digiGeek.ch
Industry 4.0, AI & ML

AI, ML & Co.
by digiGeek

"Enhance your human capacity with AI, ML and PA ! See your opportunities rather than barriers. Be mindful and take ethics into account." - by digiGeek

Artificial Intelligence (AI)

Artificial Intelligence (AI) has been used since 1956 for Machine Intelligence (MI) to distinguish Natural Intelligence (NI) of humans or animals from intelligence displayed by machines.
Basic Concepts & Methods of AI were already there
- mathematics and mathematical optimization
- statistics and probability
- decision trees and search algorithms
- economics
but since computers of those days were expensive, data storage limited and their AI outcome poor, AI was nothing but Science Fiction for more than a generation.

Today, thanks to high-speed internet, intelligent sensors, affordable multi-core- and in-core-processing, and smartphones everywhere, we capture, store and process vast quantities of data and have them available and processed on the go for no price.

We now have Intelligent Agents (IA) that perceive their environment and take actions such as learning, problem solving, reasoning, planning, natural language processing, knowledge, perception and the ability to move and manipulate objects.

AI has become routine technology in
- Interpreting complex data, images and videos
- Optical character recognition
- Competing at a high level in strategic systems
- Understanding human speech
- Gaming and simulations
- Military simulations and war as a game

Machine Learning (ML)

Machine Learning (ML) & Computational Learning (CL) are methods within the field of Data Analytics and ML is a computer science based on pattern recognition and computational learning.

ML is related to
- Computational statistics
- Mathematical optimization
- Data mining, unsupervised learning
- Establish baseline behavioral profiles, unsupervised learning

ML is used where designing and programming explicit algorithms with good performance is difficult or infeasible, such as
- Malicious insiders working towards a data breach
- Optical character recognition (OCR)
- Data contents and E-Mail filtering
- Detection of network intruders
- Learning to rank
- Computer vision

Effective machine learning is difficult because finding patterns is hard as often not enough training data is available.
So, machine-learning programs may fail to deliver.

ML algorithms give computers the ability to learn from sample data by building models and make sample data-driven decisions or predictions (prediction-making) without being explicitly programmed. This is also known as Predictive Analytics.

Automated Data Analytics (aDA) & Predictive Analytics (PA)

Machine Learning (ML) is a method within the field of Data Analytics. As these ML algorithms are automated, it's automated Data Analytics (aDA). Such aDA algorithms give computers the ability to learn from sample data by building models and make sample data-driven decisions or predictions (prediction-making) without being explicitly programmed. This is also known as Predictive Analytics.

Predictive Analytics is aDA and used to produce decisions & results through learning from historical trends in the data.

Do not be afraid ! Let your data scientists understand your business well (incl. business issues).
Develop trust on recognized results based on approved algorithms, and then step-by-step automate manual steps to enhance AI.
Improve trust of executives being prepared to trust decisions from algorithms.
The development will take time and might require cultural change, internally and externally.

In case of questions, don't hesitate to contact us from www.digiGeek.ch !

John

Matthias Seiler

CEO & Founder of digiGeek

Related Topics:
"Company Readyness for AI"-Survey 2017