​The Love Machine(Learning)

🔗 The Love Machine(Learning)
“The machines have probably already planned our wedding and everything”

​Бег с протезами: некстген симуляция движения человека с помощью мышц, костей и нейросети

Сотрудники Сеульского университета опубликовали исследование о симуляции движения двуногих персонажей на основе работы суставов и мышечных сокращений, использующей нейросеть с Deep Reinforcement Learning. Под катом перевод краткого обзора.
https://habr.com/ru/company/pixonic/blog/459208/

🔗 Бег с протезами: некстген симуляция движения человека с помощью мышц, костей и нейросети
Сотрудники Сеульского университета опубликовали исследование о симуляции движения двуногих персонажей на основе работы суставов и мышечных сокращений, использующ...

​24 Ultimate Data Science Projects To Boost Your Knowledge and Skills (& can be accessed freely)

🔗 24 Ultimate Data Science Projects To Boost Your Knowledge and Skills (& can be accessed freely)
This article list data science projects, taken from various open source data sets solving regression, classification, text mining, clustering

​Point-Voxel CNN for Efficient 3D Deep Learning

Authors: Zhijian Liu, Haotian Tang, Yujun Lin, Song Han

Abstract: We present Point-Voxel CNN (PVCNN) for efficient, fast 3D deep learning. Previous work processes 3D data using either voxel-based or point-based NN models. However, both approaches are computationally inefficient.
https://arxiv.org/abs/1907.03739

🔗 Point-Voxel CNN for Efficient 3D Deep Learning
We present Point-Voxel CNN (PVCNN) for efficient, fast 3D deep learning. Previous work processes 3D data using either voxel-based or point-based NN models. However, both approaches are computationally inefficient. The computation cost and memory footprints of the voxel-based models grow cubically with the input resolution, making it memory-prohibitive to scale up the resolution. As for point-based networks, up to 80% of the time is wasted on structuring the irregular data which have rather poor memory locality, not on the actual feature extraction. In this paper, we propose PVCNN that represents the 3D input data in points to reduce the memory consumption, while performing the convolutions in voxels to largely reduce the irregular data access and improve the locality. Our PVCNN model is both memory and computation efficient. Evaluated on semantic and part segmentation datasets, it achieves much higher accuracy than the voxel-based baseline with 10x GPU memory reduction; it also outperforms the state-of-the-ar

🎥 Topcoder Neptune - Facial Detection & Re-Identification Marathon Match – Мирас Амир
👁 1 раз ⏳ 1550 сек.
Мирас Амир рассказывает про решения двух контестов на платформе Topcoder: March Madness Series: Neptune - Facial Detection Marathon Match и Neptune - Facial Re-Identification Marathon Match. В каждом соревновании он занял второе место, решив задачи по детекции и реидентификации лиц. Из видео вы сможете узнать:
- Про формат соревнований
- Особенности датасета
- Подробности решений первого и второго места

Узнать о текущих соревнованиях можно на сайте http://mltrainings.ru/

Узнать о новых тренировках и видео

​ResNeXt models pre-trained on Instagram hashtags stand out in their
in their ability to generalized to the 'ImageNetV2' test set
#PyTorch

https://colab.research.google.com/github/rwightman/pytorch-image-models/blob/master/notebooks/GeneralizationToImageNetV2.ipynb/

🔗 Google Colaboratory

🎥 Beyond the Hype. Real Companies Doing Real Business with AI - Alyssa Rochwerger | ODSC West 2018
👁 1 раз ⏳ 1796 сек.
AI - everyone is talking about it but who is actually doing it (and generating business results). This session takes an industry by industry perspective on true AI adoption disambiguating the hype from the reality, the theoretical from the practical and the research labs from ROI.
This presentation provides:
Showcase companies getting actual real value from leveraging artificial intelligence and discuss ideas around how any company, from SMB to enterprise, can use artificial intelligence within their own bu

​Predicting the Generalization Gap in Deep Neural Networks

Наш телеграм канал - tglink.me/ai_machinelearning_big_data

🔗 Predicting the Generalization Gap in Deep Neural Networks
Posted by Yiding Jiang, Google AI Resident Deep neural networks (DNN) are the cornerstone of recent progress in machine learning, and ...

🎥 Teaching a Machine to Code
👁 1 раз ⏳ 2565 сек.
At Prodo.AI, we’re teaching machines to write code for humans. Using the latest in Deep Learning techniques, we can generate code that’s not just functional, but beautiful. Our goal is to make the computer do the heavy lifting so you can concentrate on the important things: being creative, solving problems, and having fun.We’ve tried a hundred different ways of encoding the knowledge of how to write code. In this talk, Samir will take you through a tour of the different techniques, architectures and optimis

​Keras Callbacks Explained

🔗 Keras Callbacks Explained
A gentle introduction to callbacks in Keras. Learn about EarlyStopping, ModelCheckpoint, and other callback functions with code examples.

​Web Scraping news articles in Python

🔗 Web Scraping news articles in Python
Building a web scraping application in Python made simple

​Knowledge Quadrant for Machine Learning - Towards Data Science

🔗 Knowledge Quadrant for Machine Learning - Towards Data Science
Transfer Learning, Uncertainty Sampling, and Diversity Sampling to improve your Machine Learning models.

