The book "Pandas. Working with data »

The book "Pandas. Working with data » Hello! We have finally prepared the release of our book " Pandas . Work with data! " Much of her material is devoted to classes Series and DataFrame , which is provided by pandas for storing and manipulating data, the configuration topic is also affected. pandas , work with time series, visualization of data, etc.
 
 
At the moment " Pandas . Working with data " is available only on our website, in the future, we hope that it will be available at large sites selling e-books.
 
Introduction to pandas and its ...
+ 0 -

Computer vision and machine learning in PHP using the library opencv

Computer vision and machine learning in PHP using the library opencvHello. This is my jubilee article on the habr. For almost 7 years I wrote 10 articles (including this one), 8 of them - technical. The total number of views of all articles is about half a million.
 
I made the main contribution to two hubs: PHP and Server Administration. I like to work at the junction of these two areas, but the scope of my interests is much broader.
 
Like many developers, I often use the results of someone else's work (articles on the habr, code on the githaba, ), so I'm always glad to share with the community my results in response. Writing articles is not only a return of ...
+ 0 -

Neural machine translation Google

The report was written in December 2017.
 
It's not who has the best algorithms that wins. It's who has the most data. The winner is not the one who has the better algorithm, but the one who has more data. Andrew Ng, a lecturer in the machine learning course at Coursera.
If you scale up both the size of the model, and you can learn finer distinctions or more complex features. These models can usually take a lot more context. Jeff Dean, an engineer assisting the research at Google. If you increase the size of the model and give it more data for learning, it will begin to distinguish more subtle ...
+ 0 -

Organization of the university course on signal processing

Organization of the university course on signal processingPedagogy interested me a very long time and, for many years, I, being a student, brought up, but at the same time harassed and jerked by the existing training organization, thought about how to improve it. Recently, I have been given an opportunity to check out some of the ideas in practice. In particular, this spring I was given an opportunity to read the course "Signal Processing" at the Polytechnic University (SPbPU). His organization, especially the organization of reporting - the first experiment, the results of which seem to me to be somehow successful, and in this article I want to tell ...
+ 0 -

Open webinar: "Neural networks for pharmaceutical problems"

Hello!
 
 
In this open lesson, Arthur Kadurin tells you how neural networks help save lives.
 
 
Also we suggest you to learn more about our Machine Learning course at The day of the open doors Today at 20:00 Moscow time!
 
 
Hello!  In this open lesson, Arthur Kadurin tells you how neural networks help save lives.  Also we suggest you to learn more about our Machine Learning course at The day of the open doors Today at 20:00 Moscow time!   ...
+ 0 -

Kaggle: Amazon from Space - tricks and khaki for learning neural networks

Kaggle: Amazon from Space - tricks and khaki for learning neural networks  
 
Last summer, was over. competition on the kaggle site, which was devoted to the classification of satellite imagery of the Amazon forests. Our team took 7th place out of 900+ participants. Despite the fact that the competition ended a long time ago, almost all the methods of our decision are applicable so far, not only for competitions, but also for training neural networks for selling. For details under cat.
 
albu [/i]
 
At the time of participation in the competition, he directed all ml of the direction in the company Geoscan. But since then ...
+ 0 -

AI created his first film, and this is a horror film

AI created his first film, and this is a horror film  
 
Artificial intelligence led by an engineer from Google independently made the first motion picture. Not just folded some of the foreign films with each other, or gave out the script's letters, but created a full-length short film. He wrote the story himself, wrote all the dialogues himself, chose the scenes and determined the expressions of actors' faces (and even, apparently, he wrote the headline for the newspaper shown in the film - although it turned out to be so ironic that it's hard to believe). The voice of characters, music and editing - of course, also for AI. But the resulting black and white ...
+ 0 -

A review of gradient methods in problems of mathematical optimization

A review of gradient methods in problems of mathematical optimization

Foreword


 
In this article, we will talk about methods for solving mathematical optimization problems based on the use of the gradient of a function. The main goal is to collect in the article all the most important ideas that are somehow connected with this method and its various modifications.
 
 

Note from the author


 
At the time of writing, I defended my thesis, the task of which required me to have a deep understanding of theoretical fundamentally methods of mathematical optimization. Nevertheless, I still (like everyone else) have their eyes blown away from the terrible long formulas...[/h][/h]
+ 0 -

Counting the bees by neural network on Raspberry Pi

Published May 1? 2018
 
 
Immediately after the installation of the beehive I thought: "I wonder how to calculate the number of arriving and departing bees?"
 
 
A small study has shown: it seems that until now no one has devised a good non-invasive system that solves this problem. But it would probably be useful to have such information to check the health of the hive.
 
 
First, you need to collect sample data. Raspberry Pi, standard Pi camera and solar panel: this simple equipment is enough to record one frame every 10 seconds and save 5000+ images per day (from 6am to 9pm).
 
 
Counting the bees by neural network on Raspberry Pi ...
+ 0 -

Preliminary program PyConRu-2018: three Python Core Developer'a, speakers from Google, Yelp, Red Hat, Yandex

Preliminary program PyConRu-2018: three Python Core Developer'a, speakers from Google, Yelp, Red Hat, YandexHello, Habr!
 
 
July 22-23 in 95 km. from Moscow there will pass the sixth Russian conference for python-programmers PYCON RUSSIA 2018 .
 
 
Already in the program: Yuri Selivanov (Python Core Developer, EgdeDB, Canada), Andrey Svetlov (Python Core Developer, Ukraine), Christian Heimes (Python Core Developer, Red Hat, Germany), Melanie Warrick (Google, USA), Stephan Jaensch Yelp, Germany), Kate Heddleston (Shift, USA), Alejandro Saucedo (Eigen Technologies, United Kingdom), Vadim Pushtaev (Mail.Ru), Marina Kamalova (Yandex). If you want to join as a speaker, there are a few more days to ...
+ 0 -