The LFM architecture opens a new design space for foundation models. This is not restricted to language, but can be applied to other modalities: audio, time series, images, etc. It can also be optimized for specific platforms, like @Apple
, @AMD
, @Qualcomm
, and @cerebras
MACHINE LEARNING
-
LFM Architecture Enables New Foundation Model Design Space
By
–
-
LFM-3B Achieves 32K Context Window with Promising RULER Scores
By
–
In this preview release, we focused on delivering the best-in-class 32k context window. These results are extremely promising, but we want to expand it to very, very long contexts. Here are our RULER scores (
https://
github.com/hsiehjackson/R
ULER
…) for LFM-3B ↓ -
LFM optimization outperforms transformers at 1B to 40B scale
By
–
We optimized LFMs to maximize knowledge capacity and multi-step reasoning. As a result, our 1B and 3B models significantly outperform transformer-based models in various benchmarks. And it scales: our 40B MoE (12B activated) is competitive with much bigger dense or MoE models.
-
What is AI? Exploring Artificial Intelligence and Machine Learning
By
–
What is AI?
#AI #AIio #BigData #ML #NLU #Futureofwork @TopCyberNews @SpirosMargaris @MarshaCollier @MHiesboeck
@MHcommunicate @Fisher85M @MikeQuindazzi @NealSchaffer -
OpenAI Launches Advanced AI Models for Science and Mathematics
By
–
OpenAI has announced the launch of its latest AI models last week, designed to significantly enhance problem-solving capabilities in science, coding, and mathematics. These advanced models boast improved reasoning and critical thinking skills, enabling them to tackle complex
-
Genetic Algorithms and Evaluation Cost Trade-offs in AI
By
–
Yes I've tried that https://
x.com/maximelabonne/
status/1767124527551549860
… (never trust genetic algorithms when the cost of eval is so high however) -
MATLAB R2024b Released: 1900+ Updates Including Deep Learning
By
–
Have you downloaded R2024b yet? MATLAB R2024b has 1,900+ updates! Highlights: MATLAB Package Manager Data sonification Improved ODE solving New plot types Deep learning segmentation More in the blog https://
spr.ly/6012o2cco Let us know what you think! -
Meta challenge on video with 65% accuracy and $1M
By
–
If you want some literature, here's a meta challenge on VIDEO, where the best algorithm only achieved 65% accuracy. (The random baseline is 50%) And on video, it should be even easier, right? (With a $1M prize pool!!!)
-
Critique of AI detection: 90% vs 98% in simple processing
By
–
The worst part is they must be all happy: "Hey look René, we managed to detect photos at 90%." For info, you use dlib and opencv and check eye alignment with a threshold without AI, just with image processing on thispersondoesntexist, you get 98%.
-
Incompetence: AI trained on 20,000 random photos to detect deepfakes
By
–
Encore une belle preuve d’incompétence de la @ComCyberMI sur les sujets d’IA la semaine dernière au 20h de @TF1.
— Defend Intelligence (Anis Ayari) (@DFintelligence) 30 septembre 2024
Les « experts » expliquent avoir entraîné une IA pour détecter des deepfakes… AVEC 20 000 photos de thispersondoesnotexist… MDRRRR !
Ils ont entraîné un… pic.twitter.com/59LsnB3luLAnother fine proof of incompetence from @ComCyberMI on AI topics last week on @TF1's 8pm news. The 'experts' explain that they trained an AI to detect deepfakes… USING 20,000 photos from thispersondoesnotexist… LOL! They trained a discriminator on GAN methods, yes