Best article I've read so far on understanding how Neural ODEs work. Well and clearly explained by solving a regression problem. If you have read the paper (Link: https://lnkd.in/dfDnJJz) and had difficulties understanding it then you should definitely read this blog post. #deeplearning #machinelearning

Link: https://lnkd.in/dhMhwyb

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

The Matrix Calculus You Need For Deep Learning

By Terence Parr and Jeremy Howard : https://lnkd.in/dC5MqZM

#ArtificialIntelligence #BigData #DeepLearning #MachineLearning #NeuralNetworks

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

HOW TO IMPROVE YOUR SKILL ON TEXT DATA?

Rubens Zimbres, PhD compile amazing resources on Machine Learning, NLP, and Computer Vision. On NLP Side he cover pretty much every common topic on NLP, this is very useful because as data scientist we often dealing with text data.

Yo can see the repository here https://lnkd.in/fyyvZYt

#repository #machinelearning #patternrecognition #artificialintellegence

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

Your Data Strategy is your AI Strategy

Unless you’re one of the massive digital-first companies your AI strategy needs to address these truths:

1. Right now Artificial Intelligence is really Machine Learning. Comprised of three elements; traditional ML (e.g., XG-Boost, SVM, etc.), Deep Learning, Reinforcement Learning
2. You’re not going to create new algorithms that significantly advance the science of ML. This is because it’s very difficult to compete with the likes of Google, Apple, etc. for the best algorithm developers and the billions they are investing in ML.
3. All of your competitors are working on their AI/ML strategy and soon your investments in AI/ML will be table stakes, not a sustainable competitive advantage.

The only way to compete in the above environment is to focus on your one area of competitive advantage; your data. Whether is be customer or internal process data, it is the first place to go for creating value and to do that you’ll need a data strategy.

More on data strategy in my next post.


#artificialintelligence #machinelearning #digitization #datascience


✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

100 Days of Machine Learning Coding as proposed by Siraj Raval

#tutorial #machine_learning #python #infographic #100DaysOfMLCode
@pythony

github.com/Avik-Jain/100-Days-Of-ML-Code

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

A large, biased sample in some ways is worse than a small, biased sample. It will be more credible in the eyes of many decision-makers by virtue of its size and more analytic damage can be done with it - weak associations may be statistically "significant" in large samples and many analytic procedures will not run on very small samples, for example.

The potential silver lining is when the data contain information that helps us understand and, possibly, adjust for sources of bias. This is much easier to accomplish when the sample is large.

How big is large? This depends on many things, such as the number of variables that will be used in the analysis and, more importantly, the purpose of the research.

In general, however, statisticians consider a sample of a few hundred "large" by historical definitions.

A "sample" of a few dozen that actually represents the population avoids having to make statistical inferences about the population from the sample, though often researchers are really generalizing beyond that population, e.g., to employees at the same company in five years. "Populations" can be elusive. Moreover, there are analytic limitations to very small data even when it is not a sample.

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

Make music with GANs!

GANSynth is a new method for fast generation of high-fidelity audio.

🎵 Examples: https://lnkd.in/enSDBJZ
⏯ Colab: https://lnkd.in/eDt_S3w
📝 Paper: https://lnkd.in/eCsZbx2
💻 Code: https://lnkd.in/eN3B5xc
⌨️ Blog: https://lnkd.in/eXks33i

Join Us and Share With Your Friend

#artificialintelligence #deeplearning #generativeadversarialnetworks

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

HOW TO WRITE BETTER CODE FOR DATA SCIENCE

1. Strictly follow style standards
-> https://lnkd.in/gKZUjVa

2. Use a linter to enforce style standards
-> https://lnkd.in/d_prybR

3. Write modular, generic, object-oriented code -
-> https://lnkd.in/gsynW6Q
-> https://lnkd.in/dx53u53

4. Write unit tests to test your functions and methods
-> https://lnkd.in/dsy-bPu

5. Organize your code base
-> https://lnkd.in/dviGffH

6. Separate exploration and production development, and develop production code using test-driven development (TDD)
-> https://lnkd.in/dMn-s32

This list curated by DSDJ Founder, Kyle McKiou

If you'd like some real code examples, DSDJ got 5 end-to-end data science projects with instructions, data, code, and complete video walkthroughs as part of the course.

