1V0

I created a new programming language and REPL environment, 1V0, running on a variety of platforms interactively. The idea is an "mainframe batch-like sort of computation", where you set up what you want beforehand and then get results. The only data type is the floating point (or, depending on the implementation, fixed point) number. The only input facility you need is something akin to an old cellphone numeric keypad. It runs presently on ATTiny85, Arduino UNO/MEGA2560/DUE/..., ESP8266 etc. You also get an implementation in C and one in Java, to experiment on your desktop. There are also two midlets (one more "modern", one more "primitive") that should enable any Java-MIDP-enabled device to run it. The aim was to have a tool that lets you use practically any sort of electronic, including old, nearly discareded phones, into "computers", in the style of the 1940s-1960s. It has been tested on Siemens M50 and SonyEricsson D750i, and of course, on all the Arduinos.

YouTube:

https://youtu.be/dZ4dWOTq1ac

GitHub:

https://github.com/KedalionDaimon/1V0

VI. Swarm and Evolution

Ladies and Gentlemen!

The topic of this video is multiplying useful agents and using them as a "swarm" as a scaling strategy.

http://www.youtube.com/watch?v=E6p-dtpYDHo



"Swarm" intelligence here focuses on the idea of using multiple copies of any agent, whereby I endeavour to keep it neutral as to "what kind of agent" it is and "what kind of approach" it uses. I am outlining ideas for swarm evolution. Aim is the increase of capacity. In the end, I present my view that every swarm may itself be seen as an "individual" for an even "higher" swarm.






This is the last video of the preparatory theoretical block. I am planning to show you from now on experiments fortnightly until March 2017. (And I really have to find out how this Github-thing works...) - Next shall come a few swarm intelligences operating with logical triangulation (Common Lisp, Fortran, and the "best" one is in Scheme.)







More information about my swarm approaches you can find here:

https://sites.google.com/site/logicaltriangulation/home/on-patterns-and-learning-through-interaction

Kindly let me know what you think.

V. Fragmentation Sets

The topic of this video is the recognition of similarity.

https://www.youtube.com/watch?v=Uqgge5GQvpQ

"Recognition of similarity" is a both necessary and sufficient condition for finding analogies, which I regard as central for intelligent reasoning. I propose an easy method to recognise similarities of arbitrary patters, which is (i) not a set comparison and (ii) not a vector comparison, but shall combine the advantages of both - i.e. both "flexibility", "precision" and "speed", for SYMBOLIC reasoning.



WHAT the patterns are is irrelevant (similar to most - slower - connectionist approaches); it is only required that they are patterns along some commensurable extension or dimensions. - This is going to be a central idea in a lot of the experiments which I will be showing soon.



The original theory according to which this video was created you can find here:

https://sites.google.com/site/logicaltriangulation/home/on-patterns-and-learning-through-interaction

- Nino Ivanov (Kedalion Daimon)

IV. Reasoning with Patterns and Interaction

Ladies and Gentlemen,

Here comes my fourth video of my series on general artifical intelligence:

https://www.youtube.com/watch?v=bDbU1fHivmc







I propose a method of reasoning with patterns (and introducing thereby a simplification compared to logical triangulation). I am proposing how a generally intelligent agent can learn proper interaction with the environment by observation of challenges and replies, and how context may be implemented as a "sliding window" of the "present".







The original theory according to which this video was created you can find here:

https://sites.google.com/site/logicaltriangulation/home/on-patterns-and-learning-through-interaction

The video mentioned in the video above concerning AI on Arduino is this one:

https://www.youtube.com/watch?v=okItWjaQ-N4





Kindly let me know what you think. More videos will come, later on at a rate of some one or two per month until, likely, March 2017. I understand that different levels of technicality are appropriate to different people, but as it is a whole series, I am intent on showing you all of it in order to avoid having to clarify "gaps" later on.

- Nino Ivanov (Kedalion Daimon)

III. Logical Triangulation

Dear Ladies and Gentlemen,

I would like to continue to share with you my new series on Artificial Intelligence.

I this third video I explain a new logic based on patterns, which I invented in order to analyse patterns purely based on their perception: and in particular, which patterns should be "given more weight" (and used more often). Evaluation is done on compositions of three relations, hence, "Logical Triangulation". It is a very general logic independent of the content of the patterns. This mechanism of evaluation may then be used for symbolic evaluation of data perceived through sensors, as I will show in the future.

