A Guide to Everything You Need to Know About AIOps
Computers have advanced in many ways since their inception and arguably the most important dimension is in the computational power they provide. It is this increase that ultimately enables machine learning solutions. Machine Learning is an application of artificial intelligence that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. ML brings new techniques across various aspects of business and across many different industries. It brings the ability to automate manual and repetitive tasks and improve the process behind decision making.
ML vs AI
You’ve heard the buzz, but you may be asking yourself, what exactly is machine learning? And how does it differ from AI?
ML is a type of computer program or algorithm with the ability to teach itself by analyzing data (inputs) and coming up with a solution (output). What makes machine learning algorithms valuable are the feedback loops–a well-designed algorithm continues to learn from new input data to increase the accuracy of the solution it comes up with. For example, machine learning algorithms in recruiting are used to assess candidates’ personalities, job fit, and resume. ML is essentially statistics and correlations used to make predictions. These predictions become more accurate as more inputs are fed into the system.
3 Types of ML
1. Supervised Learning
In supervised learning, the algorithm is given a set of correctly labeled input/output example pairs in order to train. The two major types of supervised learning are regression and classification. Regression involves predicting a quantity by figuring out which features are important for the outcome. Classification involves assigning observations to different categories.
This allows for questions in three formats. Two-class classification (A or B?), multi-class classification (A, B, or C?), and anomaly detection (is this abnormal?). Using these classifications you can ask and answer questions such as; Is this an image of a cat or dog? What is the mood of this tweet? Or, Is this pressure reading atypical?
2. Unsupervised Learning
While supervised learning finds patterns in data sets where we know the correct answers, unsupervised learning finds patterns in data sets where we don’t. Unsupervised learning allows us to ask questions about how data is organized and how to compress it. This is achieved through techniques such as clustering and dimensionality reduction. For example real estate websites can group their housing listings into neighborhoods so that users can navigate listings easier.
Unsupervised learning allows us to answer two types of questions:
How is this data organized?
How can we represent this data in a compressed format?
3. Reinforcement Learning
In reinforcement learning, we do not provide the machine with examples of correct input-output pairs, but we do provide a method for the machine to quantify its performance in the form of a reward signal. The machine tries a bunch of different things and is rewarded when it does something well. Reinforcement learning is useful in cases where the solution space is enormous or infinite, and typically applies in cases where the machine can be thought of as an agent interacting with its environment. An algorithm that can play video games is an example of reinforcement learning.
Reinforcement Learning is all about the actions the machine can take. The RL algorithm chooses an action based on the factors it has learned to make high reward actions. This process was inspired by how humans and rats respond to rewards and punishment. This makes RL terrific for automated systems. For example an air system can learn to pre-refrigerate the upper floors of an office building before the day gets too hot and takes longer to cool down. This saves money in the long run by having the system use less electricity because it is running and working less.
Here at Blue Orange we have used ML to automate resume screening and shortlist and grade candidates by learning from existing employees’ resumes. First we used a natural language processing ML algorithm to turn the unstructured resume text into relational data. Then we built another ML algorithm that trained itself on prior employees to learn which resume data points (inputs) are correlated with successful employees to produce a shortlist of qualified candidates for the position (output). Instead of just scanning for keywords, we are able to make predictive hiring suggestions to HR. In addition, for firms who use digitized interviews, we can use machine learning technology to assess candidates’ personality and job fit by learning from successful candidates’ facial expressions and word choices.
As technology further integrates into our everyday life, the need for collecting, storing, and analyzing the data we have will only grow in importance. This may seem daunting now, but with foresight ML can be utilized by businesses of all sizes across all industries. So let us tame your business’s data and make it start working for you.