​Log Book — Guide to Distance Measuring Approaches for K- Means Clustering machine learning

https://towardsdatascience.com/log-book-guide-to-distance-measuring-approaches-for-k-means-clustering-f137807e8e21?source=collection_home---4------1-----------------------

🔗 Log Book — Guide to Distance Measuring Approaches for K- Means Clustering
In this guide I have tried to cover the different types and features of distances that can be used in K-Means Clustering

​P-values Explained By Data Scientist - Towards Data Science

🔗 P-values Explained By Data Scientist - Towards Data Science
For Data Scientists

Chainer: A Deep Learning Framework for Fast Research & Applications | SciPy 2019 | Seiya Tokui

https://www.youtube.com/watch?v=w2n4hJWi4qA

🎥 Chainer: A Deep Learning Framework for Fast Research & Applications | SciPy 2019 | Seiya Tokui
👁 1 раз ⏳ 1842 сек.
Chainer is a deep learning framework for flexible and intuitive coding of high performance experiments and applications. It is designed to maximize the trial-and-error speed with its Define-by-Run paradigm, which provides Pythonic programming of auto-differentiated neural networks. The framework can accelerate performance with multiple GPUs in distributed environments and add-on packages enable quickly jumping into specific domains. In this talk, we introduce the abstract of Chainer’s API, its capabilities

​Cheat sheet: TensorFlow, an open source software library for machine learning
https://www.techrepublic.com/article/tensorflow-googles-open-source-software-library-for-machine-learning-the-smart-persons-guide/ …
#machinelearning

🔗 Cheat sheet: TensorFlow, an open source software library for machine learning
TensorFlow is an open source software library developed by Google for numerical computation with data flow graphs. This TensorFlow guide covers why the library matters, how to use it, and more.

​Как распознать текст с фото: новые возможности фреймворка Vision

Теперь фреймворк Vision умеет распознавать текст по-настоящему, а не как раньше. С нетерпением ждём, когда сможем применить это в Dodo IS. А пока перевод статьи о распознавании карточек из настольной игры Magic The Gathering и извлечении из них текстовой информации.

#Машинноеобучение

🔗 Как распознать текст с фото: новые возможности фреймворка Vision
Теперь фреймворк Vision умеет распознавать текст по-настоящему, а не как раньше. С нетерпением ждём, когда сможем применить это в Dodo IS. А пока перевод статьи...

​Improving Deep Lesion Detection Using 3D Contextual and Spatial Attention
https://deepai.org/publication/improving-deep-lesion-detection-using-3d-contextual-and-spatial-attention by Qingyi Tao et al.
#ComputerVision #PatternRecognition

🔗 Improving Deep Lesion Detection Using 3D Contextual and Spatial Attention
07/09/19 - Lesion detection from computed tomography (CT) scans is challenging compared to natural object detection because of two major reas...

​NVIDIA + BERT = 🔥
BERT — нейросеть для обработки естественного языка (Natural Language Processing, NLP). Если вы давно мечтали создать свою виртуальную Алису или Олега, то у нас хорошие новости: не так давно NVIDIA выложила в открытый доступ скрипты, позволяющие использовать BERT для рекомендательных систем и приложений «вопрос-ответ». Мы расскажем, в чём преимущество этой нейросети и как её обучить для конкретных задач.

https://www.reg.ru/blog/nvidia-plus-bert-is-fire/

🔗 NVIDIA + BERT = 🔥 AI да GPU – Блог REG.RU
BERT — нейросеть для обработки естественного языка (Natural Language Processing, NLP). Если вы давно мечтали создать свою виртуальную Алису или Олега, то у нас хорошие новости: не так давно NVIDIA выложила в открытый доступ скрипты, позволяющие использовать BERT для рекомендательных систем и приложений «вопрос-ответ». Мы расскажем, в чём преимущество этой нейросети и как её обучить для конкретных задач. В конце прошлого года команде NVIDIA удалось достичь четырёхкратного …

​Adversarial Objects против автономных систем вождения на основе LiDAR
https://arxiv.org/abs/1907.05418

