neural network – Page 20 – Andrew Fairless, Ph.D.

What I Read: Identifying Document Types

By Andrew Fairless on August 11, 2021July 22, 2021

https://tech.scribd.com/blog/2021/identifying-document-types.html Identifying Document Types at ScribdJonathan RamkissoonJuly 12, 2021 “User-uploaded documents have been a core component of Scribd’s business from the very beginning, understanding what is actually in the documentContinue readingWhat I Read: Identifying Document Types

What I Read: Prompting, Language Models, NLP

By Andrew Fairless on August 9, 2021July 22, 2021

https://thegradient.pub/prompting/ Prompting: Better Ways of Using Language Models for NLP Tasks03.Jul.2021Tianyu Gao “The giant model size of GPT-3 is an important factor for its success, while the concept of promptsContinue readingWhat I Read: Prompting, Language Models, NLP

What I Read: Gradient Pseudo-Swap

By Andrew Fairless on August 3, 2021July 12, 2021

https://deepnote.com/@martin-molnar/Gradient-Pseudo-Swap-F2u_aTbqTOisNYCH_JORwg Martin MolnarGradient Pseudo-Swap “When we have layers in a neural network we want to train with gradient descent, but those layers don’t have smooth gradients that can be used,Continue readingWhat I Read: Gradient Pseudo-Swap

What I Read: SwAV method

By Andrew Fairless on August 2, 2021July 12, 2021

https://m0nads.wordpress.com/2021/01/08/the-swav-method/ The SwAV methodAntonio Ferraioli “The SwAV method is a clustering-based online method whose goal is to learn visual features in an online fashion without supervision.”

What I Read: CNN Heat Maps, Class Activation Mapping

By Andrew Fairless on July 30, 2021July 12, 2021

https://glassboxmedicine.com/2019/06/11/cnn-heat-maps-class-activation-mapping-cam/ CNN Heat Maps: Class Activation Mapping (CAM)Date: June 11, 2019Author: Rachel Draelos “Class Activation Mapping (CAM) is one technique for producing heat maps to highlight class-specific regions of images.”

What I Read: Representation quality, complexity

By Andrew Fairless on July 29, 2021July 12, 2021

https://willwhitney.com/representation-quality-and-the-complexity-of-learning.html Will WhitneyRepresentation quality and the complexity of learning “In the last few years, there’s been an explosion of work on learning good representations of data…. However, defining precisely whatContinue readingWhat I Read: Representation quality, complexity

What I Read: Parallelizing neural networks, GPU, JAX

By Andrew Fairless on July 28, 2021July 12, 2021

https://willwhitney.com/parallel-training-jax.html Will WhitneyParallelizing neural networks on one GPU with JAX How you can get a 100x speedup for training small neural networks by making the most of your accelerator.“Most neuralContinue readingWhat I Read: Parallelizing neural networks, GPU, JAX

What I Read: Better computer vision models, Transformers, CNNs

By Andrew Fairless on July 27, 2021July 12, 2021

https://ai.facebook.com/blog/computer-vision-combining-transformers-and-convolutional-neural-networks/ Better computer vision models by combining Transformers and convolutional neural networksJuly 8, 2021 “We’ve developed a new computer vision model… which combines… convolutional neural networks (CNNs) and Transformer-based models…Continue readingWhat I Read: Better computer vision models, Transformers, CNNs

What I Read: Same or Different? The Question Flummoxes Neural Networks

By Andrew Fairless on July 19, 2021July 2, 2021

https://www.quantamagazine.org/same-or-different-ai-cant-tell-20210623/ Same or Different? The Question Flummoxes Neural Networks.John PavlusContributing WriterJune 23, 2021 “One of the most powerful classes of artificial intelligence systems, known as convolutional neural networks or CNNs…Continue readingWhat I Read: Same or Different? The Question Flummoxes Neural Networks

What I Read: Do Multi-Task Learning Intelligently

By Andrew Fairless on July 15, 2021July 2, 2021

https://thegradient.pub/how-to-do-multi-task-learning-intelligently/ How to Do Multi-Task Learning Intelligently22.Jun.2021 “Traditionally, a single machine learning model is devoted to one task… There are some advantages, however, to training models to make multiple kindsContinue readingWhat I Read: Do Multi-Task Learning Intelligently

Tag: neural network