🎥 V-Sense Seminar: Deep Learning
👁 1 раз ⏳ 6668 сек.
Artificial Intelligence (AI) has made it from science fiction into everyday life. Machine Learning (ML) enabled breakthroughs due to availability of massive data and computational resources. Deep Learning (DL) in particular disrupted all areas of visual computing (VC), including computer vision/graphics and image/video processing.

The V-SENSE team of Trinity College Dublin adopted this challenge and opportunity, and made a number of significant contributions to the field of DL for VC over the last 2 years,

​Задача: извлечь ключевые выражения из текста на русском языке. NLP на Python
Что было нужно в самом начале:

программа, «выуживающая» из сырого текста на русском языке уникальные названия продукции по определенной отрасли. Сырой текст — текст, который писал человек, просто излагая свои мысли и не заботясь о формировании или выделении какого-либо списка слов;
автоматически получаемый список слов;
минимальная ручная или автоматизированная обработка для преобразования списка в набор хештегов или ключевых слов к тексту.

Полагаю, что неявно с проблемой многие сталкиваются ежедневно, после написания или анализа статьи, поста, комментария, заметки, отчета и т.д. Вот и мне по роду деятельности приходилось сталкиваться с данной проблемой по многу раз в день. Поэтому, можно сказать, к идее автоматизации меня привела «лень», в хорошем смысле этого слова.

Сейчас, когда я пишу эту статью, сохранилась идея, но набор данных конечного результата сильно изменился:

выбираются не слова, а ключевые выражения и в том числе слова;
список ключевых выражений размечен 16-ю различными маркерами;
все слова текста (в том числе и не ключевые) лемматизированы – приведены в начальную форму или унифицированы под выведенный формат;
каждое слово в тексте имеет дополнительную аналитику, относящуюся к положению по тексту и числу повторений.

Результаты работы ПО nrlpk (Natural Russian Language Processing by the Keys) подготавливают данные для:

анализа текстов неограниченного круга тематик и отраслей (разработка и тестирование проводилось по материалам тематики промышленности и ВПК — Военно-Промышленного Комплекса);
автоматической рубрикации, классификации, каталогизации, предметизации материалов (online площадки);
контроля и фильтрации по содержимому с настройками реакции системы (службам и системам безопасности в замкнутых контурах или online);
многослойной разметки текстов (ИИ).

Качество

🔗 Задача: извлечь ключевые выражения из текста на русском языке. NLP на Python
Что было нужно в самом начале: программа, «выуживающая» из сырого текста на русском языке уникальные названия продукции по определенной отрасли. Сырой текст —...

​Применение сиамских нейросетей в поиске

Всем привет! В этом посте я расскажу, какие подходы мы в Поиске Mail.ru используем для сравнения текстов. Для чего это нужно? Как только мы научимся хорошо сравнивать разные тексты друг с другом, поисковая система сможет лучше понимать запросы пользователя.

Что нам для этого нужно? Для начала строго поставить задачу. Нужно определить для себя, какие тексты мы считаем похожими, а какие не считаем и затем сформулировать стратегию автоматического определения схожести. В нашем случае будут сравниваться тексты пользовательских запросов с текстами документов.

🔗 Хабр

😻 3 :) 😻 3 :) 😻 😻

🎥 SQL For Data Science Tutorial | Learn SQL Database For Data Science | Edureka
👁 1 раз ⏳ 2342 сек.
** ** Data Science Master Program: https://www.edureka.co/masters-program/data-scientist-certification **
This Edureka session on SQL for Data Science will help you understand how SQL can be used to store, access and retrieve data to perform data analysis.
Here’s a list of topics covered in this session:

1. Introduction To Data Science
2. Why Is SQL Needed For Data Science?
3. What Is SQL?
4. Basics Of SQL
5. Installing MySQL
6. Hands-On

