🎥 Машинное обучение 12. RNN
👁 1 раз ⏳ 5060 сек.
Лекция 30.04.2019
Лектор: Радослав Нейчев

Снимал: Александр Гришутин
Монтировал: Александр Васильев

🎥 Занятие 8 | Машинное обучение
👁 1 раз ⏳ 2158 сек.
Преподаватель: Власов Кирилл Вячеславович
Материалы курса: https://github.com/ml-dafe/ml_mipt_dafe_minor
Дата: 27.04.2019

​GitHub Autocompletion with Machine Learning

🔗 GitHub Autocompletion with Machine Learning
Written by Óscar D. Lara Yejas and Ankit Jha

​Comprehensive Introduction to Turing Learning and GANs: Part 1

🔗 Comprehensive Introduction to Turing Learning and GANs: Part 1
Want to turn horses into zebras? Make DIY anime characters or celebrities? Generative adversarial networks (GANs) are your new best friend.

​Dataset bridges human vision and machine learning

🔗 Dataset bridges human vision and machine learning
Neuroscientists and computer vision scientists say a new dataset of unprecedented size—comprising brain scans of four volunteers who each viewed 5,000 images—will help researchers better understand ...

🎥 Deep Neural Networks step by step final prediction model #part 5
👁 1 раз ⏳ 1737 сек.
In this tutorial we will use the functions we had implemented in the previous parts to build a deep network, and apply it to cat vs dog classification. Hopefully, we will see an improvement in accuracy relative to our previous logistic regression implementation. After this part we will be able to build and apply a deep neural network to supervised learning using only numpy library.

Full tutorial code and cats vs dogs image data-set can be found on my GitHub page: https://github.com/pythonlessons/Logistic-r

🎥 Deep Learning and Modern NLP - Zachary S Brown
👁 1 раз ⏳ 5393 сек.
In this tutorial, we’ll cover the fundamental building blocks of neural network architectures and how they are utilized to tackle problems in modern natural language processing. Topics covered will include an overview of language vector representations, text classification, named entity recognition, and sequence to sequence modeling approaches. An emphasis will be placed on the shape of these types of problems from the perspective of deep learning architectures. This will help to develop an intuition for id

​DeepMind's AI Learned a Better Understanding of 3D Scenes

🔗 DeepMind's AI Learned a Better Understanding of 3D Scenes
Backblaze: https://www.backblaze.com/cloud-backup.html#af9tk4 📝 The paper "MONet: Unsupervised Scene Decomposition and Representation" is available here: htt...

​How is Data Science Changing the World?

🔗 How is Data Science Changing the World?
In this article, you will go through the role that a Data Scientist plays. There is a veil of mystery surrounding Data Science. While the…

🎥 Tutorial 2019 || Cryptocurrency-predicting RNN intro - Deep Learning w/ Python, TensorFlow and Keras
👁 1 раз ⏳ 1204 сек.
Tutorial 2019 || Cryptocurrency-predicting RNN intro - Deep Learning w/ Python, TensorFlow and Keras p.8

🎥 Introduction to Deep Learning Recitation 12
👁 1 раз ⏳ 3180 сек.

🎥 Introduction to Deep Learning Lecture 29
👁 1 раз ⏳ 4704 сек.

🎥 Leveraging NLP and Deep Learning for Document Recommendations in the CloudGuoqiong Song Intel
👁 1 раз ⏳ 1254 сек.
Efficient recommender systems are critical for the success of many industries, such as job recommendation, news recommendation, ecommerce, etc. This talk will illustrate how to build an efficient document recommender system by leveraging Natural Language Processing(NLP) and Deep Neural Networks (DNNs). The end-to-end flow of the document recommender system is build on AWS at scale, using Analytics Zoo for Spark and BigDL. The system first processes text rich documents into embeddings by incorporating Global

🎥 Introduction to Deep Learning Lecture 28
👁 1 раз ⏳ 4406 сек.

​Наш телеграм канал - tglink.me/ai_machinelearning_big_data

https://towardsdatascience.com/the-digital-survivors-dc3311cb9602?source=collection_home---4------4---------------------

🔗 The Digital Survivors
“Soon, there will be no survivors to tell their stories, and your children would hear about the Holocaust only by books and videos”

​Neural Networks and Fibonacci Numbers

🔗 Neural Networks and Fibonacci Numbers
Does Fibonacci Initialized ANN overpowers our standard random initialization approach ? I don’t know !! Let’s experiment and see…

​TensorFlow Tutorials - YouTube

🔗 TensorFlow Tutorials - YouTube

​Representing music with Word2vec?

🔗 Representing music with Word2vec?
Machine learning algorithms have transformed the field of vision and NLP. But what about music? These last few years, the field of music…

🎥 TWiML x Fast ai v3 Deep Learning Part 2 Study Group - Lesson 12 - Spring 2019 1080p
👁 1 раз ⏳ 5738 сек.
**SUBSCRIBE AND TURN ON NOTIFICATIONS** **twimlai.com**

This video is a recap of our TWiML Online Fast.ai Deep Learning Part 2 Study Group.

In this session, we had a mini presentation on Deep Representation Learning for Trigger Monitoring and a review Lesson 12 of the Fast.ai v3 Deep Learning Part 2 course.

It’s not too late to join the study group. Just follow these simple steps:

1. Head over to twimlai.com/meetup, and sign up for the programs you're interested in, including either of the Fast.ai study

​Digging Into Self-Supervised Monocular Depth Estimation

https://arxiv.org/abs/1806.01260

https://github.com/nianticlabs/monodepth2

🔗 Digging Into Self-Supervised Monocular Depth Estimation
Per-pixel ground-truth depth data is challenging to acquire at scale. To overcome this limitation, self-supervised learning has emerged as a promising alternative for training models to perform monocular depth estimation. In this paper, we propose a set of improvements, which together result in both quantitatively and qualitatively improved depth maps compared to competing self-supervised methods. Research on self-supervised monocular training usually explores increasingly complex architectures, loss functions, and image formation models, all of which have recently helped to close the gap with fully-supervised methods. We show that a surprisingly simple model, and associated design choices, lead to superior predictions. In particular, we propose (i) a minimum reprojection loss, designed to robustly handle occlusions, (ii) a full-resolution multi-scale sampling method that reduces visual artifacts, and (iii) an auto-masking loss to ignore training pixels that violate camera motion assumptions. We demonstrate t