What are Intelligent Machines?

Machines that use artificial intelligence to learn from the environment and past experiences to perform better in the presence of variability and uncertainty are called intelligent machines. Intelligence develops according to environment and experience. It is human-like behavior. When the computer system or any programmed machine learns from the inputs and corresponding outputs, it can be called an intelligent machine.

Examples of intelligent machines are:

• Virtual personal assistants such as Siri, Alexa, Cortana, and Google Assistant using speech recognition to follow directions.
• Industrial robots having built-in sensors.
• Debit and credit card fraud detection.
• Speech recognition and voice synthesis software on computers.
• Self-driving vehicles using eyesight rather than road markings.
• Diagnostics in healthcare.
• Weapons that can identify targets (smart weapons).

Expert systems are important for intelligent computers because of their complexity of control and ability to emulate human logic skills.

Difference between MI, ML and AI

The terminologies such as Artificial Intelligence (AI), Machine Intelligence (MI) and Machine Learning (ML) are related, but they are not the same.

It's vital that we all grasp the subtle differences between them so that we may better decide which one, or combination of them, is best for a certain activity.

What is Machine Intelligence?

Machine intelligence is a type of sophisticated computing that allows a technology (machine, device, or algorithm) to interact with its environment logically. This means it may perform activities to increase the likelihood of attaining its objectives.

Machine intelligence is a broader term that has machine learning, artificial intelligence and other concepts available as part of the subject.

What is Artificial Intelligence?

Artificial Intelligence (AI) enables a system to perform actions involving cognitive capabilities like reasoning, planning, and problem-solving without requiring an explicit answer for each variant. These algorithms, procedures and approaches enable a computer system to execute activities that typically necessitate a high level of intelligence.

What is Machine Learning?

Machine learning is a set of computational techniques that use data to construct models and predict future data in a complex system. Without being expressly designed for the data they encounter, these models can learn independently and continuously adapt to the changes in settings. Many artificial intelligence systems rely on machine learning.

Note: Based on machine learning and artificial intelligence functionalities, if a computer learns to extract knowledge from various types of data, put together and arrive at its own conclusion, it is known as machine intelligence.

How big data helps in Machine Intelligence?

Large, complicated, organized, and unstructured data sets generated and communicated in real time from numerous sources are referred to as big data. It is a high-volume, high-velocity, and high-variety data asset that allows cost-effective, innovative data processing to improve insight, decision-making and process automation. 

The size of the data collected around the world is growing at an exponential rate. Data has become more meaningful and contextually relevant, taking them out of research labs and into production. The difficulty has shifted from collecting massive amounts of data to interpreting it and turning it into knowledge, conclusions, and actions. Scientific disciplines such as cognitive sciences, biology, finance, physics, social sciences, and businesses believe that data-driven and intelligent solutions can help solve many major issues.

• For improving information retrieval, large data sets of search queries can be used.
• Climate data from the past can be utilized to understand global warming and forecast weather better.
• Drought conditions can be predicted using sensor readings, spectral images of plants, insights into when and how stress affects plant growth and development. This can help combat the problem of world hunger.
• Game data can help robots grasp complicated and unstructured surroundings and develop manipulation skills, while observational data can help robots understand complex and unstructured environments and gain manipulation abilities.

Intelligent control systems

Intelligent control systems are a category of control techniques that employ artificial intelligence computing concepts such as Bayesian probability, fuzzy logic, machine learning, neural networks, and genetic algorithms.

The following are a few common concepts used in intelligent control systems:

• An intelligent control system usually has to deal with a lot of data regarding its own status and the state of its surroundings.
• An intelligent control system can deal with incomplete, inconsistent or noisy data.
• Heuristics are used in intelligent control systems. (A heuristic is a specific strategy for solving a problem that can only be applied to a small set of input parameters)
• An intelligent machine uses a knowledge base to deal with unanticipated or infrequent events.
• An algorithmic control strategy presupposes that all essential data for decision-making is available.

The idea that intelligent machines performing our work will improve our lives is incorrect. Physical exercise, social connectedness, and a feeling of purpose are profoundly linked to human well-being and longevity. It is based on evidence from neuroscience, cognitive science, and health sciences. To improve the human quality of life at the individual and social levels, we require an entirely different type of AI than that in development.

The current focus in AI, ML, and robotics research, development, and deployment is automation. The overarching objective is to replace human labor with robots in every imaginable scenario. But, what will people do if machines do all the jobs in the future? Debates regarding the future of work largely focus on labor economic ramifications, while the implications on health and wellness are just as important. In the future, one may not have reasons to get out of bed in the morning if AI handles imagination and creativity and robots perform the physical work.

