I've been advocating for architectures that can plan using a world models trained Self-Supervised Learning for a very long time.
E.g. my NIPS 2016 keynote (below, 25 minutes in). I've been working on SSL from video since 2014. But, unlike others:
1. I've always said that deep
MACHINE LEARNING
-
Self-Supervised Learning for Planning: A Decades-Long Research Journey
By
–
-
AI Systems Save Lives Through Emergency Braking Medical Imaging
By
–
No.
The overall number of deaths caused by AI is very negative: deep-learning based automatic emergency braking systems and medical image processing systems have been saving lots of lives.
The number of AI-fueled deaths is close to zero. -
Data Science Interview Questions and Answers Curated Guide
By
–
GitHub – youssefHosni/Data-Science-Interview-Questions-Answers: Curated list of data science interview questions and answers https://
bit.ly/42FHnF3 #AI #MachineLearning #DeepLearning #LLMs #DataScience -
Batch Size Optimization Trade-offs in AI Performance
By
–
Generally slowdown is 20-30%, but often for me the increased batch size means it’s faster overall.
-
AI Image Segmentation Technology: Progress Despite Limitations
By
–
Despite imperfections and limitations, the technology notably enhances image segmentation and is a significant step in AI progress, worth exploring for enthusiasts and analysts alike.
-
SAM’s Billion-Mask Dataset Revolutionizes AI Image Segmentation
By
–
SAM stands out with its "Segment Anything 1 billion" dataset, containing 1.1 billion HD masks from 11 million images—400 times larger than any other, underscoring the value of extensive, quality AI data.
-
SAM: Real-Time AI with Textual and Spatial Prompt Integration
By
–
SAM merges textual and spatial prompts with image encodings using a dedicated encoder, enabling real-time AI interactions and quick adaptability across industries without retraining.
-
Meta’s SAM: Revolutionary AI Model for Image Segmentation and Analysis
By
–
The future of image analysis: Meta's SAM, the Segment Anything Model, a groundbreaking leap in AI technology! Meta's AI SAM identifies image segments from a text prompt, enabling precise object recognition for automated editing and research applications.
-
NVIDIA Releases Makani Library for AI Weather Climate Modeling
By
–
I am so excited to see that the long-planned release of the Makani library https://
github.com/NVIDIA/makani for #AI based weather and climate modeling has finally happened! It started with our FourCastNet work almost three years ago using the Fourier neural operator, where we showed for -
Data Wrangling Functions for Education Research
By
–
Cghlewis/data-wrangling-functions: Functions used for wrangling education research data https://
bit.ly/3PKAzTo #AI #MachineLearning #DeepLearning #LLMs #DataScience