1-D encoding of 2-D spatial data for generative movement of tiger in a language model
Topics:
Keywords: Generative, movement, foundation, model
Abstract Type: Paper Abstract
Authors:
Sean C Ahearn, Hunter College - CUNY
David Smith, University of Minnesota - TC
,
,
,
,
,
,
,
,
Abstract
The paper will describe the implementation of a 1-d encoding of 2-D space using a linear quad tree, for generative movement with a Language Model. The data source for transformation and training is 4872 GPS locations of a tiger monitored at 1 hour intervals for a period of 8 months. The results are compared at different spatial scales with the original GPS data and a null model, in the form of a correlated random walk (CRW). Comparisons are made by analysis of the visitation rates of the tiger to its home range boundary. This work represents a new approach to using spatial temporal data at multiple scales, as input to Language Models (e.g. Bi-Gram, Transformers).
1-D encoding of 2-D spatial data for generative movement of tiger in a language model
Category
Paper Abstract
Description
Submitted by:
Sean Ahearn Hunter College - City University
sahearn@hunter.cuny.edu
This abstract is part of a session. Click here to view the session.
| Slides