Compression Algorithm to Quantum Computers

[BetaTester]

Quantum computers are Boltzmann machines, operating at Monte Carlo method in hardware level.

http://en.wikipedia.org/wiki/Monte_Carlo_method

http://en.wikipedia.org/wiki/Boltzmann_machine

A compression algorithm composed of neural networks, with each cell performing statistical calculations, it would have an infinite speed, and would be using the full potential of the quantum computer - Combining PPM + CTW elements.

All quantum processors so far, using a few qubits, and the rest is conventional computing.

The consistency of data would be maintained by conventional part.

[Matt Mahoney]

The problem is Boltzmann machines make nondeterministic predictions. In a conventional computer you can use a PRNG to get repeatable predictions (to decompress), but not in a quantum computer.

[vylbird]

The nondeterministic nature of quantum computers does indeed make predictive compression useless with them. But they do have their potential applications. I'm not really knowledgeable in the field, but if perfected quantum computers would have excellent search performance. That means that conventional compression could be dramatically improved by using dictionary sizes that would cripple a modern computer - just think of LZMA2 on a dictionary size of many gigabytes, or deduplication in real time on massive filesystems. I don't know if this could be extended to rapid searching for approximate matches, but if it can then there are potential applications in lossy compression for finding motion vectors and inter-frame redundancy in video.