Today, I will be talking about AI/ML best practices on weekly datascience Dialogue - Edition 3

Lets learn Django RestFramework. I have alreay explained DRF archtecture here. Will be adding subsequent tutorial video everyday. Remember i dont want to post every youtube links here .... Maximum i will try share knowldge through text and other ways.

8 reasons machine learning projects fail——

🔸 Doing ML for wrong reasons
🔸 ML not needed
🔸 Bad data
🔸 Poor problem framing
🔸 Model ≠ product
🔸 Bad infrastructure
🔸 No trust from stakeholders
🔸 Production failures

Skills to impress data science employers:

🔸 Data engineering
🔸 Model deployment
🔸 Cloud-based services
🔸 Infrastructure as code tools
🔸 Communication and storytelling

The last one is especially important

AI investment in drug design and discovery increased significantly: “Drugs, Cancer, Molecular, Drug Discovery” received the greatest amount of private AI investment in 2020, with more than USD
13.8 billion, 4.5 times higher than 2019.

Don't have enough data to train your model? Fret not! Use the synthetic one!

👉🏼 Synthetic data is artificially generated data that is not collected from real world events! It replicates the statistical components of real data without containing any identifiable information, ensuring individuals' privacy.

🧠 Synthetic data can be used for many applications:
- Privacy
- Removing Bias
- Balancing Datasets
- Augment Datasets

👉🏼 Where to generate it from and how?

Open Source Project YData Synthetic: This repository contains material on GANs for synthetic data generation, especially regular tabular data and time-series. It consists a set of different GAN architectures developed using Tensorflow 2.0. An example Jupyter Notebook is included, to show how to use the different architectures.

Link: https://github.com/ydataai/ydata-synthetic

Code generation using machine learning. You draw ✍️ and Machine can create a code for you.....

Did you know

For most Python implementations different threads do not execute at same time: they merely appear to.

Threads may be running on different processors but they merely appear to.

But how to achieve multiple tasks running simultaneously?

Just think 🤔 will share answer tomorrow if none responded to this.

Encoding methods for categorical features:

🔸 One-hot encoding
🔸 Label encoding
🔸 Mean target encoding
🔸 Weight of evidence
🔸 Embeddings

Encryption in Python. These are the modules you can use for better encryption process. I will be sharing good video and document by tomorrow

Encryption and Decryption in Python

Documentation
https://github.com/iampython-team/PythonAdvancedConcepts/blob/main/Topic_1_Encryption/Encrption.pdf

Source Code
https://github.com/iampython-team/PythonAdvancedConcepts/tree/main/Topic_1_Encryption

Video
https://youtu.be/bE7fl6qN-LY

Please follow the github account and will be adding more projects everyday including doc

pandas cheetsheet

Difference between loss function and cost function
A loss function is for a single training example. It is also sometimes called an error function. A cost function, on the other hand, is the average loss over the entire training dataset. The optimisation strategies aim at minimising the cost function.