Human augmentation (also known as "Human 2.0") is a field that focuses on enhancing cognitive and physical abilities as a natural aspect of the human body. Using active control systems to build limb prostheses with features that transcend normal human performance is one example. Let's take a closer look at it.

Human Augmentation

Human augmentation is complimentary but fundamentally separate use of AI, ML, and robotics. Intelligent technologies that augment human talents, rather than doing labor for people, strive to empower and enhance people to accomplish meaningful work. This idea includes cognitive, physical, emotional, and social augmentation.

Socially Assistive Robotics (SAR) is one of the newest and fastest-growing fields in robotics, that focuses on developing intelligent socially interactive devices to aid people through motivation and companionship rather than physical means. SAR systems help in difficult situations like rehabilitation, behavior treatment, and skill training. These tasks involve interactions with the vast and continually rising population, spanning the age and ability range, such as autism, stroke, ageing, and dementia, to name a few.

Human-machine collaboration is another difficult area of human augmentation as it necessitates more complex interactions and adaptations than automation alone. When humans are removed from a situation, the automation process can redefine and modify a problem to handle it more efficiently and cost-effectively. On the other hand, collaboration necessitates adaptation to the given context and surroundings, as well as to the collaborators. Working successfully with others is a difficult task that humanity continues to research and debate. Human-machine collaboration opens up new possibilities for using human knowledge, skills, and traits in ways that benefit both humans and machines.

Advantages and Disadvantages of intelligent machines

Machine has no inherent intelligence; it simply replies with findings entirely based on its learning. The usefulness of intelligent machines thus depends on how well they are trained, and most significantly, how well they are tested.

Many new technology implementations have been mitigated by establishing controls and processes that ensure the technology's safety. Historically, this has been accomplished by both smart design and disaster response. We can learn from the evolution of technology domains such as airplanes, nuclear power, and medical equipment.

The benefits of intelligent machines are tremendous, and they have the potential to change any industry. Let's see how:

• Human error is reduced.
• Takes risks instead of humans.
• Is available 24 hours a day, seven days a week.
• Assists with repetitive tasks.
• Makes faster decisions.

The drawbacks associated with intelligent machines are listed below:

• Development cost is high.
• It makes humans jobless and lazy.
• It lacks outside-the-box thinking.

Summary

Every discovery or breakthrough has both benefits and drawbacks, but it is up to us to balance them out and cherish only the positive aspects. Intelligent machines have many potential benefits; the only challenge is to keep the "rising of the robots" under control. Some argue that such machines can destroy human civilization if they fall into the wrong hands. Nonetheless, no artificial intelligence program developed till today can destroy or enslave humans.

Common Mistakes

Intelligent machines also make errors such as:

• Image recognition is a challenge for intelligent machines.
• They occasionally cause ethical issues among personnel in the military.
• In real-world testing, self-driving cars overran red lights.

Context and Applications

This topic is significant for the professional exams of graduate and postgraduate courses,
especially:

Bachelor of Arts in Computer Applications

Bachelor of Technology in Computer Science and Engineering

Bachelor in Computer Science

Master in Computer Science

Master of Technology in Computer Science and Engineering

Related Concepts

Machine learning: The most mathematically involved machine learning algorithm

Artificial intelligence: The dark sides of a career in AI/machine learning

Machine intelligence: If machine intelligence is created, will there be multiple machine intelligence or only one?

Practice Problems

Q.1 Which of the following does not reflect an intelligent machine?

(A) Siri

(B) Alexa

(C) Cortana

(D) Web browser

Correct Option: (D)

Q. 2 HA stands for

(A) Human Argument

(B) Human Augmentation

(C) Human Approach

(D) None of the above

Correct Option: (B)

Q. 3 Which one of the following is a drawback of intelligent machines?

(A) Assist with repetitive tasks

(B) Reduce human error

(C) Make faster decisions

(D) Lack out of the box thinking

Correct Option: (D)

Q. 4 Artificial intelligence enables a system to perform tasks such as:

(A) reasoning

(B) planning

(C) problem-solving

(D) All of the above

Correct Option: (D)

Q. 5 If a computer learns to extract various types of data, put together its own processes, and arrive at its own conclusions, it is known as:

(A) machine intelligence

(B) system intelligence

(C) computer intelligence

(D) psychomotor learning

Correct Option: (A)