​Network of Networks — A Neural-Symbolic Approach to Inverse-Graphics
One of the first ideas many people have once they get acquainted with deep learning and neural networks
https://towardsdatascience.com/network-of-networks-a-neural-symbolic-approach-to-inverse-graphics-acf3998ab3d?source=collection_home---4------0-----------------------

🔗 Network of Networks — A Neural-Symbolic Approach to Inverse-Graphics
One of the first ideas many people have once they get acquainted with deep learning and neural networks is, “what if we make a network of…

​MixNet: Mixed Depthwise Convolutional Kernels

https://arxiv.org/abs/1907.09595

🔗 MixNet: Mixed Depthwise Convolutional Kernels
Depthwise convolution is becoming increasingly popular in modern efficient ConvNets, but its kernel size is often overlooked. In this paper, we systematically study the impact of different kernel sizes, and observe that combining the benefits of multiple kernel sizes can lead to better accuracy and efficiency. Based on this observation, we propose a new mixed depthwise convolution (MDConv), which naturally mixes up multiple kernel sizes in a single convolution. As a simple drop-in replacement of vanilla depthwise convolution, our MDConv improves the accuracy and efficiency for existing MobileNets on both ImageNet classification and COCO object detection. By integrating MDConv into AutoML search space, we have further developed a new family of models, named as MixNets, which significantly outperform previous models including MobileNetV2 [19] (ImageNet top-1 accuracy +4.2%), ShuffleNetV2 [15] (+3.5%), MnasNet [25] (+1.3%), ProxylessNAS [2] (+2.2%), and FBNet [26] (+2.0%). In particular, our MixNet-L achieves a

​Learning Better Simulation Methods for Partial Differential Equations

Наш телеграм канал - tglink.me/ai_machinelearning_big_data
http://ai.googleblog.com/2019/07/learning-better-simulation-methods-for.html

🔗 Learning Better Simulation Methods for Partial Differential Equations
Posted by Stephan Hoyer, Software Engineer, Google Research The world’s fastest supercomputers were designed for modeling physical pheno...

​Трагедия не приходит одна

В 2016 году уязвимость ImageTragick в библиотеке ImageMagick наделала много шума. Как способ снижения риска предлагалось использовать GraphicsMagick — форк библиотеки ImageMagick, нацеленный на более стабильный и производительный API. Оригинальная уязвимость CVE-2016-3717, обнаруженная stewie, позволяла злоумышленнику прочитать произвольный файл на файловой системе при помощи специально созданного изображения. Сегодня я рассмотрю аналогичную уязвимость в GraphicsMagick, обнаруженную мной в ходе анализа исходного кода библиотеки.
https://habr.com/ru/company/wrike/blog/458926/

🔗 Трагедия не приходит одна
В 2016 году уязвимость ImageTragick в библиотеке ImageMagick наделала много шума. Как способ снижения риска предлагалось использовать GraphicsMagick — форк библ...

​Face Recognition App In React Native using AWS Rekognition

https://medium.com/zenofai/face-recognition-app-in-react-native-using-aws-rekognition-c10b188a6413

🔗 Face Recognition App In React Native using AWS Rekognition
In this blog, we are going to build an app for registering and verifying faces using Amazon Rekognition in React Native.

​Neuralink платформа интегрированного интерфейса мозг-компьютер с тысячами каналов
#Искусственныйинтеллект

Примечание. 16 июля 2019г. Илон Маск презентовал технологию, которая стала одной из первых серьезных попыток введения нейроинтерфейсов в клиническую практику и имеющая реальные трансгуманистические цели в долгосрочной перспективе. Данная статья является переводом, с некоторыми примечаниями, оригинальной статьи (https://www.biorxiv.org/content/10.1101/703801v2.full) описывающей технологию нейрокомпьютерного интерфейса, ссылка на которую так же размещена на сайте компании разработчика Neuralink (https://www.neuralink.com/). Статья содержит описание особенностей и характеристик основных модулей лежащих в основе данной технологии, в том числе: гибких биосовместимых микроэлектродов, роботизированного нейрохирургического манипулятора для введения электродов в мозг, а так же специализированной микроэлектроники обеспечивающей прием усиление и оцифровку сигнала поступающего от нейронной активности.
https://habr.com/ru/post/461215/

🔗 Neuralink платформа интегрированного интерфейса мозг-компьютер с тысячами каналов
Примечание. 16 июля 2019г. Илон Маск презентовал технологию, которая стала одной из первых серьезных попыток введения нейроинтерфейсов в клиническую практику и...

