Please define cognitive computers.

I also never heard about something like "cognitive computer" and it seems to me to be absolute unreal!

On the other hand I can imagine that quantum computing (not computer) will help us to solve some questions about our evolution and brain function.

In fact dealing with a great mass of data and data operations will produce new computer technologies but will never bring us one step forward understanding fundamental issues dealing with questions like cognition etc.

In what sense are you expecting quantum computers to compress information? A quantum state (e.g. on n qubits) cannot in principle store a greater amount of information than n bits; it can only store it differently. You can of course try to store multiple "pieces of information" in a quantum system simultaneously, but as soon as the total information exceeds n bits, it comes with a cost of having some probability of failure.

"Cognitive computers" would be a subset of quantum computers. All "cognitive computers" are quantum computers, but not all quantum computers are "cognitive computers". Cognition implies (and necessitates) quantum logic.

So how many quantum gates should be included in

an artificial neural network so that it deserves

to be called a cognizer?

Regards,

Joachim

I would understand 'cognitive computers' as 'human computers' = 'humans who are trained in computing and, as humans, have cognitive abilities'. The explanation (if this word is applicable here) going with the question seems not to point into this direction, though.

In my understanding, cognitive computer is defined from the functions of computers, quantum computer is defined from the strength of computer in computing capacity. If the computing capacity is stronger, the ability of computer in all cognition aspects could be stronger, as more complex or complicated computing for congnition can be implemented in real-time, hence this could enable the computer perception close to human perception.

I think that cognitive computing is more close to our way of thinking than the procedural one. Quantum computing deals with data mining process and the capacity and the speed of the computer to store and process huge data. I don't know if cognitive and quantum architectures are designed for this new generation of computers. in that case which is their Vlsi, circuits and sensor technology basis ?

As far as I know there is no way to store huge amounts of qbits yet.

And there is no hint that human long term memory stores

information in superposition state.

Regards,

Joachim

I think in the near future it may not be required to distinguish between the two. You may be familiar with the orch or model of Penrose and Hameroff, who both suggest that a quantum linear superposition within the microtubules of eucariotic nerve cells may arise for up to several milliseconds. I find the idea appealing, however, I would not be so tempting to reduce the human brain to a quantum computer, as I am of the opinion that self-organization has a role to play as well. Back to the point, take artificial neural networks into consideration, which habe been a major step forward in cognitive computing: current research suggests that processing such on a quantum computer would dramatically increase their capabilities.

Great answers and thoughts ! There is a perception that quantum computing an optimize / upscale the performance levels of certain computational tasks where as cognitive computers are for learning like human beings. In a quantum mechanical realm of universe, if there are apparent changes in the way cognition is supposed to work, what will happen to cognitive computers. How will they ( cognitive machines ) cope up with the complexities of entanglement and quantum fields.

Another question that came to my mind is : How will a cognitive computing machine behave in a quantum realm. Can a cognitive mind / function keep its stable state in quantum fields?

The cognitive computer is classical and cannot have a quantum speedup

Short coherence times needed for successfully completing complex computations such as a 'function of mind', as functionalist might say, may not necessarily be a problem. I mention this, because one of the major challenges of QC is to isolate a quantum physical system of its own irrelevant properties. When trying to implement an auto associative artificial neural network on quantum level, e.g. nuclear magnetic resonance could serve as implementation approach - coherence times of thousands of seconds have already been observed. Thus, the two worlds seem to be compatible. Also the already mentioned orch or-model of the human mind considers coherence times of that magnitude to be enough for experiencing conscious content (however, the collapse there occurs when two possible results of a computation, described by two sheets of space time, have moved away from each other by the order of a Planck length).

There are quite a few comments here already that touch upon what I consider to be the correct answer (and a couple that are not quite right).

According to IBM's website (which I assume you are familiar with :)), a cognitive computer is a natural learning type computer, a la Watson. These are (IMO) one of the most exciting developments of modern computer, and the style of processing makes non-experts like myself see the potential for an interpolation between conventional classical computing and perhaps even what could be understood as intelligence.

But it doesn't matter how powerful they become, or even if some threshold for 'consciousness' is achieved, they do not use the laws of quantum logic (ie all the laws of quantum mechanics beyond barrier tunneling and measurement) and hence cannot even be considered a quantum computer.

The simple test to distinguish a cognitive computer from a quantum computer are exactly the same tests one would use to distinguish a quantum from classical computer, or indeed a quantum computer from a human brain. We simply provide a task for which there is a quantum but no classical speed up (e.g. factorisation or boson sampling), modify the size of the problem and observe the scaling.

A few of the comments above touch upon the fact that decoherence is the issue here, and I would view this in the same way that I discount Penrose's arguments: fundamentally all physics in the brain is quantum, but the the effects required to utilise non-classical logic require long coherence times and delocalised quantum states that are usually not expected in the brain (here I distinguish processing from tasks that DO appear quantum, such as photosynthesis). Indeed it is even clearer with a cognitive computer reaches the same verifiable level of computing power as any biological brain (e.g. rat), then it appears fairly clear to me that the rat is doing classical logic, albeit with quantum bio hardware.

I'm not sure how you develop a Turing test for a rat, although I note that there is a non-zero number of Google hits for the topic - so perhaps there is more to this.

In cognitive computing, the algorithms are tuned to recognise patterns and continuously evolve for a better performance (what is known as artificial intelligence). So I feel, such computing always posses a memory of its past information and may be considered a Non Markovian process in output.

Whereas, in quantum computing information to be transferred is being stored inherently within the quantum system. The event output in this case may depend on previous events or may not. So essentially, the output may be Markovian as well as non Markovian depending on whether it is an information or some randomness. This creation of randomness perhaps can never be generated by the cognitive computer as the latter will always possess a history of events.

Please rectify me if I am wrong. I am just a novice in the field of quantum computing.

***** Watson Vs. the Electron *****

Hyper Speed Cloud based cognitive computing machines like Watson in the 2020's is likely the only way we WILL develop Quantum Computing without absolute zero chambers, and also Cognitive computing will solve the optimization issues / nonlinear dynamics involved with quantum computing, making sense of chaos / large scale ambiguous patterns... & probably solve the P versus NP problem... we will probably need something like a 'Super Watson' to continuously calculate and predict multivariable electron simultaneity... ~ "In Quantum Atom Theory it is the emission and absorption of light or Electromagnetic Radiation from one atom to another that forms the forward momentum of time."