Salvato in:
Dettagli Bibliografici
Autori principali: Hosseini, Hesam, Mighan, Ghazal Hosseini, Afzali, Amirabbas, Amini, Sajjad, Houmansadr, Amir
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2411.12589
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914233497681920
author Hosseini, Hesam
Mighan, Ghazal Hosseini
Afzali, Amirabbas
Amini, Sajjad
Houmansadr, Amir
author_facet Hosseini, Hesam
Mighan, Ghazal Hosseini
Afzali, Amirabbas
Amini, Sajjad
Houmansadr, Amir
contents Transformers have revolutionized Computer Vision (CV) through self-attention mechanisms. However, their complexity makes latent token representations difficult to interpret. We introduce ULTra, a framework for interpreting Transformer embeddings and uncovering meaningful semantic patterns within them. ULTra enables unsupervised semantic segmentation using pre-trained models without requiring fine-tuning. Additionally, we propose a self-supervised training approach that refines segmentation performance by learning an external transformation matrix without modifying the underlying model. Our method achieves state-of-the-art performance in unsupervised semantic segmentation, outperforming existing segmentation methods. Furthermore, we validate ULTra for model interpretation on both synthetic and real-world scenarios, including Object Selection and interpretable text summarization using LLMs, demonstrating its broad applicability in explaining the semantic structure of latent token representations.
format Preprint
id arxiv_https___arxiv_org_abs_2411_12589
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ULTra: Unveiling Latent Token Interpretability in Transformer-Based Understanding and Segmentation
Hosseini, Hesam
Mighan, Ghazal Hosseini
Afzali, Amirabbas
Amini, Sajjad
Houmansadr, Amir
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Transformers have revolutionized Computer Vision (CV) through self-attention mechanisms. However, their complexity makes latent token representations difficult to interpret. We introduce ULTra, a framework for interpreting Transformer embeddings and uncovering meaningful semantic patterns within them. ULTra enables unsupervised semantic segmentation using pre-trained models without requiring fine-tuning. Additionally, we propose a self-supervised training approach that refines segmentation performance by learning an external transformation matrix without modifying the underlying model. Our method achieves state-of-the-art performance in unsupervised semantic segmentation, outperforming existing segmentation methods. Furthermore, we validate ULTra for model interpretation on both synthetic and real-world scenarios, including Object Selection and interpretable text summarization using LLMs, demonstrating its broad applicability in explaining the semantic structure of latent token representations.
title ULTra: Unveiling Latent Token Interpretability in Transformer-Based Understanding and Segmentation
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2411.12589