Compositional Semantics for Open Vocabulary Spatio-semantic Representations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Karlsson, Robin, Lepe-Salazar, Francisco, Takeda, Kazuya
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910248961310720
author Karlsson, Robin
Lepe-Salazar, Francisco
Takeda, Kazuya
author_facet Karlsson, Robin
Lepe-Salazar, Francisco
Takeda, Kazuya
contents Vision-language models (VLMs) transform environment percepts into vision-language semantics interpretable by LLMs. However, completing complex tasks often requires reasoning about information beyond what is currently perceived. We propose latent compositional semantic embeddings z* as a principled learning-based knowledge representation for queryable spatio-semantic memories. We mathematically prove that z* can always be found, and that the optimal z* is the centroid for any set Z. We derive a probabilistic bound for estimating separability of related and unrelated semantics. We prove that z* is discoverable from visual appearance and singular descriptions by iterative gradient descent. We experimentally verify our findings on four embedding spaces including CLIP and SBERT. Our results show that z* can represent up to 10 semantics encoded by SBERT, and up to 100 semantics for ideal uniformly distributed high-dimensional embeddings. We introduce three new datasets with overlapping semantics to show that common VLMs trained on conventional nonoverlapping annotations discover z*. Our novel sufficient similarity inference method overcomes fundamental limitations of conventional inference, and improves higher-level overlapping semantic inference performance by 19.63 mIoU on average.
format Preprint
id arxiv_https___arxiv_org_abs_2310_04981
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Compositional Semantics for Open Vocabulary Spatio-semantic Representations
Karlsson, Robin
Lepe-Salazar, Francisco
Takeda, Kazuya
Computer Vision and Pattern Recognition
Machine Learning
I.2.10; I.2.9
Vision-language models (VLMs) transform environment percepts into vision-language semantics interpretable by LLMs. However, completing complex tasks often requires reasoning about information beyond what is currently perceived. We propose latent compositional semantic embeddings z* as a principled learning-based knowledge representation for queryable spatio-semantic memories. We mathematically prove that z* can always be found, and that the optimal z* is the centroid for any set Z. We derive a probabilistic bound for estimating separability of related and unrelated semantics. We prove that z* is discoverable from visual appearance and singular descriptions by iterative gradient descent. We experimentally verify our findings on four embedding spaces including CLIP and SBERT. Our results show that z* can represent up to 10 semantics encoded by SBERT, and up to 100 semantics for ideal uniformly distributed high-dimensional embeddings. We introduce three new datasets with overlapping semantics to show that common VLMs trained on conventional nonoverlapping annotations discover z*. Our novel sufficient similarity inference method overcomes fundamental limitations of conventional inference, and improves higher-level overlapping semantic inference performance by 19.63 mIoU on average.
title Compositional Semantics for Open Vocabulary Spatio-semantic Representations
topic Computer Vision and Pattern Recognition
Machine Learning
I.2.10; I.2.9
url https://arxiv.org/abs/2310.04981