Formal Security Thread - 0

“Machine intelligence is the last invention that humanity will ever need to make, because machines will be better at inventing than we are”

  • Nick bostrom

This thread of mine is to explain my ideas on how we can formalize our understanding and approach to the security and privacy in the field of artificial intelligance. I have been working in this field since I was an second year undergraduate student and thoruhgout this journey in every paper I have read I felt the underlying cat and mouse game going on with no real basis unlike my basic readings in cryptography, mathematics and even electronics engineering. With the results we get from formal cryptography literature is sure that no means of efficeint computation will ever achieve to break the proven secure crypto systems we have -altough specific to crypto we assume existance of PRFs or NP \neq P in each case I know until now- or in mathematics when Gödel has proven that no axiomatic system will ever achieve completeness in the sense Hilbert conjectured no one has continued to defend that idea as it first stated -but again I need to admit for this specific case people have restructured the mathematical completeness in more sophisticated ways- and in electronics engineering and specifically signal processing when Hamming has introduced his compression algorithm and proved that there will never be a more efficient algorithm for an arbitrary set of source data for a lossless compression -altough the proven algorithm is not the one we use in production right now-.

These are the types of works I admire and set myself as a goal humblely or at least to contribute for such a result in a field because this is what remains to collective understanding of humanity and contributing to collective humanity is valuable for me because it is the one and only physically realized mean of intelligance experienced by us humans so it is the only way an idea has been indeed intellectually valuable. No one remembers the name of the king of syracuse but today we have results from Archemides, tough I want to make the point that it is not important to be remembered at the end of the day both are flesh now ebing remembered or not is equally affecting both of them but one has a meaningful contirbution to a valuable intellectual result respect to this discussion I have.

Along with this admiration I left a gap for all the works listed above. They are all either rely on axioms we have never realized even further for some cases we have proven that our axiomatic systems are insufficient to prove such systems, altough proven to be worng they are being restated or not practically utilized altough having a really strong theoretical result. Because one such result the one I will try to descibe in this thread will most probably have such defects and it is totally fine and “natural”.

With these discussion I gave above I am afraid I was far from approaching to such results and all along I was disturbed with being distant to such an approach in the field I have been a part to for about 3 years -again not too long but its about three quarters of the total span of my studies-.