These examples and videos will walk you through everything that you need to take your data science coding skills to the next level.

#datascience #programming

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

MIT Introduction to Deep Learning

MIT 6.S191 Introduction to Deep Learning: https://lnkd.in/e2qmSWR

#artificialintelligence #deeplearning #machinelearning #tensorflow

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

Math Blocking you from doing Machine Learning?

#machinelearning

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

"When you first study a field, it seems like you have to memorize a zillion things. You don't. What you need is to identify the 3-5 core principles that govern the field. The million things you thought you had to memorize are various combinations of the core principles." -J. Reed

“1. Multiply things together
2. Add them up
3. Replaces negatives with zeros
4. Return to step 1, a hundred times” - Jeremy Howard

#artificialintelligence #deeplearning #machinelearning

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

If you want to learn the basics of deep learning and also TensorFlow, then check out MIT's introductory course on deep learning. The complete lecture series is now online with a lot of code examples for TensorFlow 2.0. The course covers a wide range of topics (Computer Vision, GANs, NLP etc) and there are also guest lectures from Google, NVIDIA etc. The code is designed to run seamlessly on Google colab so very easy then. Check it out! #deeplearning #machinelearning

Article: https://lnkd.in/d7yT6dU
Course page: https://lnkd.in/deaTRDZ
Github code: https://lnkd.in/d-jwcPW

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

Cheat Sheet

Subgradient Descent, Mirror Descent, and Online Learning

By Sebastian Pokutta: https://lnkd.in/eMYrh33

#artificialintelligence #deeplearning #machinelearning

✴️ @AI_Python_EN
❇️ @AI_Python
🗣 @AI_Python_arXiv

Deep Variational Reinforcement Learning for POMDPs

https://lnkd.in/eQNusdt

#ArtificialIntelligence #DeepLearning #NeuralNetworks #ReinforcementLearning

✴️ @AI_Python_EN

#DataScience Cheetsheet

✴️ @AI_Python_EN

Any refresher needed for statistics then check out this 10-page cheatsheet that covers a semester's worth of introductory statistics. Very cool it covers all the basic stuff from Bayes' rule to Markov chains. Check it out! #statistics


Github: https://lnkd.in/dgm9GcC

✴️ @AI_Python_EN

Equi-normalization of Neural Networks

Stock et al.: https://lnkd.in/eSCXsr7

#ComputerVision #PatternRecognition #MachineLearning

✴️ @AI_Python_EN

Model predictive control actuates the overhead crane to move objects of different mass. The overhead crane is converted to an inverted pendulum with a sign change in the state space model. Python GEKKO has new capabilities for system identification and state space modeling.

Overhead crane in Python and MATLAB: https://lnkd.in/e_EB23R
Inverted pendulum: https://lnkd.in/eZpgd_3

#modelpredictivecontrol #python

✴️ @AI_Python_EN

7 Steps on Data Science Interview, No 5 Is Game Changer (at least for Me)

Pranav Dar write a concise guidelines for acing the Interview (I'm also hiring, you can see here: https://lnkd.in/f4zQwEw).

Homework (no 5)is game changer, this is the place that can make company feel special.

You can see read complete article here
https://lnkd.in/f92mu9h

#datascience #interviews

✴️ @AI_Python_EN

Robust Re-identification of Manta Rays from Natural Markings by Learning Pose Invariant Embeddings

Moskvyak et al.: https://lnkd.in/eqYaqQD

#ArtificialNeuralNetworks #ComputerVision #PatternRecognition #Technology

✴️ @AI_Python_EN