The use of artificial intelligence (AI) is changing everything around us and providing organizations across industries with impressive new capabilities and insights.
Mourad Oulid-Aissa, PhD, Department Chair of South University’s Master of Science in Information Systems & Bachelor of Science in Information Technology online programs, is working on a textbook to teach technology professionals, leaders, educators, and students about this important field. The book will focus on the domains of machine learning, natural language processing and data science. Dr. Cheranellore Vasudevan, Technical Lead and Strategist of AI and Data Science at IBM (and an Adjunct Professor at South University, Austin) is co-authoring the textbook, as is Dr. Mamnoon Jamil, Associate Director of Analytics at Bristol Myers Squibb and Adjunct Professor at Rutgers University.
Although the book is still being drafted, here is an introduction to the areas the work will explore.
What is machine learning?
Machine learning enables computer systems to learn automatically without being explicitly programmed. Implementing machine learning involves creating computer programs that access and use data to learn and improve with no human assistance. Machine learning is broadly categorized as supervised or unsupervised.
Supervised machine learning
In supervised machine learning, past learnings are applied to new data to predict future outcomes or output. For this process, a system first analyzes a dataset of known input-output pairs. (This dataset is called a training dataset because it is used to train the computer system.) Based on this initial analysis, the system then attempts to predict output values for all new input.
Image recognition is one scenario where supervised machine learning is common. Take the example goal of correctly labeling images of animals. First, the system would analyze images of animals (the input) that are already associated with the correct labels like dog, monkey, cat, bird, horse, etc. (the output). From this analysis, the system would learn what traits make each animal distinct. The system would then apply these learnings to analyze and label all new images provided without labels. In our daily lives, this same technique can be used for everything from facial recognition on social media (think of when Facebook recommends you tag a friend in the picture) to identifying a medical image as benign or cancerous. Likewise, you can train systems for spam and fraud detection, voice recognition, and much more.
Unsupervised machine learning
In comparison, unsupervised machine learning occurs when the information used to train a system is neither classified nor labeled. In unsupervised learning, systems analyze unorganized
data to find and describe hidden structures within the data. Take the earlier example of image recognition. In this case, a computer system might be given a dump of images that have not been labeled in any way. The system might then look for similarities and return options for how someone might want to organize and group the images. Of course, this has many more impactful uses. For businesses, unsupervised machine learning can analyze customer actions to identify customer segments for marketing, or find online buying patterns that power recommendations to customers based on past purchases. (Amazon’s recommendation engine is one example.) In the medical field, unsupervised machine learning might identify patterns in DNA or medical images that predict for genetic diseases. As with supervised machine learning, the potential for unsupervised learning is vast.
What is natural language processing (NLP)?
Spoken and written language are essential methods of expressing human thoughts and emotions, and a major portion of worldly knowledge lives within our written text. Naturally, when computers can interact intelligently with speech and text, many opportunities open up! That’s where natural language processing (NLP) comes into play. NLP involves developing software tools and programmable techniques to process and understand written and spoken language.
Examples of NLP’s capabilities for spoken language include things like asking Amazon Alexa to turn off your lights, telling Siri to call your boss, or using automatic captioning for YouTube videos. For written documents, NLP can be used to summarize key points, assess the sentiment of a text, and extract key information such as price, location, or individuals involved. It can translate text from one language to another. NLP can even be used to field questions and provide basic customer service online via chat bots.
What is data science?
In every organization, data helps business leaders make decisions based on insights, facts, statistical numbers and trends. Data science is the collection and study of data to extract useful information and knowledge. Data scientists record, interpret and manage data to help solve complex problems, from understanding disease trends and treatment efficacy to making package delivery routes and methods more efficient. In the worlds of marketing and finance, data can help organizations predict small- and large-scale behaviors and trends. People even use data science to optimize team performance in sports. Nearly everywhere you look, there’s something that data can help improve!
Learn more about the future of technology!
Interested in learning more about information technology and how it can be used to support business needs? Earning a Bachelor of Science (BS) in Information Technology or a Master of Science (MS) in Information Systems degree may be right for you. Our MS program even includes an elective course dedicated to the topic of AI and cognitive computing. Explore our programs at South University to learn more!