[Denied]How does a Positronic Brain Work?

[Muncorn]

Type (e.g. Planet, Faction, System): Synthetic Lore

Founding/Settlement Date (if applicable): 23rd Century

Region of Space: Mars

Controlled by (if not a faction): Null

Other Snapshot information: General lore concerning Positronic Brain functions.

Long Description:

 

For a decent while there's been many theories, headcanons and semi-lore around how positronic brains actually work, but to date no full explanation of their function or structure, just half-truths and glossing over of any laws of physics that happen to be broken within any explanations that exist. I've aimed to fix most of those problems with my proposal here, something that I've had sitting around for a good while but never got round to publishing properly. So here it is, in all its (un)refined glory.

HISTORY

In the latter part of the 23rd century, massive developments in MMI devices were already being made, but some developers had started to have ideas of purely synthetic devices that could possibly achieve the same level of sentience. Over the course of a few decades, the first positronic device had been built. It was small, and had approximately 500 pathways within it. Even this allowed for the storing of close to 32000 bits of information, a number that to most seemed quite pitiful. Many scientists and roboticists were put off by the dismal result, but a group of aspiring students at the School of Robotic Science in Olympia, Mars, carried on the development and created the first positronic brain, a device similar to it's predecessor yet far larger (It was about 15 inches on all sides) and contained 12048 pathways and even had coded portions consisting of a basic arithmetic processor, memory storage and visual input. This led to the development of positronic devices used only as calculators, processors and only occasionally with any sort of remote sentience, and even then only as a novelty. It was only upon the advent and discovery of Skrell that the old AI designs were leaked to the human populous. Rather rapidly the code taken from Skrellian papers was transferred to a form useful in Positronics, and so more true AI's began to flourish.

The for them: Positronic Brains; was first coined in 2364 by Professor James Molikeev in honour of the sci-fi writer Isaac Asimov's fictional inventions of a similar nature which sparked the first investigations into the possible production of the brains in the first place. Many names were used before this, such as Guided Electron Processing Module and even such names as Electron Coding Cube was used. The name quickly caught on however, and in 2403 the first use of the abbreviated Posibrain was noted in a published paper (Investigations into Viability of Positronic Units as Storage for Financial Data by Prof. Victor Yoten).

COMPONENTS

Despite the peculiar name, the Positronic Brain actually makes no use of Positrons within it, for reasons obvious to those that know the basics of how anti-matter interacts with the regular matter in the universe. The brain instead consists of channels that electrons are fired down in certain orders much in the same way electricity would be sent through wires, but instead of simply sending information like a traditional wire would, these relays of electron transfers work in the same way as the neurons in our brains do. The Posibrain is immensely complex, and below you'll find a brief description of each piece labeled on our useful diagram.

Below is a general diagram of the NT standard positronic brain. Different models will differ in certain ways, such as varying control panels and different sizes of CPC's, but they all function in much the same way.

Central Power Core: The primary power source of the brain while it is not connected to a chassis, these can take many forms but are normally simple long lasting batteries. They are able to be charged through the connector port once attached to a chassis or other device.

Central Programming Core: The hub of work, it works as an intermediary between the positronic pathways and the output of the brain. It contains the necessary information needed to run and maintain the complex systems that the pathways use all the time, and is the most essential part of the brain. The main reason for Positronic Brains having stayed the same size and complexity for so long is due to this piece, as more positronic pathways mean exponentially more processing power is needed, which is enormously difficult to extrapolate a solution for.

Positronic Pathway Layers: This intricate and massive series of layers can consist of something between 10,000 and 100,000,000 different pathways, depending on the quality of the brain itself. These pathways are essentially the same as neurons in a brain, but process around 10 to 100 times faster. However as stated before, the number of pathways can't reach the same as the number of neurons in a brain, thus the same level of complexity and storage cannot be reached. However, all memory is perfectly locatable and never lost (Apart from through damage).

Protective Shell: A sturdy shell that wraps around the entire brain, it is always designed to be semi-transparent. This is due to reasons stated later on.

Maintenence Panel and Maintenence Port: These allow easy access to the central core whenever it is needed. The panel will pop off and the port will collapse in on itself to provide a small opening. While closed, the port is made of a lightweight, sturdy material (Normally a form of graphene) that allows easy heat escape.

