λ-2D

  • 4 Replies
  • 433 Views
*

ivan.moony

  • Trusty Member
  • ************
  • Bishop
  • *
  • 1649
    • contrast-zone
λ-2D
« on: April 05, 2022, 06:00:36 pm »
λ-2D: An Exploration of Drawing as Programming Language, Featuring Ideas from Lambda Calculus - link

« Last Edit: April 05, 2022, 08:00:41 pm by ivan.moony »

*

frankinstien

  • Replicant
  • ********
  • 618
    • Knowledgeable Machines
Re: λ-2D
« Reply #1 on: April 06, 2022, 01:56:55 am »
Interesting approach, but it's a little difficult to tell what functions are doing. There are other visual programming languages and the shapes are easily identifiable as well as colorized. In fact, many robots use visual languages such as Cozmo.  Also trying to symbolize programming logic and architecture has been a challenge and even today using UML isn't as popular as it should be.

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4639
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: λ-2D
« Reply #2 on: April 06, 2022, 05:35:43 am »
I'd like to see a more intuitive visual, such as:

Say the algorithm is a compression algorithm, that builds a tree of seen sentences (of 0-14 letters in length, all them it came across). And updates its nodes how many times seen each ex. 44 times for 'we walked', 2749 times for 'walk', etc. It then goes over a file and compresses it by predicting the likely next letter but for its error it stores a little correction to its comp pad.

So, the visual would be a big page of writing - the text file........and a little bar sliding across it sending off branch updates to a tree that counts times seen each piece......Then, when done, the other text file has a bar go across it and the bar snippet snapshotted is sent into the tree and it drops off predictions like a falling leaf to the ground, and these leaves then meet with a number thing to get the correction number, of which is added to a long pad being the compression text pad page, another 3rd page of text.

How to do it? Well you have the 3 files as input, the text to train on, the text to compress, and the output pad page. And you are selecting parts of the file, which can be visualized I think. Where these go is into a tree in this example, but how do we get a tree from a list system made up in the human's brain is a "tree" but not visually a tree? It's just a list, but it operates as a tree!! Maybe we can make something to make a visual from it! So say the first item in the list is always came across, all 256 of them, a, b, c, and so on, and so these are put into a row...the first layer. Then, the template algorithm sees ok so every time the item a is seen it sees then items in the list only at these next 256 possible positions, no where else at all! So, it associates them....this would look like a tree, each node connecting to its children accesses, the first letter of a text snippet does not access the (at least in this example) the 3rd letter ex. 'wa[l]k', so it does not jump to that item in the long list.

But what about the IF THEN? And AND? Etc? I think if we can show the main files, and structures that aren't there but are in the human's brain, we can already visualize LOTS of what's going on.

Just an idea, but, it may work!
Emergent          https://openai.com/blog/

*

MagnusWootton

  • Replicant
  • ********
  • 535
Re: λ-2D
« Reply #3 on: April 06, 2022, 09:51:24 am »
This looks like you actually draw out "hardware" for the software,  its good if you want to make a computer from scratch one day.
Doing software as a circuit works out fine,   u can put anything into one.   I do something similar,  well worth learning!!!

*

Zero

  • Eve
  • ***********
  • 1287
Re: λ-2D
« Reply #4 on: April 07, 2022, 09:28:34 am »
It made me discover Asciidots, another esolang I didn't know. These are very cool experiments.
 O0
Computer vision applied to them should allow strong memoization.

 


Users Online

45 Guests, 0 Users

Most Online Today: 68. Most Online Ever: 2369 (November 21, 2020, 04:08:13 pm)

Articles