Do subscribe to our channel and

🎥 Test Driven Machine Learning - Detlef D Nauck, Chief Research Scientist, BT Technology
👁 1 раз ⏳ 1889 сек.
Data Scientists and machine learning specialists are familiar with testing principles during the model building phase like cross-validation, but they are often unfamiliar with test-driven software engineering principles. While testing a learned model gives an idea how well it might perform on unseen data this is not sufficient for model deployment. Trying to learn from test driven software development practices we look across the machine learning life cycle to understand where we need to test and how this c

​Kaggle Live Coding: Automatically generating reports | Kaggle

🔗 Kaggle Live Coding: Automatically generating reports | Kaggle
Today we'll be working on taking the output of our text clusters and use it to generate a human-readable report. Link to paper: https://arxiv.org/pdf/1906.04358.pdf SUBSCRIBE: https://www.youtube.com/c/kaggle?sub_... About Kaggle: Kaggle is the world's largest community of data scientists. Join us to compete, collaborate, learn, and do your data science work. Kaggle's platform is the fastest way to get started on a new data science project. Spin up a Jupyter notebook with a single click. Build with our h

​Exploring wild west of natural language generation — from n-gram and RNNs to Seq2Seq

🔗 Exploring wild west of natural language generation — from n-gram and RNNs to Seq2Seq
Introduction to Natural Language Models as taught in an NLP Stanford class

🎥 Python Voice Assistant Tutorial #9 - Waking the Assistant
👁 1 раз ⏳ 486 сек.
In this python voice assistant tutorial I will cover how we can create a wake keyword for our assistant. This word will allow us to trigger the assistant. Something like "hey tim".

Text-Based Tutorial: Coming Soon...

*****
Enroll in The Fundamentals of Programming w/ Python
https://tech-with-tim.teachable.com/p/the-fundamentals-of-programming-with-python

Instagram: https://www.instagram.com/tech_with_tim
Website https://techwithtim.net
Twitter: https://twitter.com/TechWithTimm
Discord: https://discord.g

​Training Robust Deep Neural Networks via Adversarial Noise Propagation

Authors: Aishan Liu, Xianglong Liu, Chongzhi Zhang, Hang Yu, Qiang Liu

Abstract: Deep neural networks have been found vulnerable to noises like adversarial examples and corruption in practice. A number of adversarial defense methods have been developed, which indeed improve the model robustness towards adversarial examples in practice. However, only relying on training with the data mixed with noises, most of them still fail to defend the generalized types of noises. Motivated by the fact that hidden layers play a very important role in maintaining a robust model, this paper comes up with a simple yet powerful training algorithm named Adversarial Noise Propagation (ANP) that injects diversified noises into the hidden layers in a layer-wise manner. We show that ANP can be efficiently implemented by exploiting the nature of the popular backward-forward training style for deep models

https://arxiv.org/abs/1909.09034

🔗 Training Robust Deep Neural Networks via Adversarial Noise Propagation
Deep neural networks have been found vulnerable to noises like adversarial examples and corruption in practice. A number of adversarial defense methods have been developed, which indeed improve the model robustness towards adversarial examples in practice. However, only relying on training with the data mixed with noises, most of them still fail to defend the generalized types of noises. Motivated by the fact that hidden layers play a very important role in maintaining a robust model, this paper comes up with a simple yet powerful training algorithm named Adversarial Noise Propagation (ANP) that injects diversified noises into the hidden layers in a layer-wise manner. We show that ANP can be efficiently implemented by exploiting the nature of the popular backward-forward training style for deep models. To comprehensively understand the behaviors and contributions of hidden layers, we further explore the insights from hidden representation insensitivity and human vision perception alignment. Extensive experime

​Calculating Maximum Likelihood Estimation by Hand Step-by-step

🔗 Calculating Maximum Likelihood Estimation by Hand Step-by-step
I wrote this because I couldn’t find many tutorials showing the detailed math for this calculation. So I decided to write it out…

​Understanding 1D and 3D Convolution Neural Network | Keras

🔗 Understanding 1D and 3D Convolution Neural Network | Keras
When we say Convolution Neural Network (CNN), generally we refer to a 2 dimensional CNN which is used for image classification. But there…

🎥 Learn how to morph faces with a Generative Adversarial Network!
👁 1 раз ⏳ 1527 сек.
Link to Notebooks:
https://drive.google.com/open?id=1LBWcmnUPoHDeaYlRiHokGyjywIdyhAQb

Link to the StyleGAN paper: https://arxiv.org/abs/1812.04948

--------------------------------

This episode covers one of the greatest ideas in Deep Learning of the past couple of years: Generative Adversarial Networks.

