Welcome to our eNewsletter, Spotlight on IT, where we spotlight a different trend or topic in the information technology world. If you would like to receive this publication in your inbox, subscribe here.
Today’s spotlight is on Machine Learning.
Computers that teach themselves? Sounds like something straight out of a sci-fi movie. But machine learning is here and is already a part of many things we do everyday.
What is Machine Learning?
- Machine learning is a kind of artificial intelligence. It allows computers to learn without being explicitly programmed.
- It uses algorithmic models, which are trained to recognize patterns in large amounts of data, to automatically make predictions.
- In other words, machine learning allows computers to learn and adapt.
- Carleton’s School of Computer Science offers several grad courses on the topics of artificial intelligence and machine learning.
Why now? What makes machine learning more feasible?
- There is more and more data at our disposal – both in quantity and variety.
- Computational processing is becoming cheaper and more powerful.
- Data storage is becoming more affordable.
- This makes it possible to create models that analyze complex data quickly and accurately.
- Carleton researchers who need high performance computing resources have access to the Research Computing and Development Cloud.
Examples of machine learning
Many things we do every day are powered by machine learning algorithms:
Social Media
- When you click “Like” on Facebook, your feed is adjusted to display similar content in the future.
- Apps suggest “People you may know” based on your connections to others.
- Image tagging automatically detects yours or your friends’ faces in photos you upload.
Marketing
- Recommendation engines make educated guesses based on what they think you will like.
- For example, Amazon recommends similar products and Netflix makes movie suggestions.
- Web pages and mobile apps display ads based on what you have previously looked at.
Email and Search
- Spam filters learn to classify spam based on an email’s content, subject and how you choose to filter it.
- Every time you do a Google search, the program looks at how you respond to the results – if you click, what you click and if you type in a new query. Google learns to deliver more accurate results next time around.
Transportation
- Apps, such as Waze Fastest Routes, learn which driving routes are best by analyzing driver behavior and outcomes.
- Self-driving cars learn where obstacles, road edges and other cars are by analyzing video frames from an onboard camera.
- A group from BlackBerry and its software subsidiary QNX Software Systems visited Carleton this past spring, to explore collaboration in connected car and autonomous vehicle research… Read more
Machine Learning Research at Carleton University
Yuhong Guo is Carleton’s Canada Research Chair in Machine Learning. She is interested in the question “What if machines could automatically learn useful skills by interacting with data that is already freely available?” Her aim is to develop automatic data analysis and to reduce dependence on human input… Read more
5 Things We Thought you Should Know
1. CAUBO 2017 Quality and Productivity Awards
Carleton ITS and Faculty Affairs were acknowledged for their excellent work on the new online Faculty Careers System. Congrats! The system has increased the efficiency, accuracy and timeliness of managing the faculty leaves process… Read more (story on page 29)
2. Implementing IPv6 at Carleton – a case study
Carleton’s Mike Milne tackled the task of implementing IPv6 here at Carleton. His story was recently published on the TeamARIN blog… Read more
3. Carleton HR moves to intranet
All staff and faculty HR content is now available on the Intranet. The internal site contains all HR policies, documentation and forms (i.e. monthly leave forms, health claim forms, etc.) important to staff and faculty.
4. Latest web template upgrade
Web Services has migrated more than half of existing Framework sites to version 2.0 which features more layout options, streamlined front and back-end editing, and fast loading, search-engine friendly pages. See the new template in action on: Graduate Admissions and Industrial Design.
5. CUAD2017
Despite the rain, Carleton University Appreciation Day was a great success. Thanks to the many faculty and staff who came together to celebrate.