December 12, 2020  /  blog

What is AI?

Many definitions exist on Artificial Intelligence, Machine Learning and Algorithms. Hereby a brief overview of a couple of them.  Definition of AI  What does Wikipedia say?  Artificial intelligence (AI), sometimes called machine intelligence, is intelligence demonstrated by machines, unlike the natural intelligence displayed by humans and animals. Leading AI textbooks define the field as the study of “intelligent agents“: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] Colloquially, the term “artificial intelligence” is often used to describe machines (or computers) that mimic “cognitive” functions that humans associate with the human mind, such as “learning” and “problem solving”.  Well to do some proper accounting we definitely need to mimic cognitive functions that humans   Types of AI  Typically 4 types 1of AI are being defined: 
  1. REACTIVE MACHINE 
  2. LIMITED MEMORY 
  3. THEORY OF MIND 
  4. SELF-AWARENESS 
Below a brief description of the four types.  1. Reactive Machines:  The most basic types of AI systems are purely reactive, and have the ability neither to form memories nor to use past experiences to inform current decisions.  Example: Deep Blue from IBM beating chess world-champions  This type of intelligence involves the computer perceiving the world directly and acting on what it sees. It doesn’t rely on an internal concept of the world.   2. Limited Memory  This Type II class contains machines can look into the past. Self-driving cars do some of this already. For example, they observe other cars’ speed and direction. That can’t be done in a just one moment, but rather requires identifying specific objects and monitoring them over time.  These machines only use simple pieces of information about the past and these are only transient. They aren’t saved as part of the car’s library of experience it can learn from, the way human drivers compile experience over years behind the wheel.  3. THEORY OF MIND  Machines in the next, more advanced, class not only form representations about the world, but also about other agents or entities in the world. In psychology, this is called “theory of mind” – the understanding that people, creatures and objects in the world can have thoughts and emotions that affect their own behavior.  4. SELF-AWARENESS  The final step of AI development is to build systems that can form representations about themselves. Ultimately, we AI researchers will have to not only understand consciousness, but build machines that have it.  Types of Algorithms: 
  • Recursive Algorithm 
  • Divide and Conquer Algorithm 
  • Dynamic Programming Algorithm 
  • Greedy Algorithm 
  • Brute Force Algorithm 
  • Backtracking Algorithm 
  Definition of an Algorithm?  According to Wikipedia:  In mathematics and computer science, an algorithm is a finite sequence of well-defined, computer-implementable instructions, typically to solve a class of problems or to perform a computation. Algorithms are always unambiguous and are used as specifications for performing calculations, data processing, automated reasoning, and other tasks.  Types of Algorithms  Tens of types of algorithms can be found in the literature. The six most recognized2 ones are:  1. Recursive Algorithm  This is one of the most interesting Algorithms as it calls itself with a smaller value as inputs which it gets after solving for the current inputs. In more simpler words, It’s an Algorithm that calls itself repeatedly until the problem is solved.  This is not required for Accounting  2. Divide and Conquer Algorithm  This is another effective way of solving many problems. In Divide and Conquer algorithms, divide the algorithm into two parts, the first parts divides the problem on hand into smaller subproblems of the same type. Then on the second part, these smaller problems are solved and then added together (combined) to produce the final solution of the problem.  We don’t use this type of algorithm either  3. Dynamic Programming Algorithm  These algorithms work by remembering the results of the past run and using them to find new results. In other words, dynamic programming algorithm solves complex problems by breaking it into multiple simple subproblems and then it solves each of them once and then stores them for future use.  Fibonacci sequence is a good example for Dynamic Programming algorithms.  4. Greedy Algorithm  These algorithms are used for solving optimization problems. In this algorithm, we find a locally optimum solution (without any regard for any consequence in future) and hope to find the optimal solution at the global level.  The method does not guarantee that we will be able to find an optimal solution.  The algorithm has 5 components: 
  1. The first one is a candidate set from which we try to find a solution.
  2. A selection function which helps choose the best possible candidate.
  3. A feasibility function which helps in deciding if the candidate can be used to find a solution. 
  4. An objective function which assigns value to a possible solution or to a partial solution 
  5. Solution function that tells when we have found a solution to the problem. 
 5. Brute Force Algorithm  This is one of the simplest algorithms in the concept. A brute force algorithm blindly iterates all possible solutions to search one or more than one solution that may solve a function. Think of brute force as using all possible combinations of numbers to open a safe.  6. Backtracking Algorithm  Backtracking is a technique to find a solution to a problem in an incremental approach. It solves problems recursively and tries to get to a solution to a problem by solving one piece of the problem at a time. If one of the solutions fail, we remove it and backtrack to find another solution.  In other words, a backtracking algorithm solves a subproblem and if it fails to solve the problem, it undoes the last step and starts again to find the solution to the problem. 
Share