Abstract:The aim of this paper is to propose an alternative behavioural definition of computation (and of a computer) based simply on whether a system is capable of reacting to the environment-the input-as reflected in a measure of programmability. This definition is intended to have relevance beyond the realm of digital computers, particularly vis-a-vis natural systems. This will be done by using an extension of a phase transition coefficient previously defined in an attempt to characterise the dynamical behaviour of cellular automata and other systems. The transition coefficient measures the sensitivity of a system to external stimuli, and will be used to define the susceptibility of a system to being (efficiently) programmed.
What problem does this paper attempt to address?
The core problem that this paper attempts to solve is to provide a new, behavior - based definition for computation (computation), which is applicable not only to traditional digital computers but also to natural systems. Specifically, the author Hector Zenil proposed a concept centered on "programmability", defining computation by measuring the system's ability to respond to external stimuli.
### Main problems of the paper
1. **Definition of computation**:
- Although there have been multiple attempts, so far there is no widely - accepted definition of computation. Existing definitions often rely on the specific specifications or state mappings of the system, which makes it difficult to accurately describe natural computation.
2. **Difference between natural computation and traditional computation**:
- Traditional computation usually refers to the computational process in digital computers, while natural computation involves the information - processing process in natural systems such as physics and biology. How to unify the two is a challenge.
3. **Introduction of programmability**:
- The author proposed a computation definition based on system behavior, especially the "programmability" of the system. Here, programmability is not limited to programming in the traditional sense, but refers to whether the system can respond to external stimuli and change its behavior according to these stimuli.
### Specific methods
To achieve this goal, the author introduced an extended phase - transition coefficient (phase transition coefficient), which was originally used to characterize the dynamic behavior of cellular automata and other systems. This coefficient can measure the sensitivity of the system to external stimuli, thereby defining the programmability of the system.
### Main contributions
1. **Behavior - defined computation**:
- Proposed a computation definition based on system behavior rather than internal structure, which makes this definition applicable to a wider range of systems, including natural systems.
2. **Quantification of programmability**:
- Introduced the phase - transition coefficient as a tool to measure the programmability of the system, providing a method for quantitative analysis.
3. **Cross - domain applicability**:
- This definition is applicable not only to traditional computer science, but may also have an impact on the concept of computation in fields such as physics and biology.
### Formula presentation
The key formula mentioned in the paper is the algorithmic complexity (Kolmogorov - Chaitin complexity), which is defined as follows:
\[ K_T(s)=\min\{|p|, T(p) = s\} \]
where:
- \( K_T(s) \) represents the shortest program length for generating the string \( s \).
- \( p \) is a program.
- \( T \) is a universal Turing machine.
This formula is used to measure the complexity of the system, and then help evaluate the programmability of the system.
In this way, the author hopes to provide a more general and flexible definition of computation that can cover various computational forms from traditional computers to natural systems.