Machine Learning – Deep Learning
By Kaitlin Logie, December 20th 2016.
What is machine learning and what are we doing when we program machines to “learn”?
Machine learning is considered a type of artificial intelligence (AI) that provides the ability for computer’s to ‘learn’ from data they are exposed to without explicitly being programmed to do so. Machine learning focuses on the development of programs that have the ability to change when exposed to new/different data. Similar to data mining, the machine learning process searches through data and looks for patterns and uses that data to detect patterns in other data and adjust program actions accordingly. There are 2 categories of machine learning algorithms:
- Supervised – apply what has been learnt in the past to new data.
- Unsupervised – draw inferences from a data set.
So instead of trying to write a program, we try to develop an algorithm that a computer can use to look at hundreds or thousands of data samples. Then the computer uses that experience to solve the same problem in new situations.
(image by techspot)
What is deep learning?
Also known as “deep machine learning”, deep learning is a new branch of machine learning based on a set of algorithm that attempt to model high level abstractions in data. These algorithms are inspired by the structural and function of the human brain and are called “called artificial neural networks”. Deep learning can be an intimidating concept, but it’s becoming increasingly important these days and more widely used then ever before.
The most basic foundational unit in a neural network is called a neuron which isn’t as scary as you think!
(image by kdnuggets)
Each neuron has a set of inputs, each of which is given a specific weight. The neuron computes some function on these weighted inputs.
What can i use machine learning/deep learning for in my research?
You probably use many applications of machine learning and deep learning in your everyday life without even noticing it. But these methods are far from perfect and are still widely being developed on and researched around the world.
- Data modelling – a usually slow and tedious process of comparing data can be automated and run efficiently with applications of machine learning.
- Voice recognition.
- Facial recognition.
- Face detection.
- Data searching applications.
- Data sorting applications.
- Anti-virus.
- Genetics – classical data mining or clustering sequencing algorithms. Help find genes associated with a particular disease.
What are some real life applications of machine learning and deep learning?
Facebook! Facebook’s ‘newsfeed’ uses machine learning to personalise each member’s newsfeed. As the member scrolls down their newsfeed and ‘likes’ a specific post, the news feed will start to show more of those post similar to the liked posts and more posts from the page it came from earlier in the feed. Behind the scenes, the software is simply using statistical analysis and predictive analytics to identify patterns in the user’s data and use these patterns to populate the News Feed. Should the member no longer stop to read, like or comment on the posts, that new data will be included in the data set and the News Feed will adjust accordingly.
This is just one example of how Facebook uses applications of machine learning in their process of designing each member’s individual home-space. There are many more applications of artificial intelligence that Facebook uses in every aspect of their website/apps.
(image by atlantablackstar)
Google! Google defines itself as a machine learning company now. It is also a leader in this area because machine learning is a very important component to it’s core advertising and search businesses. It applies machine learning to improve search results and search suggestions. Google also now have a deep learning research project currently being worked on called “Google Brain”. Google Brain uses large scale brain simulations for machine learning and AI.
(image by si)
How do I get started?
There are many resources available to you online, or on campus at the University of Auckland. Contact the Centre for eResearch for help getting started or come along to one of our Hacky Hours and we will put you in the right direction.