​ML trainings updates – Pavel Pleskov; Kaggle Google Landmark Recognition 2019 – Ivan Sosin

🔗 ML trainings updates – Pavel Pleskov; Kaggle Google Landmark Recognition 2019 – Ivan Sosin
Pavel Pleskov talks about updates on machine learning trainings. Ivan Sosin tells about his participation in Kaggle Google Landmark Recognition 2019 competition in which his teammates were Kirill Brodt and Pavel Pleskov. The team got a gold medal. In this video you will find out: - The competition overview - Dataset specifics - Details of the solution - Problems that were encountered Find out about new competitions http://mltrainings.ru/ Find out about new machine learning trainings: VKontakte https://vk

​Org-scale analytics: Today’s startups build societies. Do it right.
Org-scale analytics uses data science to compare groups of groups, teams, and organizations.
https://towardsdatascience.com/org-scale-analytics-todays-startups-build-societies-do-it-right-4f6185e81482?source=collection_home---4------0-----------------------

🔗 Org-scale analytics: Today’s startups build societies. Do it right.
Org-scale analytics uses data science to compare groups of groups, teams, and organizations.

Built-in Machine Learning Algorithms with Amazon SageMaker - a Deep Dive

https://www.youtube.com/watch?v=yGc0qePSYig

🎥 Built-in Machine Learning Algorithms with Amazon SageMaker - a Deep Dive
👁 1 раз ⏳ 938 сек.
Learn more about Amazon SageMaker at – https://amzn.to/2ZjenDf 
Amazon SageMaker comes built-in with a number of high-performance algorithms for different use cases. Learn the fundamentals and then dive deep into these algorithms.

​The Evolution of Deeplab for Semantic Segmentation
In computer vision, a simple image segmentation is the process of partitioning a digital image into multiple segments (sets of pixels)
https://towardsdatascience.com/the-evolution-of-deeplab-for-semantic-segmentation-95082b025571?source=collection_home---4------1-----------------------

🔗 The Evolution of Deeplab for Semantic Segmentation
In computer vision, a simple image segmentation is the process of partitioning a digital image into multiple segments (sets of pixels)…

​A ConvNet that works on like, 20 samples: Scatter Wavelets
The lesser known, non-neural, convolution network

🔗 A ConvNet that works on like, 20 samples: Scatter Wavelets
The lesser known, non-neural, convolution network

​Reservoir Computing Models for Patient-Adaptable ECG Monitoring in arxiv.org/abs/1907.09504

🔗 Reservoir Computing Models for Patient-Adaptable ECG Monitoring in Wearable Devices
The reservoir computing paradigm is employed to classify heartbeat anomalies online based on electrocardiogram signals. Inspired by the principles of information processing in the brain, reservoir computing provides a framework to design, train, and analyze recurrent neural networks (RNNs) for processing time-dependent information. Due to its computational efficiency and the fact that training amounts to a simple linear regression, this supervised learning algorithm has been variously considered as a strategy to implement useful computations not only on digital computers but also on emerging unconventional hardware platforms such as neuromorphic microchips. Here, this biological-inspired learning framework is exploited to devise an accurate patient-adaptive model that has the potential to be integrated into wearable cardiac events monitoring devices. The proposed patient-customized model was trained and tested on ECG recordings selected from the MIT-BIH arrhythmia database. Restrictive inclusion criteria were

​Predicting vs. Explaining
And Why Data Science Needs More “Half-Bayesians”
https://towardsdatascience.com/predicting-vs-explaining-69b516f90796?source=collection_home---4------0-----------------------

🔗 Predicting vs. Explaining
And Why Data Science Needs More “Half-Bayesians”

Image Captioning with Keras and TensorFlow (10.4)

🎥 Image Captioning with Keras and TensorFlow (10.4)
👁 1 раз ⏳ 1616 сек.
Using multi-image recognition and natural language processing it is possible to create a neural network that can write captions for images. This video shows how to create and train an image captioning neural network for Keras. Transfer learning is used to greatly reduce training time. Makes use of glove and InceptionV3.