Lighting Arrays: Connected directly to the central core, these are used to direct positions of damage or error. When a scan is run through the brain, if an error or point of damage is found the array will shine a light through the area. Different colours mean different problems;

Blue - Missing Code/Memory

Red - Physical Damage to Pathways

Brown - Physical Damage to Central Programming Core

Pink - Physical Damage to Central Power Core

Purple - Physical Damage to Protective Shell

Green - No Problems (Will shine through whole brain for ten seconds then fade)

Orange - Physical Damage to Unfoldable Pathways

Yellow - Physical Damage to Connector/Connector Port/Exterior Panel

LightBlue - EMP Damage to Core/Pathways

Connector: A simple connector between the core and the port, all information sent out is formatted and collected by the Core and sent through this. By default, all outputted information will be in binary, but outputs can be changed in many ways.

Connector Port: A simple connector port, what style will change depending on manufacturer and intended use of the brain.

Exterior Panel: A simple panel consisting of a few buttons. The ones found on most NanoTrasen models are listed below, but other models might have differing setups.

1: Reboot Brain - Shuts off power and clears current processing data. Upon reboot, will check for any errors caused by the reboot. Do not press if important actions are being processed.

2: Scan Pathways and Core for Physical Damage - The lighting arrays come into play here. The brain will scan the brain for physical damage of any component excluding the protective shell. This is due to NanoTrasen Units not having an inbuilt function to scan for exterior damage (Shell, Panel or Connector Port) though the intelligence can perform this scan itself, with appropriate permission.

3: Scan Pathways for Network Damage - Will scan the pathways for any errors and then show appropriate error display. Always do a physical check before performing this check, as results may differ otherwise.

4: Open Maintenence Panel- Unclamps the panel and unfolds the covering sheets.

5: Scan for Personalities - In all NanoTrasen units (Those designed and manufactured by NanoTrasen for stationside use) there will be a basic connection to the extranet for the purpose of downloading synthetic personalities. Before a personality is in place, the brain is devoid of processes barring those within the core. All pathways are laid bare and unresponsive. When a personality is found, the pathways are instantly coded and the core begins bootup.

6: Display Power Level - Will display the current power level with the help of the lighting arrays. Beginning on the left of the connector port, the lights will shine in a circle around the brain. The proportion it makes around is the percentage charge. However, an easier way is to ask the personality.

[CakeIsOssim]

If the term "Positronic Brain" was coined in 2364, but AI was not in use until 2437 (seventy-three years between the two), why was it acknowledged as a "Brain" if it was really no different than a specialized computer? We don't really call our CPUs "brains" colloquially or otherwise.

[Muncorn]

As that's the only thing you've brought up, that bodes well overall. As for the name, let me make some things clear, mainly for myself.

By AI, I'm presuming you mean AI in the level of complexity that Skrellian technology allowed us to have. Presuming that, before that point there will have still been rudimentary AI as a lack of it would make very little sense as we already have it in present day to a certain extent. The most common usage of this AI will have been fairly trivial, perhaps to help answer basic questions or as a cheap alternative to an MMI or Cyborg, as those have existed since 2200 or so. The Positronic Brain will have come about as a replacement for microchips and not in particularly as a sole facilitator of AI. Due to its function being both similar and inspired by Asimov's Positronic Brain, and it having a fairly large, cuboid shape which very vaguely resembles a brain, the term was coined and used even if it didn't do anything like you'd expect a brain to do. The eventual relevance of it being called a brain would likely have been quite unknown to Professor James Molikeev at the time, much as Asimov would never have expected his dreams to become reality.

I hope that covers it.

[DrHobo]

Right I absolutely love this application (it's beautiful and sets a good example for future applications) and it really got me thinking but here's the but, actually positron-emission tomography is an example of how matter-antimatter annihilation isn't all that dangerous or bad, and can be incredibly useful! I don't think you should all out abandon the idea of actual positrons being used in it, as when it's powered (with electrons) it just produces a couple gamma rays which could then interfere with other gamma rays, which interact with others, and others (possibly making some kind of ultracomplex Feynman diagram of wave-particle interactions as the basis for computation).

