Error correction is at the core of all cognition: synthetic and biological.
But error correction cannot be learned without failure.
But error correction cannot be learned without failure.
The difference between computers and biological cognition is that error correction is built into computers while it is learned through experience in biology.
The deepest enigma about biology (and life) is how did it discover the digital mechanism of DNA to reduce errors in replication?
To devise an error correction algorithm, the critical first step is recognizing what an error is.
But does biology recognize error or does it only recognized the difference between self and non-self?
Extinction is the error that biology attempts to correct. The two mechanisms it has discovered is replication and homeostasis.
Both mechanisms, in an abstract sense, preserve the self across time. Biology response to entropy is the generation of objects that are momentarily invariant in time.
This 'momentarily invariant in time' property is a relative property. Something appears invariant if it is relatively slower to change than everything else around it. Invariance is achieved by fast processes generating slow processes.
But what makes a process fast and what makes it slow?
The answer here is that it relates to time reversibility. The dynamics of subatomic particles are the same regardless of time flowing forward or backward.
Emergent properties and thus complexity is a consequence of micro-processes that generate macro-processes that are less reversible.
The macro-process will tend toward entropy without the micro-processes. That is why useful energy is need by the micro-processes to sustain a complex system.
But what 'motivates' a more complex system to maintain itself? Why does hydrogen coalesce to form stars?
Emergent properties arise as a consequence of fast micro-processes that *interact* with each other. Gravity, when formulated as an entropic force, explains why a star tends to be formed.
Emergent phenomena at large scales is a consequence of the interaction of its smaller parts. There is no emergence if there is no interaction of parts.
For biological beings, the parts (i.e. cells) are motivated towards survival. A multicellular being is constructed with the motivation of collective survival. Entire ecosystems band together to ensure their collective survival.
Error correction emerges at multiple scales. The error correction algorithms required of cells are different from error correction required for brains. But in all cases, this error correction algorithm is learned from its respective scale.
It is useful then to compare the designs for error-correction in human designs and that of evolutionary design. We shall see the difference in these design explain the greater adaptability of biological design.
There are two classes of error correction methods. Pessimistic and optimistic. The first kind is preventive and eager. The second kind is compensatory and lazy. Which kind do you think biology designs for?
@threadreaderapp unroll
