https://lnkd.in/edT8m4y #machineleaning
“PyText is a deep-learning based NLP modeling framework built on PyTorch”

For those unfamiliar with the man: Schmidhuber is one of the creators of LSTM. Well worth your time.
https://lnkd.in/dYgXxMm

✴️ @AI_Python_EN

Machine learning models are being increasingly used to make decisions that affect people’s lives. With this power comes a responsibility to ensure that the model predictions are fair and not discriminating.

#machinelearning #bias
https://towardsdatascience.com/is-your-machine-learning-model-biased-94f9ee176b67?gi=d3c7b468df95

✴️ @AI_Python_EN

Check out this beautifully drawn notes from the excellent Coursera specialization by Andrew Ng!

Source: https://lnkd.in/dw5_Fmm

✴️ @AI_Python_EN

▪️Data Science Projects:

☆ The Data Science IPython Notebooks:
=> This repository is filled with IPython notebooks that cover different topics, going from Kaggle competitions to big data and deep learning.
[ https://lnkd.in/dW3WBi6 ]

☆ The Pattern Classification:
=> Tutorials and examples to solve and understand machine learning and pattern classification tasks.
[ https://lnkd.in/d9PGxHm ]

☆ Deep Learning In Python:
=> This repository is the way to go!
[ https://lnkd.in/d-hNVCD ]

More ...
▪️Data Science News
▪️Data Science Books
▪️Data Science Talks
▪️R for Data Science Talks
▪️Python for Data Science Talks
▪️Big Data Talks
▪️Data Science Podcasts
▪️Data Science Webinars
▪️Data Science Tutorials
▪️Data Science Community
▪️Data Science Courses

Refer to the full article
[ https://lnkd.in/dmKmx_D ]

✴️ @AI_Python_EN

Although there have been notable exceptions, quantitative marketing research on the whole has historically been quite basic.

In some respects, MR is still living in the early 90s, just before data warehousing, data mining and predictive analytics began to boom - developments that largely bypassed or were ignored by MR.

Though AI, machine learning, automation, etc., are now the buzz in MR, many marketing researchers are struggling to adapt the new world of marketing and marketing analytics.

52 Things About Customer Analytics (Sherman and Sherman) is a good place to start if you'd like to learn more about some of these new developments. Data Mining Techniques (Linoff and Berry) and Introduction to Algorithmic Marketing (Katsov) delve into these and related topics in more detail.

Data Science and Advanced Analytics aren't for everyone, though, and we need to bear this in mind. Many marketing researchers who have leaned on selling and people skills over the years are going to find the road ahead very bumpy. More and more, you have to be able to walk the walk if want to talk the talk.

✴️ @AI_Python_EN

DeepMind and Google: the battle to control artificial intelligence

Blog by Hal Hodson: https://lnkd.in/eVYhX2e

#artificialintelligence #deeplearning #reinforcementlearning

✴️ @AI_Python_EN

Did you know that Google has recently open-sourced their federated learning framework as part of the TensorFlow ecosystem? It's called TensorFlow Federated or TFF. They also created a simple example which uses the classic MNIST dataset. In particular, they're simulating the data on each client devices and then sample a random subset of the clients that are involved in the training, usually in multiple rounds. The cool thing is that we can use Keras to construct a model and then it can be plug into TFF seamlessly. Very cool! Check it out! Code is also provided! #deeplearning #machinelearning

Article: https://lnkd.in/dM-upwn
TF Federated Learning Page: https://lnkd.in/d4FnthY
Tutorial: https://lnkd.in/dw5DMUW

✴️ @AI_Python_EN

François Chollet Tweeted:
Are you a deep learning researcher? Wondering if all this #TensorFlow 2.0 stuff you heard about is relevant to you?

This thread is a crash course on everything you need to know to use TensorFlow 2.0 + Keras for #deeplearning research. Read on! https://t.co/dFNI2E6yjF https://lnkd.in/fiqShbx

✴️ @AI_Python_EN

An All-Neural On-Device Speech Recognizer #DataScience #MachineLearning #ArtificialIntelligence http://bit.ly/2VWExJQ

✴️ @AI_Python_EN

Creating a good machine learning powered product is a team sport. There is not one unicorn 🦄 that can do it all but you need several roles:

1. 👔 business -> product (product manager/designer)
2. 💡 product -> ML problem (product manager/designer/research scientist)
3. 📚 ML problem -> research paper (research scientist)
4. 👨🏻‍💻 research paper -> code (research scientist/research engineer)
5. 🤖 code -> production (research engineer/data engineer/software engineer)
6. 🔁 repeat 1-5

So don't make the mistake to hire one data scientist and hope that you will become an AI company tomorrow. #deeplearning #machinelearning

✴️ @AI_Python_EN

Before buying or purchasing GPU, do take a look at how they perform do always take a look at what is the balance between what you pay and how much compute power you get.