​Online poker — When’s the Money?

🔗 Online poker — When’s the Money?
The best time to play and how much it’s worth

​10 фич для ускорения анализа данных в Python
https://habr.com/ru/post/457302/

🔗 10 фич для ускорения анализа данных в Python
Источник Советы и рекомендации, особенно в программировании, могут быть очень полезны. Маленький шоткат, аддон или хак может сэкономить кучу времени и серьёзно у...

​BTGym
https://github.com/notadamking/RLTrader/

Scalable event-driven RL-friendly backtesting library. Build on top of Backtrader with OpenAI Gym environment API.
Backtrader is open-source algorithmic trading library:
GitHub: http://github.com/mementum/backtrader
Documentation and community:
http://www.backtrader.com/
OpenAI Gym is..., well, everyone knows Gym:
GitHub: http://github.com/openai/gym
Documentation and community:
https://gym.openai.com/

🔗 notadamking/RLTrader
A profitable cryptocurrency trading environment using deep reinforcement learning and OpenAI's gym - notadamking/RLTrader

​Kaggle: Google AI Open Images – Константин Гаврильчик

🔗 Kaggle: Google AI Open Images – Константин Гаврильчик
Константин Гаврильчик рассказывает про соревнования Kaggle: Google AI Open Images, задача которых заключалась в детектировании объектов и связей между ними на изображениях. Решение Константина принесло серебряную медаль. Из видео вы сможете узнать: - Какая метрика использовалась - Особенности разметки и датасета - Подходы к решению и наилучшие решения Узнать о текущих соревнованиях можно на сайте http://mltrainings.ru/ Узнать о новых тренировках и видео можно из групп: ВКонтакте https://vk.com/mltrainin

​Sarcasm Detection: Step towards Sentiment Analysis - DIGVIJAY SINGH - Medium

🔗 Sarcasm Detection: Step towards Sentiment Analysis - DIGVIJAY SINGH - Medium
Humans have a social nature. Social nature means that we interact with each other in positive, friendly ways, and it also means that we…

​Sean Carroll: The Nature of the Universe, Life, and Intelligence | Artificial Intelligence Podcast

🔗 Sean Carroll: The Nature of the Universe, Life, and Intelligence | Artificial Intelligence Podcast
Sean Carroll is a theoretical physicist at Caltech, specializing in quantum mechanics, gravity, and cosmology. He is the author of several popular books: one on the arrow of time called From Eternity to Here, one on the Higgs boson called The Particle at the End of the Universe, and one on science and philosophy called The Big Picture: On the Origins of Life, Meaning, and the Universe Itself. He has an upcoming book on Quantum Mechanics that you can preorder now called Something Deeply Hidden. He writes one

Deep Learning

Наш телеграм канал - tglink.me/ai_machinelearning_big_data
AI и естественный язык
Краткая история криптографии
Курс по криптографии. Вводное занятие
Курс по криптографии. Симметричные шифры

#video

🎥 Дмитрий Коробченко: Deep Learning
👁 2 раз ⏳ 6062 сек.
Руководитель проектов и инженер в области машинного обучения, компьютерного зрения и обработки сигналов в Исследовательском Центре Samsung Дмитрий ...

🎥 Сергей Марков: AI и естественный язык
👁 1 раз ⏳ 7969 сек.
1 июня 2016 в антикафе «Кочерга» (http://kocherga-club.ru/) автор одной из сильнейших российских
шахматных программ, специалист по методам машинног...

🎥 Сергей Владимиров: Краткая история криптографии
👁 1 раз ⏳ 8259 сек.
8 сентября 2016 Сергей Владимиров рассказал в Кочерге (http://kocherga-club.ru ) об истории криптографии.

Сергей Владимиров (https://vk.com/vlser...

🎥 Дмитрий Яхонтов: Курс по криптографии. Вводное занятие
👁 1 раз ⏳ 3940 сек.
Курс расскажет об основных понятиях современной криптографии. Будут рассмотрены протоколы шифрования, проверки подлинности, обмена ключами, а также...

🎥 Дмитрий Яхонтов: Курс по криптографии. Симметричные шифры
👁 1 раз ⏳ 4271 сек.
Тема лекции - симметричные шифры:
- Основные принципы построения криптосистем.
- Протоколы, использующие один и тот же ключ для шифрования и для р...

​Blind Universal Bayesian Image Denoising with Gaussian Noise Level Learning. arxiv.org/abs/1907.03029

🔗 Blind Universal Bayesian Image Denoising with Gaussian Noise Level Learning
Blind and universal image denoising consists of a unique model that denoises images with any level of noise. It is especially practical as noise levels do not need to be known when the model is developed or at test time. We propose a theoretically-grounded blind and universal deep learning image denoiser for Gaussian noise. Our network is based on an optimal denoising solution, which we call fusion denoising. It is derived theoretically with a Gaussian image prior assumption. Synthetic experiments show our network's generalization strength to unseen noise levels. We also adapt the fusion denoising network architecture for real image denoising. Our approach improves real-world grayscale image denoising PSNR results by up to $0.7dB$ for training noise levels and by up to $2.82dB$ on noise levels not seen during training. It also improves state-of-the-art color image denoising performance on every single noise level, by an average of $0.1dB$, whether trained on or not.