oswaldoludwig/visually-informed-embedding-of-word-VIEW-

Visually informed embedding of word (VIEW) is a tool for transferring multimodal background knowledge to NLP algorithms.

40
/ 100
Emerging

This project helps natural language processing (NLP) researchers and practitioners improve how their algorithms understand spatial relationships described in text. It takes textual descriptions of visual scenes (like image captions) and produces specialized word embeddings that capture visual and spatial context. These embeddings can then be combined with standard word embeddings to enhance algorithms that need to interpret where objects are in relation to each other.

No commits in the last 6 months.

Use this if you are developing or evaluating NLP models, especially for tasks like Spatial Role Labeling, and need to improve their comprehension of spatial language.

Not ideal if your NLP task does not involve understanding spatial relationships between objects, or if you prefer to use pre-trained models without custom training and embedding generation.

natural-language-processing spatial-reasoning text-understanding computational-linguistics multimodal-AI
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 17 / 25

How are scores calculated?

Stars

29

Forks

11

Language

Python

License

BSD-2-Clause

Last pushed

Sep 18, 2016

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/oswaldoludwig/visually-informed-embedding-of-word-VIEW-"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.