#deeplearning #computation #performance #gpu

❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_EN

🔸 100 Days of ML Coding
#Machine_Learning #Ml
#machine_learning_algorithms

👉 Link Review

〰️〰️〰️〰️〰️
©️ @Code_Community

WANT TO UNDERSTAND MACHINE LEARNING BETTER? LEARN IT VISUALLY AND TRY SOME PROJECTS

1. LEARN IT VISUALLY
Visual speak louder. Visualization helps us to imagine things and and retain a concepts.

This is will useful to Machine learning concepts. People can easily grasp the idea of an algorithm when presented in an interactive way.

R2D3(http://www.r2d3.us/) is build statistical thinking with interactive design. R2D3's goal is to make machine learning accessible to a broad group of people, including folks who don't speak English. #machinelearning #datascience

You can see here http://www.r2d3.us/

2. TRY SOME PROJECTS

Here's some projects taht writrn by Harun-Ur-Rashid (Shimanto) for his #100DaysMLCOde, you can start also
a.Naive Bayes (NB)(https://goo.gl/xSJZCZ)
b.Sentiment Analysis (https://goo.gl/xSJZCZ)
c. Support Vector Machine(SVM)(https://goo.gl/xSJZCZ)
d. Artificial Neural Networks (ANN)(https://goo.gl/xSJZCZ)
✴️ @AI_Python_EN

Andrej Karpathy’s character level RNN model - is a masterpiece! It is a sufficiently trained model using text generators that gives some eye-popping results. This article by Pranjal Srivastava introduces us to #TextGenerators and their applications in creating a Machine Learning model that can write sonnets just like a Poet !
https://bit.ly/2u7kigq
✴️ @AI_Python_EN

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

Predicting how the stock market will perform is one of the most difficult things to do given the sheer number of factors involved. Can #machinelearning & #deeplearning make a difference? Here are 3 articles to get you started:

1. Stock Prices Prediction Using #ML and Deep Learning Techniques (with #Python codes) - https://bit.ly/2TKqABa
2. Comparative Stock Market Analysis in R using Quandl & tidyverse (Part 1) - https://bit.ly/2u5HlbG
3. Bollinger Bands and their use in Stock Market Analysis (using Quandl & tidyverse in R - Part 2) - https://bit.ly/2JaiFct

✴️ @AI_Python_EN

Glad to see that our #GAN research works enable people to "generate realistic dance videos of NBA players for in-game entertainment." #pix2pixHD, #vid2vid https://medium.com/@getxpire/how-we-used-ai-to-make-nba-players-dance-2fdbe6c63a97

✴️ @AI_Python_EN

Whoa: "Run VS Code on a remote server" https://github.com/codercom/code-server … I haven't been able to fight Emacs+screen+mosh addiction for ~20 years now (well, mosh was recent addition) because I loved being able to move between machines & have my state preserved. I've been waiting for this!

✴️ @AI_Python_EN

In analytics, data typically cannot be used as is, even when "structured." Statisticians and data scientists often spend considerable time on data preparation.

Categorical data, for instance, often needs to be recoded. For example, we may need collapse it into a smaller number of categories for modeling purposes. The new variable must be interpretable as well as useful in the modeling.

Variables are often combined into new ones. This can be done judgmentally (with care) or based on exploratory data analysis (EDA).

Logarithmic and other transformations of data are also frequently necessary for modeling purposes.

Missing data are nearly always a concern, and how they are handled is seldom inconsequential. Some academic statisticians have spent significant parts of their careers on this one topic.

All the foregoing examples require subject matter knowledge and background pertinent to the project. If the project is repeated, then the learnings and code can be leveraged to reduce or even automate data preparation, but the first time around it can consume the majority of our time.

The upside is that we can learn a lot from EDA, as an archeologist can from the location of artifacts and their proximity to others.

✴️ @AI_Python_EN