Big data, machine learning, predictive analysis, chatbots … are new concepts that we must know and understand before jumping into ‘data’ to explore the new types of artificial intelligence linked to the field of human resources.
As we enter into unknown territory we must be well equipped to understand the concepts of artificial intelligence.
But before talking about the terms, knowing what artificial intelligence is and its different types seems like a fundamental step.
Thus, a definition of artificial intelligence to approach this concept is “the discipline that unites sciences such as logic, computation or philosophy with the aim of creating artificial entities capable of solving problems by themselves”.
But is there a way to simplify the definition of AI? Yes, indeed. Machines that think like humans.
Once this is clear we can then identify the most important types of artificial intelligence coined by Stuart Russell and Peter Norvig:
- Systems that think like humans– These are the systems that try to emulate human thinking, such as problem-solving or learning.
- Systems that act like humans– The clearest example of this type of artificial intelligence is robotics. This case studies how computers usually perform tasks linked to humans.
- Systems that think rationally– Those who try to imitate the rational logical thinking of the human being.
- Systems that act rationally– It is the type of system that try to achieve specific objectives based on beliefs. They have qualities such as reasoning, learning or perception.
On the other hand, what are some of the concepts you should be aware of?
Big data: Literally “big data” is defined as a set of data of such magnitude that traditional IT management tools are not able to manage them. As a recruiter, you are one of the professionals that manage large amounts of data: employee administrative information, salaries, skills assessment,…
Machine learning: People often talk about learning by example. This is the basis of machine learning, which represents the set of processes that allow intelligent machines to learn according to the rules pre-established by algorithms that feed on examples. As in any formation, one always begins by imparting the theory, examples are given, the rules are marked, and sometimes the exceptions applied to the rule. Then, it is time for practice and, the more experience we accumulate, the more we learn. Well, machine learning works exactly in the same way.
Deep learning: It is a subfield of machine learning that gives autonomy to artificial intelligence machines to treat a large volume of data hierarchically. To do this, the system begins by analyzing the data in a simplified way and gradually refines the search until reaching more concrete results. During an interview, the first things that are detected are the basic social cues. You will know what we mean when we talk about these small gestures of communication that arouse your attention quickly: a stutter, a frown, a silence or a well-chosen word …
These elements of artificial intelligence will accumulate in your mind and you will begin to establish associations that will lead you to form an opinion about the candidate. This will ultimately motivate your decision during the recruitment process. Deep learning can reproduce the same process.
Decision tree: It is the graphic representation of the rules that lead to decision making in the form of a tree. Each branch represents the different possible decisions, which in turn can lead to other options and so on, until the final conclusion. For each of the branches, there are associated probabilities. In HR, when you classify CVs, you apply specific screening filters for each position.
Supervised or unsupervised learning: what distinguishes these two methods is the intervention or absence of the human hand in the machine learning process. In the first type of artificial intelligence, a person writes down or classifies the data to create samples that will guide the machine. In the second, the computer will have to search on its own among a large volume of data from diverse sources. As an HR professional, you probably have many hiring processes behind your back.
In supervised learning, all the information, from the pre-selection of curricula to the interview with the final candidate, are examples of which an algorithm can be fed to try to find the logic behind your decisions, and thus, reproduce them.
On the contrary, the unsupervised one will not take into account the HR actions and will organize the CVs in different categories (business profile, technical profile, junior, senior …) with the risk of making classification by irrelevant categories from a practical viewpoint.
Neural networks in artificial intelligence: They are algorithms that can schematically mimic the biological neural networks, initially designed to model data processing. These operational rules are based on statistics and are learned as they occur. For example, they are applied in the recognition of shapes and images, stock markets or medical diagnoses. In an interview, when you receive a candidate, he offers you a lot of information (their degrees, experience, technical skills …). Each of your neurons will apply mysterious mathematical formulas that will weigh each information based on your conscious and unconscious expectations to allow you to make a decision- the candidate continues in the process or not. Natural language processing (PLN) is the form of artificial intelligence that becomes the crossroads between linguistics and computer science. It is the field that combines information technologies (among which is artificial intelligence or machine learning) and applied linguistics. The purpose of making possible the computer-assisted understanding and processing of information expressed in human language for tasks such as chatbots or automatic translators.
Cognitive sciences: Its objective is to describe, explain and imitate the mechanisms of human thought and knowledge. According to the Cervantes Institute, cognitive sciences are the meeting point between cognitive psychology and artificial intelligence; “Its objective is to study the way in which people and machines assimilate new data, process it and act accordingly.”
Conversational agent or chatbot: They are interactive dialogue systems that manage the interactions between man and machine through artificial intelligence. The interface is capable of conducting a dialogue, asking questions and providing answers according to predefined rules. In HR, the recruiting agents can use these virtual agents capable of interpreting the language to collect information from the candidates, formulate selection questions and answer precise questions, which can be sent directly to the Applicant Tracking System (ATS).
Predictive analytics: It consists of the application of statistics and methods of different types of artificial intelligence to predict future events or the evolution of variables. It is based on the predictive hypothesis of assuming that several equal situations will experience the same evolution if this has been the case in the past. This is the type of artificial intelligence that allows us to anticipate hiring at a given year according to the time and rate of rotation, or even determine, within a group of candidates, who will be the most interesting candidates for your company in the future.
So how do all these elements come together?
Now that you know a little more about the concepts and types of artificial intelligence, here we explain how they interact and a summary of all this.
AI combines machine learning, deep learning, and decision trees. Any AI can be supervised or unsupervised and uses algorithms and types of artificial intelligence, such as neural networks, which are the basis of deep learning. In this way, artificial intelligence systems such as chatbots or predictive analyses are based on data or even big data.
In addition to this, there can be no precise deep learning without big data.
Welcome to the Future of HR!