AI, Machine Learning, Deep Learning — What are the differences?

Source: Deep Learning on Medium

BEGINNER’S GUIDE

AI, Machine Learning, Deep Learning — What are the differences?

If you are new to data science or confused about the definitions of Artificial Intelligence, Machine Learning, Deep Learning.

Artificial Intelligence

AI is the abbreviation for Artificial Intelligence. Throughout the history of AI, researchers have divided into 4 distinct approaches:

  • Acting humanly
  • Thinking humanly
  • Thinking rationally
  • Acting rationally

The most challenging task in this study is programming computers to effectively mimic human behaviors, in both actions and thoughts. AI researchers also try to instill the following features:

  • Reasoning: The ability to solve problems in a logical way.
  • Knowledge: The ability to digest information in a context (like analyzing how many objects, events, obvious situations in real-world and classifying based on the characteristics of corresponding objects).
  • Planning: The ability to create and pursue a goal based on the knowledge.
  • Communication: The ability to understand human written and verbal communication.
  • Perception: The ability to understand the context from images, audio, and other kinds of input.

Turing Test — Acting Humanly

Picture by SearchEnterpriseAI.

In 1950, a famous mathematician named Alan Turing proposed an approach to determine if AI acts like a human. The proposal was known as the “Turing Test”.

According to Turing, the purpose was to examine how smart the system acted. The tests were performed in discussion-like way: one person asks a question, another person answers the question and the computer is expected to respond to the question. During the discussion, if the questioner cannot differentiate whether the answers come from human or machine, the machine can be considered “smart”.

This approach has been bringing remarkable achievements:

  • Natural Language Processing: A computer is able to understand text or human voice and respond back.
  • Knowledge Representation: The ability to memorize knowledge via vision, audio, or document.
  • Automated Reasoning: The machine is capable of using its knowledge to answer a question or make a decision.
  • Machine Learning: The computer is able to adapt to a context, then conclude new principles from acquired knowledge and use it for decision making purposes.
  • Computer Vision: The capability of observing and determining objects.
  • Robotics: The machine is able to interact with nearby objects as well as moving in an area.

Cognitive Science — Thinking Humanly

Thinking like a human is another approach. Researchers will need a mega library to work on this field. If we want a computer to think like a human, we first need to understand how humans think.

There are 3 ways to examine:

  • Observe the thinking process
  • Observe a person’s actions
  • Observe the brain activities

If we give the computer a human similar input and receive a human similar output, we may consider it is thinking like humans at some point. There is a belief that if we are able to combine the results from the Turing test and cognitive science, we can achieve the AI goal faster.

Logic — Thinking Rationally

However, it is a broad and challenging task to say we want a computer to think like a human, so we narrow down the problem to: thinking rationally is enough.

In the 19th century, the logicians invented mathematical notations to describe a statement or relationship among objects in the world, which were useful in training computers later on. However, this approach was a rough path. The inventors faced 2 challenges:

  • First, the computer is incapable of digesting informal knowledge and expressing it in formal terms, especially the knowledge was not assured of 100% certainty.
  • Second, there is a big difference between solving a problem in theory and solving it in the real world. We need a large dataset for the input, which can exhaust the resources of any computers, even if the input is a problem with few hundred facts.

Agent — Acting Rationally

Obviously, computer programs must do something, but computer agents are expected to satisfy more requirements. They must operate autonomously, collect information from the surrounding environment, survive for a long time period, adapt to any changes, make plans, and try to achieve them.

A typical example is the NASA robots which were launched to planets with unpleasant conditions and received few or no commands from human. How can they collect the soil, survive in the stormy, windy climate, make its own plan for charging the solar battery in order to continue the mission?

Machine Learning (ML) and Deep Learning (DL)

ML is just a part of the “thinking like a human” approach, and this approach is just one of 4 ways to research AI.

To reach the ML goal, researchers have been inventing algorithms as well as different ways to deal with problems:

  • Supervised learning: decision tree, k-NN, Naive-Bayes, SVM, neural network, deep learning, etc.
  • Unsupervised learning: k-means, hierarchical clustering
  • Reinforcement learning: passive/active/generalization

Once again, we can see deep learning is only one of the methodologies to solve problems using supervised learning. So why is deep learning such an important topic in our industry?

Deep Learning:

Strength: Deep learning is a technique to implement machine learning to achieve state-of-the-art results of computer vision or speech recognition. Deep neural networks work effectively with images, audio, and text data, capable of updating the model with new data through batch propagation. The architecture of the model (quantity and structure of layers) can be applied to different problems, particularly, the hidden layers play a role in reducing the cost of feature engineering.

Weakness: Deep learning is not a Swiss army knife, it needs a large dataset of a certain area to be trained. Actually, deep learning can be defeated by tree ensemble method in some basic machine learning tasks. In addition, training a deep learning model is expensive. It requires powerful resources as well as high-quality specialists to perform tuning hyperparameters (number of layers, number of nodes per layer, learning rate).

Conclusion

I hope my readers have a clear picture of the definition of AI, ML, and DL. If you want to step into the AI world for your career, ask yourself which approach you aim at (thinking or actions) and the level of human-like (just enough or near-human). Those AI systems require a lot of effort to complete but also a high risk of failure. However, if they are successful in the mission of serving humankind, you will change the world. AI is the future!