​6 февраля в 20:00 мск OTUS приглашает на бесплатный вебинар онлайн-курса «Machine learning» - «Как я взял золото на Kaggle. Разбор задач» для аналитиков и разработчиков. Для регистрации пройдите тест (необходимо знание Python и математики): https://otus.pw/MQ5z/

На вебинаре Валерий Бабушкин (Competitions Grandmaster на Kaggle, top 30) разберёт несколько задач с соревнований, на которых он взял золото, и поделится секретами побед.

Будут разобраны задачи: IEEE kaggle camera identification; carvana masking challenge. Не упустите возможность задать преподавателю все интересующие вопросы!

Чтобы попасть на курс с welcome-скидкой прямо сейчас, пройдите вступительный тест: https://otus.pw/ubps/

🔗 Machine learning: современные методы анализа данных и инструменты Data Science
Курс Machine learning в OTUS с возможностью трудоустройства!

🎥 Probabilistic Deep Learning in TensorFlow: The Why and How | ODSC Europe 2019
👁 1 раз ⏳ 1740 сек.
Bayesian probabilistic techniques allow machine learning practitioners to encode expert knowledge in otherwise-uninformed models and support uncertainty in model output. Probabilistic deep learning models take this further by fitting distributions rather than point estimates to each of the weights in a neural network, allowing its builder to inspect the prediction stability for any given set of input data. Following a slew of recent technical advancements, it's never been easier to apply probabilistic model

​How to succeed in code (kernel) competitions | Dmitry Gordeev | Kaggle Days

🔗 How to succeed in code (kernel) competitions | Dmitry Gordeev | Kaggle Days
Content note: The presenter references 'Kaggle Kernels' which have recently been renamed to 'Notebooks'. Kaggle Days Tokyo took place on December 11-12, 2019 at Mori Tower, Roppongi Hills, Tokyo. This was the 6th edition of our signature two-day event featuring Master and Grandmaster speakers, and a live, all-day Kaggle Competition. This edition was sponsored by Google Cloud (GCP), Data Science Dojo, and DeNA. Kaggle Days is produced by LogicAI and Kaggle. About LogicAI: LogicAI is a boutique Data Scie

​Методы регрессионного анализа в Data Science
Накануне запуска курса «Математика для Data Science. Продвинутый курс» мы провели открытый вебинар на тему «Методы регрессионного анализа в Data Science». На нём познакомились с понятием линейных регрессий, изучили, где и как их можно применять на практике, а также узнали, какие темы и разделы математического анализа, линейной алгебры и теории вероятностей используются в этой области. Преподаватель — Петр Лукьянченко, преподаватель НИУ ВШЭ, руководитель технологических проектов.

Если мы говорим о математике в контексте Data Science, мы можем выделить три наиболее часто решаемые задачи (хотя задач, разумеется, больше):

🔗 Методы регрессионного анализа в Data Science
Накануне запуска курса «Математика для Data Science. Продвинутый курс» мы провели открытый вебинар на тему «Методы регрессионного анализа в Data Science». На нём...

​Augmenting Self-attention with Persistent Memory

It includes LSH attention, reversible network, and chunking. It has been validated with an auto-regressive task (enwik8). It also includes additional features to make the entire network pure attention all the way down.

https://github.com/lucidrains/reformer-pytorch

https://arxiv.org/abs/1907.01470v1
Наш телеграм канал - tglink.me/ai_machinelearning_big_data

🔗 lucidrains/reformer-pytorch
Reformer, the efficient Transformer, in Pytorch. Contribute to lucidrains/reformer-pytorch development by creating an account on GitHub.

​AdelaiDet is an open source toolbox for multiple instance-level detection applications.

https://github.com/aim-uofa/adet

🔗 aim-uofa/adet
AdelaiDet is an open source toolbox for multiple instance-level detection applications. - aim-uofa/adet

​On implementing Deep Learning library from scratch in Python

🔗 On implementing Deep Learning library from scratch in Python
A beginner’s guide to understanding the fundamental building blocks of deep learning platforms

​How to extract facial expressions, head pose, and gaze from any Youtube video

🔗 How to extract facial expressions, head pose, and gaze from any Youtube video
A tutorial using Google Colab & OpenFace to extract facial features from Youtube videos without installing a single program on your laptop.

