Computer Science

AI and Liberalism

I stumpled upon a German languge article in the NZZ by Slavoj Žižek that deals with the topics of liberalism, humanism and digitalization. It is declared as a translation but I didn't find the source, so I can only provide you with the German text. While searching for the translation I found a greater number of text by the slowenian philosopher dealing with similar topics, so I guess this is more of a general reading recommendation. 

Eben weil die Maschine, die uns liest, als mechanischer Algorithmus blind und bewusstseinslos ist, kann sie Entscheidungen treffen, die nicht nur der äusseren Wirklichkeit angemessener sind als unsere eigenen Entscheidungen. Sie sind es vor allem auch in Bezug auf unsere eigenen Wünsche und Bedürfnisse. Die Maschine kann alle Widersprüche eruieren, Inkohärenzen messen und mit ihnen auf weitaus rationalere Weise umgehen, als unser fiktives Selbst dies vermag.

NZZ: Digitalisierung und künstliche Intelligenz: Das Ende der Menschlichkeit

Image: Zizek in Liverpool By Original photographer: Andy Miah , cropped by User:Michalis Famelis

A new kind of knowledge

In an article on Google Translate I mentioned that there is evidence, that the Google Translate artifical intelligence developed an "Interlingua" to translate between languages, i.e. a language not spoken by anybody that serves as a bridge between different langauges - by saving natural language information in a way never done before and impossible to be defined or understood my humans. This is a pretty good example of the new kind of knowledge that appears already and will appear more often on all areas, artifical intelligence is applied on. This is due to the approach, modern AI is using to understand and elaborate on topics: Neural networks. Another popular case was the Go game first one by an AI: It was impossible for human players to predict or even understand in retrospective the "winning move". It was made by an alien intelligence not working similar to our own. The article "Our Machines Have Knowledge We'll Never Understand" by David Weinberger is an inspiring musing on this topic.

"We are increasingly relying on machines that derive conclusions from models that they themselves have created, models that are often beyond human comprehension, models that “think” about the world differently than we do.

But this comes with a price. This infusion of alien intelligence is bringing into question the assumptions embedded in our long Western tradition. We thought knowledge was about finding the order hidden in the chaos. We thought it was about simplifying the world. It looks like we were wrong. Knowing the world may require giving up on understanding it." - David Weinberger

Neurosciencing a microprocessor

Now this is a topic that is rather fascinating to me: The scientists Eric Jonas and Konrad Paul Kording are applying the methods of neuroscience to a simulated microprocessor (a MOS-6502-Chip similar to the one used in the Commodore C64) in order to describe the observed behavior. Is it possible to get an idea of functionality? Apperantly, in 2002 there was a former study "Can a biologist fix a radio" (by Juri Lazebnik) with a very similar topic.

Here we will examine three different “behaviors”, that is, three different games: Donkey Kong (1981), Space Invaders (1978), and Pitfall (1981). Obviously these “behaviors” are qualitatively different from those of animals and may seem more complicated. However, even the simple behaviors that are studied in neuroscience still involve a plethora of components, typically including the allocation of attention, cognitive processing, and multiple modalities of inputs and outputs. As such, the breadth of ongoing computation in the processor may actually be simpler than those in the brain.

As it turns out in both cases: No, the methods of neuroscience/biology are not sufficient to understand or describe the behavior of the respective system. Does this mean anything? Yes and no. They are not designed to understand technology. Vice versa, an expert in reverse engineering probably would not understand a specified lifeform by the application of his methods aswell. But on the other hand the study reveals that we do not know for sure if the methods and the results they generate are useful for the purpose of understanding e.g. the brain. Do we have language centers in the brain or is this comparable to the misconception of space invaders centers in the micro processor?

Eric Jonas, Konrad Paul Kording: Could a Neuroscientist Understand a Microprocessor?

Golem: Könnten Hirnforscher einen C64 verstehen?


Image: Image of the circuit board of a Commodore 64 showing some important MOS Technology circuits: the 6510 CPU (long chip, lower left) and the 6581 SID (right). The production week/year (WWYY) of each chip is given below its name. Found on Wikipedia by Jef-Infojef

Computer Science Fun with Excel

There are a lot of pretty cool things one can do with Excel. Today I stumpled upon two of them at Hacker News:

First, Felienne has implemented a Turing Machine in Excel which basically means that one can process "every possible operation" in Excel, too.

The second one is a ENIGMA machine implemented in Ecxel as you may see in this video:

You can download the sheets on the respective pages.

IBM: Big Data, Speech Processing and Machine Translation

For a machine to truly process speech data, it needs cognitive computing – a system with architecture that imitates how the human brain understands information. IBM Watson’s ability to understand natural language is just a first piece to a complex cognitive computing puzzle. But as cognitive computing is applied to Big Data, it will also revolutionize speech recognition and speech translation.

IBM Research: Dimitri Kanevsky Translating Big Data