What is Machine Learning?
Machine learning is a field of computer science concerned with programs that learn.
The field of machine learning is concerned with the question of how to construct computer programs that automatically improve with experience.
— Machine Learning, 1997.
Nik Bear Brown is an Assistant Professor at Northeastern University. His PhD was in computer science from UCLA. His major field was computational, and systems biology and his minor fields were artificial intelligence and statistics. His post-doc was in deep learning at Harvard Medical School. He has taught computer science, statistics, applied mathematics, programming, 3D visual effects, web programming, server administration, networking and game programming at Northeastern University, UCLA, Santa Monica College, ITT and the Art Institutes – Hollywood.
Q. How can people start learning Machine Learning?
NIK: Start with Udacity Nanodegree programs. They have several excellent offerings in python, AI and machine learning and are reasonably priced. Do not do these expensive pseudo-schools like Thinkful, Metis and General Assembly. They are expensive and in my opinion not worth the price tag. Find at least one person to work with. Find a group and enter Kaggle competitions. Once you are certain that this is for you, consider a Masters program in a top-50 school.
Machine Learning (Deep Learning & Artificial Neural Networks, ANN)
Neural networks are a set of algorithms, that are designed to recognize patterns, similar to how a brain functions. They interpret sensory data through a kind of machine perception, labeling or clustering raw input and as with “humans” this learning process allows the machines neural network to learn from it’s mistakes getting better over time.
There are three types of deep learning with ANN
- Unsupervised Learning – using only inputs. Network attempts to group similar inputs and put them into categories.
- Reinforcement Learning -training sets using inputs but then it might also be offered additional info on whether the results were “right or “wrong” and the nature of the mistake that was made.
- Supervised Learning – training sets uses both inputs and the desired “goal” or output. In this way, the computer can learn from it’s mistake and make the necessary adjustments to obtain the desired outcome.
Below is an example of a machine learning how to park two cars using the same neural network. Self-driving cars see the world using sensors.
Q. What are the main places to learn machine learning?
NIK: Meet-up, YouTube, Udacity, online platforms such as CGCircuit and online Masters Programs from top-50 schools.
Khan Academy – for the math needed. Calculus, Linear Algebra, Probability & Statistics.
Fast.ai – Offers Practical Deep Learning for Coders, lots of great resources for Python etc., and a great place to start.
CGCircuit – Offers courses on Python more! As well as, Nik “Bear” Brown will be teaching a series of tutorials soon on Deep Learning in Game Analytics. So stay tuned!
Google’s Python class – free class for people with a little bit of programming experience who want to learn Python
MIT Online 6 week course in Machine Learning.
Stanford – (Coursera.org) – offers free course on Machine Learning.
LA Machine Learning (meet-up) a large variety of machine learning topics such as classification, clustering, neural networks, graphical algorithms, information retrieval, search, game theory, computational learning theory, reinforcement learning, collaborative filtering etc. More recently, a Data Science Track
Annual Big Data and Artificial Intelligence Conference by RMDS (meet-up) This is a group for anyone interested in AI/ ML/ Data Science, Big Data, NoSQL, Data Analytic, Spark, Data Mining, Data Visualization & more. All skills levels are welcome. This group was started to meet other Data Enthusiasts.
Q. How are companies using Machine Learning?
NIK : There are far too many applications to list. Any time a prediction or decision is being made AI and Machine Learning are increasingly being used to make or assist in those decisions. Does the spot look cancerous? Should a buy this or that? Who should I hire? How do I get from here to there?
Q. Examine how Games are using Machine Learning from a business point of view?
NIK: To create content. To create automated NPC behavior. To sell and advertise games. To understand the sticking and easy parts of a game.
Q. Examples of how companies should use Machine Learning in their business?
NIK: Depends on the business. Whatever decisions a business makes a lot are a candidate for machine learning and AI. Hedge funds use to decide what to buy and sell. Ad companies use it to optimize return on ad buys. At the Computational Radiology Laboratory at Harvard it is used to assist radiologists with finding tumors in image data. At the Broad Institute of MIT and Harvard the Cancer Genome Atlas Program has analyzed over 2.5 petabytes of genomic, epigenomic, transcriptomic, and proteomic data to improve our ability to diagnose, treat, and prevent cancer.
SOCIAL MEDIA INFO
Nik Bear Brown, PhDIf you have any questions or feedback on this blog, we’d love to hear it in the comment section below or email me at lori@cgcircuit.com
About the Author:
Lori Hammond, CG Circuit
Author & Content Producer
Experienced multi-talented Artist/Designer/Blogger with an extensive background in the Arts & Entertainment Industry(Animation, VFX, Game & Product Design)
