Foundation models are enabling breakthrough AI capabilities that are delivering value that was not previously possible. https://
fastcompany.com/90827380/the-b
iggest-tech-trends-of-2023-according-to-over-40-experts
…
RESEARCH
-
Foundation Models Deliver Unprecedented AI Value and Capabilities
By
–
-
Free Harvard Course on Probability for Machine Learning
By
–
"Probabilities" for machine learning. From Harvard.
For Free. You'll learn: • Concepts of probability
• Monte Carlo simulations
• Central Limit Theorem
• and more… https://
edx.org/course/data-sc
ience-probability
… -
Humans and AI Understanding Each Other Better Than Ever
By
–
Humans & #AI Will Understand Each Other Better Than Ever https://
wired.co.uk/article/artifi
cial-intelligence-data-models
… @pierrepinna @sallyeaves @PawlowskiMario @Xbond49 @gvalan @psb_dc @SpirosMargaris @JimMarous @HaroldSinnott @Shi4Tech @mikeflache @Nicochan33 #MachineLearning #DeepLearning #Fintech #DataScience -
Inpainting AI: From Near-Perfect to Flawless Image Generation
By
–
Wow! Por mucho que estas herramientas estén ya en nuestras manos, es imposible no asombrarse con los resultados.
— Carlos Santana (@DotCSV) 28 décembre 2022
Para 2023 sólo pido que los inpaintings dejen de ser imágenes *casi* perfectas, a obtener resultados que no contengan ni un sólo error. ¿Llegaremos?🌠 https://t.co/DUFFeShVxbWow! Por mucho que estas herramientas estén ya en nuestras manos, es imposible no asombrarse con los resultados. Para 2023 sólo pido que los inpaintings dejen de ser imágenes *casi* perfectas, a obtener resultados que no contengan ni un sólo error. ¿Llegaremos?
-
Training ResNets Without Batchnorm: Achieving Strong Results
By
–
Pretty amazing that we can get such good results even without batchnorm — and not using a proper resnet init like Fixup either! Maybe I should try lsuv with batchnorm-free resnet sometime…
-
Bug Fix Improves Model Accuracy to 94.9 Percent
By
–
h/t to @PiotrCzapla for noticing I had a bug that meant there's no batchnorm in those models! After fixing, my new best results: Epochs / c
5 / 93.0
20 / 94.5
50 / 94.9 -
Active Learning: The Year of Breakthrough for Machine Learning Systems
By
–
Active learning can enable machine learning systems to: 1) Adapt to changing conditions 2) Learn from fewer labels 3) Keep a human-in-the-loop for most difficult examples 4) Achieve higher performance. AL has been around for decades, 2023 might finally be the takeoff year.
-
Generative AI to Revolutionize Active Learning Through Uncertainty
By
–
With recent advances in generative AI, Active Learning is primed for a major breakthrough. When an algorithm is unsure of correct label for some part of encoding space, it can actively generate from that section to get input from a human.
-
Reforming Academic Metrics: Beyond Peer-Review Point Systems
By
–
I guess I'm glossing over the 2nd use of peer-reviewed publications though, which is the point-scoring system widely used for academic career development. Maybe that should be replaced with something more personal and less metric-driven.
-
OpenAI’s Point-E Generates 3D Point Clouds From Text Prompts
By
–
OpenAI’s Point·E: Generating 3D Point Clouds From Complex Prompts in Minutes on a Single GPU https://
syncedreview.com/2022/12/27/ope
nais-pointe-generating-3d-point-clouds-from-complex-prompts-in-minutes-on-a-single-gpu/
…