The evolution from artificial intelligence to machine learning to data science
Photo by DeepMind on Unsplash
The past few years have seen many breakthroughs and discoveries in the fields of artificial intelligence (AI), machine learning (ML) and data science. These domains intersect so much that they have become synonymous. Unfortunately, this created some ambiguity.
This guide aims to clear up the confusion by defining terms and explaining how they are applied to business and science. We won’t cover them in depth; however, by the end of this article, you should be able to distinguish between these concepts.
As a field, AI focuses on creating flexible automated systems. The ultimate goal of AI is to build systems that can operate intelligently and independently like human beings can. As such, AI must be able to mimic some of the senses that human beings possess.
They must at least be able to hear, see and sometimes smell touch and smell. The AI must then be able to interpret the stimuli received by these senses and react accordingly. So, different fields and branches under the AI umbrella are dedicated to giving machines and systems these capabilities.
Main branches of AI
The main branches of AI are:
- Machine Learning (ML)
- Deep Learning (DL)
- Natural Language Processing (NLP)
- Fuzzy logic
- Expert systems
- Neural networks
These concepts are not separate areas of artificial intelligence, but make modern and future implementations of AI possible.
The three phases/stages of AI are:
- Artificial Narrow Intelligence (ANI) is the current stage of artificial intelligence. It is also known as weak AI and describes AI systems capable of performing a limited set of defined tasks.
- Artificial General Intelligence (AIG): We are slowly approaching this stage, also known as strong AI. He describes AI as able to reason as well as human beings. Some academics believe that the AGI label should be limited to sentient AI.
- Artificial Super Intelligence (ASI): This is a hypothetical stage of AI where the intelligence and capabilities of computers exceed those of human beings. For now, ASI does not exist outside the realms of science fiction.
The above information may seem a bit jargon-laden and esoteric for professional users. How does this translate to the real world and how is AI applied?
Common applications of AI
- Image processing functions in photo editing software
- Customer Engagement Service
- Social media algorithms
- Online advertising platforms
- Translations provided by natural language processing
- Robotic Process Automation (RPA)
- Marketing and analysis of product usage
- Non-playable characters and enemies in video and text-based games
- AI improvements in augmented reality (AR)
- Sales and trend predictions
- Autonomous cars
- Traffic detection
The term machine learning (ML) is often used interchangeably with artificial intelligence. Although they are not the same thing, they are closely related.
Applications and software run on mostly fixed code. This code contains a limited set of parameters that can only be changed when a programmer changes or adds them. Machine learning aims to make computing more flexible, allowing software to modify its source code at will. It’s similar to how when a person learns something new, they change their brain structures in subtle and drastic ways.
Main branches of ML
The four major branches of machine learning are:
- Supervised teaching
- Semi-supervised learning
- Unsupervised learning
- Reinforcement learning
Of course, there are subsets and new paradigms such as reinforcement learning, dimensionality reduction, etc. Machine learning is usually implemented using a model.
Types of Machine Learning Models
- Artificial neural networks
- Decision trees
- Support vector machines
- Regression analysis
- Bayesian networks
- Genetic algorithms
- Federated learning
- Reinforcement learning
Deep learning is one of the most well-known and widely used subsets of machine learning. It basically consists of a multi-layered neural network. Neural networks attempt to mimic cognition by closely mimicking the structure of the human brain. They are considered the most viable routes to AGI.
Applications of machine learning in business
Here are some examples of the use of machine learning in commercial and consumer products:
Product recommendations are arguably one of the most popular applications of ML and AI, especially in e-commerce. In this app, a merchant’s website or app tracks your behavior based on your activities using machine learning. These activities may include your previous purchases, search patterns, clicks, shopping cart history, and more. The merchant will then use an algorithm to create personalized product recommendations.
With the implementation of machine learning in finance and banking, financial institutions have been able to uncover hidden patterns, detect suspicious activity, and anticipate clerical errors before it is too late. Capgemini, a technology consultancy, claims that a well-trained machine learning solution can reduce all fraud incidents by 70% while increasing transaction accuracy by 90%.
Machine learning has improved the rate of anomaly detection in medical diagnostics, allowing doctors to make more accurate diagnoses. Recently, ML-powered software has been shown to diagnose patients more accurately than experienced physicians. It does this by processing medical records and evaluating changing parameters in real time. Its ability to quickly adapt to changes in the environment is one of the biggest advantages of machine learning in healthcare.
Data science is a broad term that refers to all facets of data management, including collection, storage, analysis, etc. As such, it is a field that involves several disciplines, including:
- Computer science
- Data analysis
- Domain knowledge
- Information science, etc.
An estimated 2.5 quintillion bytes of data are generated daily (globally). Much of this data is unstructured and noisy. Much of the effort of data scientists goes into structuring, sorting, and gaining insights from that data.
Because data science is a multidisciplinary science and not a concept, it cannot be categorized in the same way as artificial intelligence and machine learning. However, let’s expand on the different professions involved in data science before discussing how it can be used in a business context.
Most Important Professions in Data Science
Some of the most common types of data scientists include:
- Machine learning scientists
- Data Engineers
- Software engineers
- Actuarial scientists
- Digital analysts
- Business analysts
- Spatial Data Scientists
- Quality analysts
It is recommended that Data Scientists be able to develop software (code), use analysis tools and software, develop predictive models, analyze data integrity and quality, and optimize the flow of data collection.
Applications of data science in business
Data science has been an extremely useful tool for businesses. Much of the data generated daily is data on potential consumers. For example, a machine learning implementation might process old medical records or observe and gather information about user behavior. It is a form of data mining. Other ways to apply data science in business include:
- Targeted Advertising: Companies such as Google, Facebook, and Baidu derive the bulk of their revenue from digital advertisements. Whether you run a blog or an online store, you can use data science to perform customer segmentation or grouping before publishing targeted advertising campaigns. The best way to perform clustering and grouping is to use an unsupervised ML model.
- Sales forecasting for inventory management: You can use predictive data science models to predict future sales. Predictive models attempt to predict future sales based on historical data.
- Recommendation engine for e-commerce: You can use data science to create personalized product recommendations for loyal customers by looking at their purchase history.
Data science (mainly implemented by data analytics) can also be used in business intelligence. Businesses can extract valuable information from data warehouses and use it to make informed business decisions.
The above guide serves as a simple introduction that mainly highlights the differences between artificial intelligence, machine learning, and data science and how they can be applied in a business context. To learn more about these topics, you can check out one of KDnuggets’ many guides and articles on these topics.
Nahla Davies is a software developer and technical writer. Before dedicating her full-time job to technical writing, she managed — among other intriguing things — to serve as a lead programmer at an experiential brand organization Inc. 5,000 whose clients include Samsung, Time Warner, Netflix, and Sony.