🔗 Adversarial Objects Against LiDAR-Based Autonomous Driving Systems
Deep neural networks (DNNs) are found to be vulnerable against adversarial examples, which are carefully crafted inputs with a small magnitude of perturbation aiming to induce arbitrarily incorrect predictions. Recent studies show that adversarial examples can pose a threat to real-world security-critical applications: a "physical adversarial Stop Sign" can be synthesized such that the autonomous driving cars will misrecognize it as others (e.g., a speed limit sign). However, these image-space adversarial examples cannot easily alter 3D scans of widely equipped LiDAR or radar on autonomous vehicles. In this paper, we reveal the potential vulnerabilities of LiDAR-based autonomous driving detection systems, by proposing an optimization based approach LiDAR-Adv to generate adversarial objects that can evade the LiDAR-based detection system under various conditions. We first show the vulnerabilities using a blackbox evolution-based algorithm, and then explore how much a strong adversary can do, using ou

Hyperparameter Tuning and Experimenting - Training Deep Neural Networks
https://www.youtube.com/watch?v=ycxulUVoNbk

🎥 Hyperparameter Tuning and Experimenting - Training Deep Neural Networks
👁 1 раз ⏳ 709 сек.
Welcome to this neural network programming series. In this episode, we will see how we can use TensorBoard to rapidly experiment with different training hyperparameters to more deeply understand our neural network.

We'll learn how to uniquely identify each run by building and passing a comment string to the SummeryWriter constructor that will be appended to the auto-generated file name.

We'll learn how to use a Cartesian product to create a set of hyper parameters to try, and at the end, we'll consider h

​Пример простой нейросети, как результат разобраться что к чему

Нейросети — это та тема, которая вызывает огромный интерес и желание разобраться в ней. Но, к сожалению, поддаётся она далеко не каждому. Когда видишь тома непонятной литературы, теряешь желание изучить, но всё равно хочется быть в курсе происходящего.

В конечном итоге, как мне показалось, нет лучше способа разобраться, чем просто взять и создать свой маленький проект.
#Машинноеобучение,
#Искусственныйинтеллект
https://habr.com/ru/post/459822/

🔗 Пример простой нейросети, как результат разобраться что к чему
Нейросети — это та тема, которая вызывает огромный интерес и желание разобраться в ней. Но, к сожалению, поддаётся она далеко не каждому. Когда видишь тома непон...

Object Detection with Python, #Tensorflow and ImageAI within 5 minutes

https://www.youtube.com/watch?v=SazWuIIxwZc

🎥 Object Detection with Python, Tensorflow and ImageAI within 5 minutes
👁 1 раз ⏳ 363 сек.
Object Detection like Human, By-cycle, moto-cycle, truck etc.

Install the dependencies:

1) Download and install Python 3 from official Python Language website
https://python.org

2) Install the following dependencies via pip:
i. Tensorflow
pip install tensorflow

ii. Numpy
pip install numpy

iii. SciPy
pip install scipy

iv. OpenCV
pip install opencv-python

v. Pillow
pip install pillow

vi. Matplotlib
pip install matplotlib

vii. H5py
pip install h5py

viii. Keras
pip install keras

ix. ImageAI
pip3 ins

​Recent Advances in Modern Computer Vision
Computer Vision beyond object classification

🔗 Recent Advances in Modern Computer Vision - Towards Data Science
Computer Vision beyond object classification

​Maximizing group happiness in White Elephants using the Hungarian optimal assignment algorithm
Learn about the Hungarian (Munkres) optimal assignment algorithm with Python code to maximize group over individual preferences.

🔗 Maximizing group happiness in White Elephants using the Hungarian optimal assignment algorithm
Learn about the Hungarian (Munkres) optimal assignment algorithm with Python code to maximize group over individual preferences.

Введение в Deep Learning. "Копаем глубже"

🎥 10. Введение в Deep Learning. "Копаем глубже"
👁 2 раз ⏳ 5142 сек.
- Предпосылки.
- Проблемы глубоких сетей.
- Предобучение, ограниченные машины Больцмана.
- Нормализованная инициализация весов.
- ReLU активация.
- Проклятье размерности.
- Dropout регуляризация.

