deepai.org
Supervised Dimensionality Reduction and Classification with Convolutional Autoencoders
20+ min · · 08/25/22 - The joint optimization of the reconstruction and classification error is a hard non convex problem, especially when a non linear m... · Shared by 4, including andre hollstein
adept.ai
ACT-1: Transformer for Actions
3 min · · At Adept, we are building the next frontier of models that can take actions in the digital world—that’s why we’re excited to introduce our first large model, Action Transformer (ACT-1). · Shared by 28, including Nando 💪🇦🇺🇦🇷, Gio Tiongson, Nathan Benaich, 瑞拿頭, andre hollstein, 🇺🇦Evan Kirstel #B2B #TechFluencer
deepai.org
An Embarrassingly Simple Approach to Semi-Supervised Few-Shot Learning
· 09/28/22 - Semi-supervised few-shot learning consists in training a classifier to adapt
to new tasks with limited labeled data and a fixed qu... · Shared by 4, including andre hollstein
keras.io
Keras documentation: High-performance image generation using Stable Diffusion in KerasCV
9 min · · Authors: fchollet, lukewood, divamgupta Date created: 2022/09/25 Last modified: 2022/09/25 Description: Generate new images using KerasCV's StableDiffusion model. View in Colab • GitHub source… · Shared by 18, including Real Python, andre hollstein
jalammar.github.io
The Illustrated Stable Diffusion
9 min · · AI image generation is the most recent AI capability blowing people’s minds (mine included). The ability to create striking visuals from text descriptions has a magical quality to it and points… · Shared by 87, including Matt Shaffer, Vikram Dutt, Colin Wright, Mike Tamir, PhD, andre hollstein, Jesús Torres 👩💻, Sebastian Raschka, AGI.Eth | MTL.Eth, James Young, hardmaru, Lynn Cherny, Nando de Freitas 🏳️🌈, Robert Scoble
nature
Discovering faster matrix multiplication algorithms with reinforcement learning
20+ min · · A reinforcement learning approach based on AlphaZero is used to discover efficient and provably correct algorithms for matrix multiplication, finding faster algorithms for a variety of matrix sizes. · Shared by 391, including Yann “不停” Heurtaux ⏚ @shalf@mastodon.social, Marcel Salathé, AGI.Eth | MTL.Eth, Sebastian Raschka, Demis Hassabis, Dr. Ganapathi Pulipaka 🇺🇸, Eric Topol, Nico Müller 🇺🇦, andre hollstein
outerbounds.com
Introducing Metaflow Resources for Data Science and Engineering
4 min · · A knowledge base of free, practical data science and machine learning materials, including how to set up a modern data science infrastructure. · Shared by 4, including andre hollstein
AI Papers
Elucidating the Design Space of Diffusion-Based Generative Models
1 min · · We argue that the theory and practice of diffusion-based generative models are currently unnecessarily convoluted and seek to remedy the situation by presenting a design space that clearly separates… · Shared by 10, including andre hollstein, Yisong Yue
sites.google.com
SERAC: Memory-based Model Editing at Scale
1 min · · SERAC: Memory-based Model Editing at Scale Eric Mitchell, Charles Lin, Antoine Bosselut, Christopher D. Manning, Chelsea Finn paper | code coming soon! | interactive demo A computationally efficient… · Shared by 8, including andre hollstein, Stanford NLP Group
OpenAI
Techniques for Training Large Neural Networks
7 min · · Large neural networks are at the core of many recent advances in AI, but training them is a difficult engineering and research challenge which requires orchestrating a cluster of GPUs to perform a… · Shared by 154, including andre hollstein, Nico Müller 🇺🇦, Nando de Freitas 🏳️🌈, Dr. Ganapathi Pulipaka 🇺🇸, Nicolas Babin, Christine McLeavey, Nando 💪🇦🇺🇦🇷, Elena Neira, ipfconline in vacation mode :) will be back Sept.4, Sam Altman
cs.cornell.edu
ARF: Artistic Radiance Fields
· 1Cornell University 2Adobe Research · Shared by 4, including andre hollstein
minedojo.org
MineDojo | Building Open-Ended Embodied Agents with Internet-Scale Knowledge
2 min · · MineDojo is a new framework built on the popular Minecraft game for embodied agent research. MineDojo features a simulation suite with 1000s of open-ended and language-prompted tasks, where the AI… · Shared by 11, including andre hollstein, AGI.Eth | MTL.Eth, Nando de Freitas 🏳️🌈
Noema Magazine
What AI Can Tell Us About Intelligence
10 min · · Can deep learning systems learn to manipulate symbols? The answers might change our understanding of how intelligence works and what makes humans unique. · Shared by 220, including andre hollstein, Kevin Mitchell @WiringtheBrain@ mstdn .social, Terence Mills, Paul Roetzer, Bob E. Hayes, Dorothea Baur (Dr.), Dr. Marigo Raftopoulos, Sebastien Meunier, Yann LeCun
AI Papers
Contrastive Learning of Medical Visual Representations from Paired Images and Text
1 min · · Learning visual representations of medical images is core to medical image understanding but its progress has been held back by the small size of hand-labeled datasets. Existing work commonly relies… · Shared by 9, including andre hollstein, Alexander Seifert, Stanford NLP Group
PyImageSearch
Building image pairs for siamese networks with Python
16+ min · · In this tutorial you will learn how to build image pairs for training siamese networks. We’ll implement our image pair generator using Python so that you can use the same code, regardless of whether… · Shared by 8, including andre hollstein
TensorFlow
TensorFlow Addons Losses: TripletSemiHardLoss
2 min · · This notebook will demonstrate how to use the TripletSemiHardLoss function in TensorFlow Addons. Resources: TripletLoss As first introduced in the FaceNet paper, TripletLoss is a loss function that… · Shared by 2, including andre hollstein