plant in South Korea, the world's largest steel mill
103
2,098
271
10,906
[ Dé-va-li-sés ] : nous n'avons plus aucun exemplaire d'ActuIA N°4 version papier en stock pour les futurs abonnements. En revanche il sera dispo dès demain dans tous vos kiosques préférés. 😍
1
4
9
L’American College of Radiology publie une enquête sur les usages de l’intelligence artificielle en radiologie actuia.com/actualite/lameric… #IA #IntelligenceArtificielle #esante #ehealth #hcsmeufr #sante
📌Ne manquez pas le magazine de l' #IntelligenceArtificielle N°4, entre autres : l' #IA & le #sport, l'évolution des métiers de l'IA, #recherche, #ReinforcementLearning, le @ParisMLgroup, le #LS2N Pour le recevoir => boutique.actuia.com/shop/mag…
5
4
🔎Où trouver #ActuIA N°3 en kiosque ? 👉web2store.mlp.fr/produit.asp… De nouveaux points de vente très prochainement. #IA #IntelligenceArtificielle #AI #ArtificialIntelligence #TransfoNum #VendrediLecture
6
8
Hubert Wassner retweeted
There are two distinct types of careers: - One where the payoff follows a normal distribution (doctor, programmer, plumber, architect). - The other where the payoff follows a Pareto distribution (musician, author, startup founder, youtuber). ...
35
374
46
1,863