I first explain how a generative adversarial network (GAN) really works. After this general overview, we go into the specific objective function that is optimized during training. We then dive into Nvidi

​Discrete Probability Distributions for Machine Learning

https://machinelearningmastery.com/discrete-probability-distributions-for-machine-learning/
Наш телеграм канал - tglink.me/ai_machinelearning_big_data

🔗 Discrete Probability Distributions for Machine Learning
The probability for a discrete random variable can be summarized with a discrete probability distribution. Discrete probability distributions are used in machine learning, most notably in the modeling of binary and multi-class classification problems, but also in evaluating the performance for binary classification models, such as the calculation of confidence intervals, and in the modeling …

​🗣 Using AI-generated questions to train NLP systems

https://ai.facebook.com/blog/research-in-brief-unsupervised-question-answering-by-cloze-translation/

code: https://github.com/facebookresearch/UnsupervisedQA

paper: https://research.fb.com/publications/unsupervised-question-answering-by-cloze-translation/

🔗 facebookresearch/UnsupervisedQA
Unsupervised Question answering via Cloze Translation - facebookresearch/UnsupervisedQA

​Лучшие 50 визуализаций matplotlib — The Master Plots (с полным кодом на Python)
Подборка 50 графиков matplotlib, наиболее полезных для анализа и визуализации данных. Этот список позволяет вам выбрать, какую визуализацию показывать для какой ситуации, используя библиотеки python matplotlib и seaborn.

🔗 Лучшие 50 визуализаций matplotlib — The Master Plots (с полным кодом на Python)
Подборка 50 графиков matplotlib, наиболее полезных для анализа и визуализации данных. Этот список позволяет вам выбрать, какую визуализацию показывать для какой...

​One Shot learning, Siamese networks and Triplet Loss with Keras

🔗 One Shot learning, Siamese networks and Triplet Loss with Keras
Introduction

🎥 9/20/2019 - Using Big Data to Detect Clinical Deterioration
👁 1 раз ⏳ 3496 сек.
Matthew Churpek, MD, MPH, PhD, presents a talk about data-driven risk stratification can help predict clinical deterioration in hospital settings. Dr. Churpek is a visiting associate professor in the Division of Allergy, Pulmonary and Critical Care at the University of Wisconsin-Madison Department of Medicine. He is a board-certified clinical informaticist whose research program focuses on developing and implementing prediction models to detect early clinical deterioration in order to improve patient outcom

​New Face Swapping AI Creates Amazing DeepFakes

🔗 New Face Swapping AI Creates Amazing DeepFakes
📝 The paper "FSGAN: Subject Agnostic Face Swapping and Reenactment" is available here: https://nirkin.com/fsgan/ ❤️ Pick up cool perks on our Patreon page: https://www.patreon.com/TwoMinutePapers 🙏 We would like to thank our generous Patreon supporters who make Two Minute Papers possible: Alex Haro, Andrew Melnychuk, Angelos Evripiotis, Anthony Vdovitchenko, Brian Gilman, Bruno Brito, Bryan Learn, Christian Ahlin, Christoph Jadanowski, Claudio Fernandes, Daniel Hasegan, Dennis Abts, Eric Haddad, Eric Mart

​Will Machines Ever Be Capable of Empathy?

🔗 Will Machines Ever Be Capable of Empathy?
Humans are magical creatures and no machine can replicate this magic.