Code for This Video:
https://github.com/jeffheaton/t81_558_deep_learning/blob/master/t81_558_class_10_4_captioning.ipynb

Course Homepage: https://sites.wustl.edu/jeffheaton/t81-558/

F

Как повысить безопасность с помощью больших данных от сбора до анализа с помощью машинного обучения
https://www.youtube.com/watch?v=yh3jQvmEoVQ

🎥 Как повысить безопасность с помощью больших данных от сбора до анализа с помощью машинного обучения
👁 1 раз ⏳ 2787 сек.

​Introducing TensorFlow Addons

https://medium.com/tensorflow/introducing-tensorflow-addons-6131a50a3dcf

🔗 Introducing TensorFlow Addons
Posted by: Sean Morgan (Two Six Labs), Yan Facai (Alibaba), Moritz Kröger (RWTH Aachen University), Tzu-Wei Sung (National Taiwan…

​DL соревнования — гуси, пайплайны, кулстори – Артур Кузин

🔗 DL соревнования — гуси, пайплайны, кулстори – Артур Кузин
Артур Кузин рассказывает про переход и профессиональное развитие в Data Science. А также несколько кулстори и рекомендации начинающим дата сайентистам. Из видео вы сможете узнать: - С чего начать, если хочешь стать дата сайентистом - Зачем участвовать в соревнованиях и почему это весело - Про атрибуты хорошего пайплайна - Интересные кейсы из соревнований *Доклад с тренировки в 2018 году Узнать о текущих соревнованиях можно на сайте http://mltrainings.ru/ Узнать о новых тренировках и видео можно из групп:

​HUGE2: a Highly Untangled Generative-model Engine for Edge-computing
https://arxiv.org/abs/1907.11210

🔗 HUGE2: a Highly Untangled Generative-model Engine for Edge-computing
As a type of prominent studies in deep learning, generative models have been widely investigated in research recently. Two research branches of the deep learning models, the Generative Networks (GANs, VAE) and the Semantic Segmentation, rely highly on the upsampling operations, especially the transposed convolution and the dilated convolution. However, these two types of convolutions are intrinsically different from standard convolution regarding the insertion of zeros in input feature maps or in kernels respectively. This distinct nature severely degrades the performance of the existing deep learning engine or frameworks, such as Darknet, Tensorflow, and PyTorch, which are mainly developed for the standard convolution. Another trend in deep learning realm is to deploy the model onto edge/ embedded devices, in which the memory resource is scarce. In this work, we propose a Highly Untangled Generative-model Engine for Edge-computing or HUGE2 for accelerating these two special convolutions on the edge-computing

Artificial Intelligence with #Python | Artificial Intelligence Tutorial using Python | Edureka
https://www.youtube.com/watch?v=7O60HOZRLng

🎥 Artificial Intelligence with Python | Artificial Intelligence Tutorial using Python | Edureka
👁 1 раз ⏳ 6144 сек.
* Machine Learning Engineer Masters Program: https://www.edureka.co/masters-program/machine-learning-engineer-training *
This Edureka video on "Artificial Intelligence With Python" will provide you with a comprehensive and detailed knowledge of Artificial Intelligence concepts with hands-on examples.


Python Course: https://www.youtube.com/watch?v=vaysJAMDaZw
Statistics and Probability Tutorial: https://www.youtube.com/watch?v=XcLO4f1i4Yo
Check out the entire Machine Learning Playlist: https://bit.ly/2NG9t

ML на службе threat hunter

🎥 ML на службе threat hunter
👁 1 раз ⏳ 2868 сек.
Тема: Технический доклад

Спикеры: Алексей Тараненко

Рассказ о том, как машинное обучение помогает аналитику кибербезопасности и офицеру по расследованиям инцидентов в их повседневной работе. Докладчик поделится опытом использования ML в Microsoft. Покажет примеры расследований инцидентов ИБ с помощью Azure Notebook.

​Image Panorama Stitching with OpenCV
Image stitching is one of the most successful applications in Computer Vision.

https://towardsdatascience.com/image-panorama-stitching-with-opencv-2402bde6b46c?source=collection_home---4------0-----------------------

🔗 Image Panorama Stitching with OpenCV
Image stitching is one of the most successful applications in Computer Vision. Nowadays, it is hard to find a cell phone or an image…