‘Can we apply the linguistic rule to Boid for creating new poetry?’ >NLP? or ‘Nature’? ‘Language’?’Processing’?

Project Overview

This project visualizes the lyrics of "Close To You" by The Carpenters as an interactive system where words become autonomous agents moving on a canvas. Using natural language processing and boid algorithms, the project creates a dynamic representation of language structure where:

Words fly around like birds in a flock
Words can land on "wires" to form sentences
Words with grammatical relationships tend to move together

I feel the lyrics of ‘close to u’ is really sweet and fits to the context of ‘Boid algorithm’ Especially the rule of alignment. So I want to remix the lyrics using Boid by perceiving each words as the single agent, ‘What if the poetry composition processing is driven by physics?’I want to explore the intertextuality and materiality of language emergent

ref：https://medium.com/@eschaddel/the-materiality-of-language-what-does-it-mean-feebca6e06b

Why do birds suddenly appear
Every time you are near?
Just like me, they long to be
Close to you

Why do stars fall down from the sky
Every time you walk by?
Just like me, they long to be
Close to you

On the day that you were born the angels got together
And decided to create a dream come trueSo they sprinkled moon dust in your hair of goldAnd starlight in your eyes of blue
That is why all the girls in townFollow you all aroundJust like me, they long to beClose to you
On the day that you were born the angels got togetherAnd decided to create a dream come trueSo they sprinkled moon dust in your hair of goldAnd starlight in your eyes of blue

That is why all the girls in townFollow you all aroundJust like me, they long to beClose to you
Just like me, they long to be 
Close to you
Wa, close to you Wa, close to you
Ha, close to you La, close to you

Technical Foundation

1. Word Vectors and Semantic Space

The initial position of each word is determined by BERT word vectors reduced to 2D space. This means words with similar meanings appear closer together in the visualization - creating a "semantic landscape" where the words naturally organize into meaningful clusters.

2. Dependency Relations

https://spacy.io/api/dependencyparser#init

A key innovation is using dependency parsing to influence word behavior: