The differences between ARTIFICIAL INTELLIGENCE(AI), MACHINE LEARNING(ML) AND DEEP LEARNING (DL)

Since the improvement of the personal computer in the 1940s, it is a well-known fact that computers can be modified to complete extremely complex errands — such as, finding proofs for scientific hypotheses or even playing chess — with extraordinary capability.

What is the meaning of Artificial Intelligence (AI)?

AI (Artificial Intelligence) is a branch of computer science that focuses on the study and production of highly intelligent machines that work and respond as humans do. Artificial Intelligence is designed to perform various human activities such as:

• Recognition of speech
• Learning
• Solving problems
• Critical thinking

AI branches into:

• Artificial General Intelligence (AGI),
• Artificial Narrow Intelligence (ANI), and
• Artificial Super Intelligence (ASI).

Artificial General Intelligence; also known as Deep or Strong AI, this is when an AI is able to mimic human intelligence and behavior to indistinguishable level. A lot of Computer science experts hold the belief that Deep AI is possible.

Artificial Narrow Intelligence; it can also be called weak or Narrow AI, the AI is programmed to mimic a narrow range of parameters and contexts in relation to human intelligence and behavior. Apples’ Siri is a good example of ANI, it is important to note that even thou the AI parameters are narrow, it is still a complex technology with billions in investment.

Artificial Super Intelligence; this is simply when an AI not only learn to mimic human behavior and intelligence but also surpasses it. At this stage, an AI will only count as ASI if it performs things only humans are capable of such as arts and human relationship.

What is Deep learning?

Deep learning is an aspect of machine learning that show computers how to perform tasks that humans do naturally i.e. learning by following an example. Deep learning is a key technological innovation that drives the technology like driverless cars, allowing them to identify a stop sign, or differentiate between a lamppost and a pedestrian. It is also behind voice control technology we find in devices such as mobile phones, smart TVs, tablets, etc. Deep learning is been given more focus recently for the right reasons as it is reaching goals previously thought impossible.

When we talk about deep learning, we are talking of computer model learning to carry out classification tasks on its own based on text, sound, or images. They are known for their astonishing accuracy, which sometimes supersedes the performance of humans. Deep learning models are trained to utilize a large amount of categorized data as well as neural network architectures that are made of several layers.

How deep learning functions.

Deep learning is considered a subcategory of machine learning that makes use of hierarchical layers of neural networks to perform the task of machine learning. These neural network architectures contain neuron nodes in a web-like connection imitating the human brain design. The way deep learning model is designed allows it to process information in a nonlinear manner.

For example, the conventional way of detecting money laundering or fraud might depend on the amount of money transacted. However, with deep learning nonlinear approach, additional information will be processed like time, IP address, geographic location, and other information that can be used to identify fraudulent activity. The information is processed in layers, the first layer process raw information such as the amount of money transacted then pass it to the next level as an output data. The processing continues at the second layer where extra data are added like location and passes the output to the next layer and so the process continues across all layers of the neuron network until the final output.

What is Machine learning?

Machine learning is the utilization of Artificial Intelligence (AI) that gives systems the capacity to naturally take in and enhance as a matter of fact without being unequivocally programmed. Machine learning centers around the improvement of computer programs that can get to information and utilize it to learn for themselves.

The way toward learning starts with perceptions or information, for example, cases, coordinate understanding, or direction, keeping in mind the end goal to search for designs in information and settle on better choices later on in view of the illustrations that we give. The essential point is to permit the computers to learn naturally without human intercession or help and change activities in like manner.

Some machine learning techniques

Machine learning algorithms are regularly sorted as administered or non-directed.

• Administered machine learning algorithms can apply what has been realized in the past to new information utilizing named cases to foresee future occasions. Beginning from the investigation of a known preparing data set, the learning algorithm creates a deduced capacity to make expectations about the yield esteems. The structure can give focuses on any new effort after competent preparing. The learning algorithm can likewise contrast its yield and the right, proposed output and discover mistakes with a specific end goal to alter the model appropriately.
• Conversely, non-directed machine learning algorithms are utilized when the data used to prepare is neither grouped nor named. non-directed learning considers how systems can derive a capacity to depict a concealed structure from unlabeled information. The structure doesn't take cognizance of the correct output, however, it analyzes the information and can draw deductions from data sets to portray concealed structures from unlabeled information.
• Semi-directed machine learning algorithms fall someplace in the middle of managed and non-directed learning since they utilize both marked and unlabeled information for preparing — commonly a little measure of named information and a lot of unlabeled information. The systems that utilize this technique can impressively enhance learning exactness. For the most part, semi-managed learning is picked when the obtained named information requires talented and pertinent assets with a specific end goal to prepare it/gain from it. Something else, acquiring unlabeled information, for the most part, doesn't require extra assets.
• Reinforcement machine learning algorithms is a learning strategy that collaborates with its environment by creating activities and finds blunders or rewards. Experimentation look and deferred compensation are the most significant attributes of reinforcement learning. This method enables machines and programming expert to consequently decide the perfect behavior inside a particular setting with a specific end goal to augment its execution. Basic reward input is required for the operator to realize which activity is ideal; this is known as the fortification flag.