Instead of electron pathways, I'd think include a positron-emitting radionuclide of some kind (maybe vary on personality complexity: Radium, Iridium, Radium-Iridium, Iridium-Phoron?) in the CPU which works kinda like a crystal processor clock and have the gamma-rays directed like neural pathways in the positronic pathway layer (would call it a matrix or matrix layer as well cause computational matrices and it sounds cool). The protective casing would be some lead-tungsten superalloy or the like to prevent it leaking and any heat build up. "Guided Electron Processing Module" and "Electron Coding Cube" could be "Guided Gamma Processing Module" and "Gamma Coding Cube". Electrons are so passé anyway

As for Cake's response which is a good question, Muncorn is right that AI before would have been trivial compared to then but also as it uses pathways to replicate computational processes like a brain doesn't necessarily make it intelligent much like the petri dish neural networks we have now, they in no way produce a conscious entity on their own as is. It would still take a fair while to learn how to properly utilize simpler functions and the like before more increasingly intelligent algorithms would be developed, and 73 years seems like a good amount of time.

tl;dr +1 but please keep positronic brains positronic!

[Muncorn]

Like, I'm not fully sure of what half the stuff you just said means, and that is much of my point in making this. I didn't want to go overcomplicate this, as I want the largest numver of people to understand this and thus use it ingame. I'm under the impression that making things overly complex alienates those that wish to involve themselves in the lore without having to spend hours researching practical or theoretical physics.

[DrHobo]

Ok, I'm sorry I can be really, really bad for overcomplicating stuff but fear not I'll try again! Forget about alloys and all that or Feynman diagrams, just the principle of positronic computing.

When it says about matter-antimatter not being used for obvious reasons, that's not actually true cause PET brain scanners use the principle on a daily basis in hospitals all over the world and even bananas emit tiny, tiny amounts of positrons. So they could work with a "positron-emitting radionuclide" (I didn't choose the name, it's just science) which would work like the crystal in a modern computer processor except it produces gamma rays from the electrons coming from the power cell, so that could be a little thing in the CPU, a PER crystal.

Those gamma rays could be used where you say it uses of channels of electrons like electricity and form those neural pathways probably a lot more efficiently than electrons (in a sci-fi future defo, but could actually be feasible one day), it would just be channels of gamma rays like electricity. That's what the protective casing could stop from escaping and protect IPCs from radstorms.

The only change I'm suggesting is the start of Components to change electrons to gamma-rays from a PER crystal, and the History with "Gamma Cubes" etc instead of the electron equivalent. 500 years of electronics is far too long and positronics is good. It's good science! Don't fear the positron! Also there's nothing wrong with giving folks a real science taste to go and spend a few hours learning practical and theoretical physics, I'm sure there are plenty folk that play do.

[CakeIsOssim]

Kyres and I have discussed this a bit, and we've broken it down to three major criticisms. Here they are:

1. Accurately describes the processing power and function of a positronic brain.

This is something we actually want to avoid doing. A marginal amount of people who play or play around synthetics are going to be either computer scientists or theoretical physicists, so we can allow ourselves some creative freedom here. This freedom, when expressed correctly, allows people to come up with their own explanations for filling the gaps. Not really 'gaps' per se, but something that is more meaningful to a reader or player other than the exact scientific function of the brain. It can be reduced to a few simple concepts, I'm sure.

I can see you have already realized the importance of not alienating people. That's good, but there's another goal; how much of this will be relevant in-game? How often will any of it actually be brought up in roleplay? If I may be blunt, how much of this will people give a damn about, other than that marginal percentage?

2. "Has anyone actually seen your degree, Doctor Freeman?"

"The brain instead consists of channels that electrons are fired down in certain orders much in the same way electricity would be sent through wires, but instead of simply sending information like a traditional wire would, these relays of electron transfers work in the same way as the neurons in our brains do."

You got half of what I covered in number one here. Half scientific technobabble, half reality-based science. I'd like to lean more towards "technobabble," but with less babble. Let me clarify here that there is nothing really wrong with describing the function of a positronic brain. The trick is to just not reduce it to an exact science. There's another thing that ties into this one but I'll cover that in number three.

3. The specifications of a positronic brain are not really necessary.

Just about every megacorporation builds or owns synthetics. They all have varying models, different makeups, and whatever else. Positronic brains aren't really that different, so we'd like to ebb away from over-specifying parts of the brain. While I personally originally liked the diagram and outlining of the different parts, I'm now against it for the sake of brevity. Brevity is the soul of wit, and this is what I was alluding to at the end of number two. It's also somewhat trivial.

While I do really appreciate your effort, I do not think it's likely for this to be accepted in the current form that it takes. We want to open discussion about this, however. What are your thoughts on these criticisms?

[Coalf]

No activity since Feb 16th, locking and archiving.