While it may look innocuous, it is something really new, and with this, I wish to set the ground for presenting my implementations later on:

https://www.youtube.com/watch?v=IUFWNiKoDHA





(FYI, this:

https://www.youtube.com/watch?v=cW1BFJSZpd4

... was the previous and this ...

https://www.youtube.com/watch?v=-yVTMb4-B-A

... was the first video.)

The full-text description you shall find here:

https://sites.google.com/site/logicaltriangulation/

(Scroll down for the PDF for all the nitty-gritty details.)







Kindly let me know what you think. More videos will come until, likely, March 2017. I understand that different levels of technicality are appropriate to different people, but as it is a whole series, I am intent on showing you all of it in order to avoid having to clarify "gaps" later on.

- Nino Ivanov (Kedalion Daimon)

II. On the General Nature of Intelligence

Ladies and Gentlemen,

First of all: today's video:

https://www.youtube.com/watch?v=cW1BFJSZpd4





Welcome to the continuation of my series on artificial intelligence, after quite a hiatus to set up things and see what people say elsewhere. All is looking great. I will outline herein my theories and experiments, beginning with ideas and conjectures and continuing to various implementations.



The Logical Triangulation method I presented in my website under https://sites.google.com/site/logicaltriangulation/ was a first step. But there are many "dimwits with a theory", right? - Babblers, "communities"... only ACTUAL SYSTEMS are "a little" harder to find.



We shall, therefore, not "leave it at that". - It shall be my very good pleasure to share with you my theories and actual, practical experiments and implementations in the field of general artificial intelligence (which some seem to prefer to name "AGI"). In the course of this series, you shall see new stuff - MY stuff. I is open-source. Recognising patterns, chatting, and driving an Arduino robot will be part of it.



One thing I say ex ante: I do not give a damn about groveling before "so called authorities" (or "random babblers", if we are frank, and irrespective of their workplace), "which company attempts what" (which is either "hot air" or "money-making" - but will not TEACH you anything), etc. I am not going to sacrifice a goat in order to get a neural network going (though I shall show you how to make them learn hundreds of times faster, with regard to classical multi-layer perceptron style ones that learn with backpropagation). You shall see actual systems and experiments, and soon, a virtual machine with Linux and ready-to-run things as well. I have done my philosophical homework, and the result shall be shown to the world, for the benefit of everyone who is interested.



---

The original essay according to which this video was created you can find here:

https://sites.google.com/site/logicaltriangulation/home/philosophic/intelligence.pdf?attredirects=0&d=1

In this second video I explain my view that intelligence means the application of past experiences to new challenges and the fundamental role of analogy for achieving this. Reasoning by analogy hereby means treating DIFFERENT things THE SAME due to their COMMON properties. I present the conclusion that a general artificial intelligence thereby may well be intelligent, but not in the way humans are.

With this I wish to set the ground for presenting my implementation ideas in future videos. I believe that "recognising similarity" is SUFFICIENT as a process for general AI.



I shall publish new videos about twice a month, though the first few may come in more often - I have a backlog to handle. In the beginning, I will outline some philosophical basics and theoretical approaches, and later on I will show you practical implementations working on these principles. I understand that different levels of technicality are appropriate to different people, but as it is a whole series, I am intent on showing you all of it in order to avoid having to clarify "gaps" later on.

Kindly let me know what you think.


- Nino Ivanov (Kedalion Daimon)

EXPERIMENTS IN ARTIFICIAL INTELLIGENCE: I. ON THAT WHAT CANNOT BE KNOWN NOR THOUGHT

NEW SERIES ON

EXPERIMENTS IN

ARTIFICIAL INTELLIGENCE

Dear Ladies and Gentlemen,

It shall be my very good pleasure to share with you my theories and experiments in the field of artificial intelligence.

I shall publish new videos about twice a month within the course of a year. In the beginning, I will outline some philosophical basics and theoretical approaches, and later on I will show you practical implementations working on these principles. I understand that different levels of technicality are appropriate to different people, but as it is a whole series, I am intent on showing you all of it in order to avoid having to clarify "gaps" later on.







And here comes my first video of, as it seems, twenty-five (very much on the philosophical side):

https://www.youtube.com/watch?v=-yVTMb4-B-A





 

Kindly let me know what you think.

- Nino Ivanov (Kedalion Daimon)