🎥 Stanford ICME Lecture on Deep Learning
👁 1 раз ⏳ 4554 сек.
Random Matrix Theory (RMT) is applied to analyze the weight matrices of
Deep Neural Networks (DNNs), including production quality, pre-trained
models and smaller models trained from scratch. Empirical and theoretical
results indicate that the DNN training process itself implements a
form of self-regularization, evident in the empirical spectral density (ESD)
of DNN layer matrices. To understand this, we provide a phenomenology
to identify 5+1 Phases of Training, corresponding to increasing amounts of
i

​Understand Neural Networks & Model Generalization

🔗 Understand Neural Networks & Model Generalization
The Challenge of Model Generalization, Overfitting and Regularization Methods for Deep Neural Networks

​Gated Path Selection Network for Semantic Segmentation

🔗 Gated Path Selection Network for Semantic Segmentation
01/19/20 - Semantic segmentation is a challenging task that needs to handle large scale
variations, deformations and different viewpoints. In...

​Deep Learning–based Prescription of Cardiac MRI Planes
https://escholarship.org/uc/item/8r27603c

🔗 Deep Learning–based Prescription of Cardiac MRI Planes
Author(s): Blansit, Kevin; Retson, Tara; Masutani, Evan; Bahrami, Naeim; Hsiao, Albert

​MOTS: Multi-Object Tracking and Segmentation

https://www.vision.rwth-aachen.de/page/mots

🔗 Computer Vision

​https://medium.com/the-future-is-electric/machine-learning-isnt-the-carbon-hog-headlines-suggested-6310b2c571dc

🔗 Machine learning isn’t the carbon hog headlines suggested
Climate research and actions related to machine learning far outweigh the overstated carbon debt

🎥 Дмитрий Бабаев - Нерешенные проблемы нейронных сетей - DataStart.ru
👁 1 раз ⏳ 2519 сек.
Билеты на следующую конференцию:
https://datastart.ru/latest

Материалы последней конференции:
https://datastart.ru/post-event

На видео – вторая конференция DataStart 07/04/2018

Дмитрий Бабаев - Нерешенные проблемы нейронных сетей

Описание доклада:
Глубокое обучение нейронных сетей явилось прорывом в развитии систем искусственного интеллекта. Многие задачи, недоступные классическим методам машинного обучения, были решены на уровне, близком к возможностям человека или даже более высоком. Тем не менее, с

NLP with Deep Learning
https://www.youtube.com/watch?v=9oTHFx0Gg3Q

🎥 Stanford CS224N: NLP with Deep Learning | Winter 2019 | Lecture 12 – Subword Models
👁 1 раз ⏳ 4530 сек.
Professor Christopher Manning, Stanford University
http://onlinehub.stanford.edu/

Professor Christopher Manning
Thomas M. Siebel Professor in Machine Learning, Professor of Linguistics and of Computer Science
Director, Stanford Artificial Intelligence Laboratory (SAIL)

To follow along with the course schedule and syllabus, visit: http://web.stanford.edu/class/cs224n/index.html#schedule

To get the latest news on Stanford’s upcoming professional programs in Artificial Intelligence, visit: http://learn.s

​This equation will change how you see the world

🔗 This equation will change how you see the world
The logistic map connects fluid convection, neuron firing, the Mandelbrot set and so much more. Fasthosts are giving UK viewers the chance to win tickets, flight, and accommodation to SXSW 2020 by answering my Techie Test question: https://www.fasthosts.co.uk/veritasium Animations, coding, interactives in this video by Jonny Hyman 🙌 Try the code yourself: https://github.com/jonnyhyman/Chaos References: James Gleick, Chaos Steven Strogatz, Nonlinear Dynamics and Chaos May, R. Simple mathematical models w

​DiffTaichi: Differentiable Programming for Physical Simulation

We present DiffTaichi, a new differentiable programming language tailored for building high-performance differentiable physical simulators. Based on an imperative programming language, DiffTaichi generates gradients of simulation steps using source code transformations that preserve arithmetic intensity and parallelism

https://github.com/yuanming-hu/difftaichi

https://arxiv.org/abs/1910.00935v2
Наш телеграм канал - tglink.me/ai_machinelearning_big_data

🔗 yuanming-hu/difftaichi
10 differentiable physical simulators built with Taichi differentiable programming (DiffTaichi, ICLR 2020) - yuanming-hu/difftaichi

​Deep Graph Matching Consensus
Implementation of Deep Graph Matching Consensus in PyTorch

Code: https://github.com/rusty1s/deep-graph-matching-consensus

Paper: https://arxiv.org/abs/2001.09621v1
Наш телеграм канал - tglink.me/ai_machinelearning_big_data

🔗 rusty1s/deep-graph-matching-consensus
Implementation of Deep Graph Matching Consensus in PyTorch - rusty1s/deep-graph-matching-consensus

​Machine Learning at VU University Amsterdam
Download: https://mlvu.github.io

Youtube: https://www.youtube.com/playlist?list=PLCof9EqayQgsORO3pFzeYZFz6cszYO0VJ

🔗 VU Machine Learning 2018 - YouTube