There are many ways to build a system that can solve problems in AI. Problem-solving techniques that can be used to solve problems in AI include:
- Searching (search)
- Reasoning
- Learning (Learning)
1. Searching (search)
problem solving with searching techniques will be easier if the objects are represented in a graph. Graph representation is done by first creating a representation of the problem object as a node and a representation of the relationship between objects by connecting the nodes. After that, each node in the graph is visited systematically (traverse).
Searching techniques are divided into 2, namely:
- Blind search
- Guided search (heuristic search).
Blind search
Some of the methods available in blind search include:
- Breadth first search
- Depth first search
BREADTH FIRST SEARCH
In the BFS method, all nodes at level n will be visited first before visiting the nodes at level n+1.
Breadth first search
In the BFS method, the visited child nodes are stored in a QUEUE. This QUEUE is used to refer to neighboring nodes to be visited according to the queue.
Following are the steps of the BFS algorithm:
- Insert the root node into the QUEUE
- Take a node from the beginning of the QUEUE, then check whether the node is a solution.
- If the node is a solution, the search is complete and the results are returned.
- If the node is not a solution, put all child nodes into the QUEUE
- If the QUEUE is empty and every node has been checked, the search is complete.
- If QUEUE is not empty, repeat the search starting from point 2.
Advantages of the BFS method
Ensuring the best solution is found (complete & optimal)
Weakness
Because BFS has to store all generated nodes, this method requires a lot of memory and time.
Case 1
- It is known: the tracking tree in the image on the side:
- Question: implement the BFS algorithm to find a solution from the initial node (start) S to node G (goal)
Case 1
Iteration to -- 1
insert node S into QUEUE
picture of the queue:
state space representation: (S)
remove S from QUEUE and check if S is a goal?
it turns out that S ≠ goal
S has children A and B, put them in QUEUE
State space representation
Iteration to -- 2
remove A from QUEUE and check if A is a goal?
It turns out that A ≠ Goal
A has children C and D, put them in QUEUE
State Space Representation
Iteration to -- 3
remove B from QUEUE and check if B is a goal?
It turns out that B ≠ Goal
B has children E and F, put them in QUEUE
State Space Representation
Iteration to -- 4
remove C from QUEUE and check if C is a goal?
It turns out that C ≠ Goal
C has no children, so none are added to the QUEUE.
State Space Representation
Iteration to -- 5
remove D from QUEUE and check if D is a goal?
It turns out that D ≠ Goal
D has no children, so there is nothing to put in the QUEUE
State Space Representation
Iteration to -- 6
remove E from QUEUE and check if E is a goal?
It turns out that E ≠ Goal
E has children H and G, put them in QUEUE
State Space Representation
Iteration to -- 7
remove F from QUEUE and check if F is a goal?
It turns out that F ≠ Goal
F has no children, so none are added to the QUEUE.
State Space Representation
Iteration to -- 8
remove H from QUEUE and check if H is a goal?
It turns out that H ≠ Goal
H has no children, so there is nothing to put in the QUEUE
State Space Representation
Iteration to -- 9
remove G from QUEUE and check if G is a goal?
It turns out G = Goal
Search Stopped.
Looking for Solutions:
- G is the son of E, and E is the son of B, and B is the son of S
- Since S is the root node, the search for a solution is stopped and a solution is obtained.
- S -- B -- E -- G
Artificial Intelligence for Industry & Business Models
The application of AI covers various fields as shown in the branches of the AI tree in the Figure, while the parent science of AI can be seen at the root of the AI tree.
What fields implement AI (Artificial Intelligence)?
Expert system
Expert Systems, namely advisory programs that try to imitate the reasoning process of an expert in solving complex problems.
Expert Systems are the oldest and most widely developed AI applications. So in this case human expertise seems to be transferred to the computer hard disk.
Computer Vision
The tool above is an ultrasound tool.
Other tools in the world of health are:
- Classification and detection (e.g. tumor detection)
- 2D/3D segmentation
- 3D reconstruction of human organs (MRI or ultrasound)
Natural Language Processing
a program that gives the computer the ability to communicate with the user using natural human language such as Indonesian, English, Japanese or others. Example; Google Translate, Google Voice, etc.
Voice Recognition (Voice Recognition)
- Voice recognition is a technique so that computers can recognize and understand spoken language.
- This process allows a person to communicate with a computer by talking to it.
- The term "speech recognition" implies that the main goal is to understand spoken words without having to know their meaning, which is the task of "speech understanding".
- In general, the process is an attempt to translate what a human says into words or sentences that can be understood by a computer.
- This is based on the experts' thoughts about what if the user only dictates sentences through the speaker and the computer can understand the meaning of the sentence being spoken.
Sensor Systems and Robotics
- Robots are a combination of elements such as: mechanical systems, vision and imaging systems, and signal processing systems.
- A robot, which is an electromechanical device programmed to perform manual tasks, is not entirely a part of AI.
- Robots that merely perform actions based on mechanical/electrical switches are said to be stupid robots that are no smarter than an elevator.
- Intelligent robots usually have sensor devices, such as cameras, that collect information about their operations and their environment. The robot's AI then interprets the information and responds and adapts to changes in the environment.
ASIMO is a humanoid robot created by Honda. It uses sensors and intelligent algorithms to avoid obstacles and navigate stairs.
Game Playing
- Game software appears like mushrooms, video games are very popular with adults and children.
- Games are a great area to analyze the intelligence of a computer.
- There are two reasons for this, namely: the game contains a structured pattern to achieve victory or defeat easily, and the game requires the right strategy to win.
- The game can be completed by searching from the starting position to the winning position.
- Before performing a search, the valid positions in the game need to be generated first by a procedure.
- GP is designed to be able to evaluate/search for solutions going forward from the starting position to the position that leads to victory.
- A popular GP is IBM's Deep Thought, a chess program that competed against Grand Master Anatoly Karpov with a draw, amazingly: a computer software can play a draw with a chess professor!
Advantages of AI
- Artificial intelligence doesn't need to sleep,
- Artificial intelligence is never bothered by biological mind problems like humans. For example, rest, toilet and eating.
- When making a decision, humans are sometimes based on emotion rather than logic. This is different from artificial intelligence, where the thinking pattern is more about logic
- Artificial intelligence applications are easy to duplicate. Human intelligence is difficult to transfer to other humans.
Disadvantages of AI
- Artificial intelligence is meant to help human work. But behind that, the implementation of artificial intelligence has caused layoffs in several countries.
- For example: in 2016, Foxconn, a supplier of Apple components, announced that it would cut 60,000 employees, or more than half of the 110,000 employees employed at its Kunshan, China factory, and the related jobs would be replaced by an army of robots.
- Research at Deloitte Touche Tohmatsu and Oxford University estimates that by 2033, 45% of human jobs will be replaced by robots, and this involves the livelihoods of 10 million people.
- Oxford University lists 10 types of jobs that will face this unfortunate fate, including assembly workers, secretaries, non-litigation lawyers, restaurant waiters and bartenders, scout soldiers, drivers, astronauts, pharmacists and nurses/baby sitters.