Введение в CNN. Учим нейронные сети видеть

🎥 11. Введение в CNN. Учим нейронные сети видеть
👁 2 раз ⏳ 4966 сек.
- Идея свёрточных нейронных сетей
- Constitutional layer, polling layer
- Обзор архитектур свёрточных сетей
- Transfer learning

​4 Underrated Great Resources For Learning Data Science That Might Surprise You
Getting tired of online MOOCs? Or simply just have a roadblock? This article is for you.
https://towardsdatascience.com/4-underrated-great-resources-for-learning-data-science-that-might-surprise-you-19316b79efbb?source=collection_home---4------2-----------------------

🔗 4 Underrated Great Resources For Learning Data Science That Might Surprise You
Getting tired of online MOOCs? Or simply just have a roadblock? This article is for you.

Applied Deep Learning with #PyTorch - Full Course

https://www.youtube.com/watch?v=CNuI8OWsppg

🎥 Applied Deep Learning with PyTorch - Full Course
👁 1 раз ⏳ 20404 сек.
In this course you will learn the key concepts behind deep learning and how to apply the concepts to a real-life project using PyTorch and Python.

You'll learn the following:
⌨️ RNNs and LSTMs
⌨️ Sequence Modeling
⌨️ PyTorch
⌨️ Building a Chatbot in PyTorch

⭐️Requirements ⭐️
⌨️ Some Basic High School Mathematics
⌨️ Some Basic Programming Knowledge
⌨️ Some basic Knowledge about Neural Networks

⭐️Contents ⭐️
⌨️ (0:00:08) Recurrent Nerual Networks - RNNs and LSTMs
⌨️ (0:35:54) Sequence-To-Sequence Models
⌨️

Rekko Challenge: построение рекомендаций для онлайн кинотеатра
https://www.youtube.com/watch?v=-eCr1K9lKxg&t=6s

🎥 Rekko Challenge: построение рекомендаций для онлайн кинотеатра – Евгений Смирнов
👁 1 раз ⏳ 1605 сек.
Евгений Смирнов рассказывает про то, как можно построить гибридную рекомендательную систему на примере соревнования Rekko Challenge, где он занял второе место. Из видео вы сможете узнать:
- Описание данных и задачи соревнования, интересные факты
- Какая метрика использовалась
- Подробности решения второго места
- Какие признаки использовались, а какие не зашли

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

Узнать о новых тренировках и видео можно из групп:
ВКонтакте https://vk.com

​R-Transformer: Recurrent Neural Network Enhanced Transformer "empirical results show that R-Transformer outperforms the state-of-the-art methods by a large margin in most of the tasks"
https://arxiv.org/abs/1907.05572
https://github.com/DSE-MSU/R-transformer

🔗 R-Transformer: Recurrent Neural Network Enhanced Transformer
Recurrent Neural Networks have long been the dominating choice for sequence modeling. However, it severely suffers from two issues: impotent in capturing very long-term dependencies and unable to parallelize the sequential computation procedure. Therefore, many non-recurrent sequence models that are built on convolution and attention operations have been proposed recently. Notably, models with multi-head attention such as Transformer have demonstrated extreme effectiveness in capturing long-term dependencies in a variety of sequence modeling tasks. Despite their success, however, these models lack necessary components to model local structures in sequences and heavily rely on position embeddings that have limited effects and require a considerable amount of design efforts. In this paper, we propose the R-Transformer which enjoys the advantages of both RNNs and the multi-head attention mechanism while avoids their respective drawbacks. The proposed model can effectively capture both local structures and global

​A General Decoupled Learning Framework for Parameterized Image Operators

https://arxiv.org/abs/1907.05852

🔗 A General Decoupled Learning Framework for Parameterized Image Operators
Many different deep networks have been used to approximate, accelerate or improve traditional image operators. Among these traditional operators, many contain parameters which need to be tweaked to obtain the satisfactory results, which we refer to as parameterized image operators. However, most existing deep networks trained for these operators are only designed for one specific parameter configuration, which does not meet the needs of real scenarios that usually require flexible parameters settings. To overcome this limitation, we propose a new decoupled learning algorithm to learn from the operator parameters to dynamically adjust the weights of a deep network for image operators, denoted as the base network. The learned algorithm is formed as another network, namely the weight learning network, which can be end-to-end jointly trained with the base network. Experiments demonstrate that the proposed framework can be successfully applied to many traditional parameterized image operators. To accelerate the pa