Improving Source-Free Target Adaptation with Vision Transformers Leveraging Domain Representation Images

2023-11-21 · Gauransh Sawhney, Daksh Dave, Adeel Ahmed, Jiechao Gao, Khalid Saleem

Unsupervised Domain Adaptation (UDA) methods facilitate knowledge transfer from a labeled source domain to an unlabeled target domain, navigating the obstacle of domain shift. While Convolutional Neural Networks (CNNs) are a staple in UDA, the rise of Vision Transformers (ViTs) provides new avenues for domain generalization. This paper presents an innovative method to bolster ViT performance in source-free target adaptation, beginning with an evaluation of how key, query, and value elements affect ViT outcomes. Experiments indicate that altering the key component has negligible effects on Transformer performance. Leveraging this discovery, we introduce Domain Representation Images (DRIs), feeding embeddings through the key element. DRIs act as domain-specific markers, effortlessly merging with the training regimen. To assess our method, we perform target adaptation tests on the Cross Instance DRI source-only (SO) control. We measure the efficacy of target adaptation with and without DRIs, against existing benchmarks like SHOT-B* and adaptations via CDTrans. Findings demonstrate that excluding DRIs offers limited gains over SHOT-B*, while their inclusion in the key segment boosts average precision promoting superior domain generalization. This research underscores the vital role of DRIs in enhancing ViT efficiency in UDA scenarios, setting a precedent for further domain adaptation explorations.

📄 PDF Abstract BibTeX arXiv:2311.12589

Code (0)

등록된 구현이 없습니다.

Tasks

Domain Adaptation Domain Generalization Transfer Learning Unsupervised Domain Adaptation

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음

Attention 설명 없음

Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.

Dense Connections Dense Connections, or Fully Connected Connections, are a type of layer in a deep neural network that use a linear operation where every input is connected to every output…

Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…

Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…

Residual Connection 설명 없음

Domain-Specificity Inducing Transformers for Source-Free Domain Adaptation

2023-08-27 · ICCV 2023 1 · Sunandini Sanyal, Ashish Ramayee Asokan, Suvaansh Bhambri, Akshay Kulkarni 외

Conventional Domain Adaptation (DA) methods aim to learn domain-invariant feature representations to improve the target adaptation performance. However, we motivate that domain-specificity is equally important since in-d…

Disentanglement Domain Adaptation Source-Free Domain Adaptation Specificity

Aligning Non-Causal Factors for Transformer-Based Source-Free Domain Adaptation

2023-11-27 · Sunandini Sanyal, Ashish Ramayee Asokan, Suvaansh Bhambri, Pradyumna YM 외

Conventional domain adaptation algorithms aim to achieve better generalization by aligning only the task-discriminative causal factors between a source and target domain. However, we find that retaining the spurious corr…

Disentanglement Domain Adaptation Privacy Preserving Source-Free Domain Adaptation

Source-Free Domain Adaptation for RGB-D Semantic Segmentation with Vision Transformers

2023-05-23 · Giulia Rizzoli, Donald Shenaj, Pietro Zanuttigh

With the increasing availability of depth sensors, multimodal frameworks that combine color information with depth data are gaining interest. However, ground truth data for semantic segmentation is burdensome to provide,…

Domain Adaptation Segmentation Semantic Segmentation Source-Free Domain Adaptation+1

CoNMix for Source-free Single and Multi-target Domain Adaptation

2022-11-07 · Vikash Kumar, Rohit Lal, Himanshu Patil, Anirban Chakraborty

This work introduces the novel task of Source-free Multi-target Domain Adaptation and proposes adaptation framework comprising of \textbf{Co}nsistency with \textbf{N}uclear-Norm Maximization and \textbf{Mix}Up knowledge …

Domain Adaptation Knowledge Distillation Multi-target Domain Adaptation Pseudo Label

Rethinking the Backbone in Class Imbalanced Federated Source Free Domain Adaptation: The Utility of Vision Foundation Models

2025-09-10 · Kosuke Kihara, Junki Mori, Taiki Miyagawa, Akinori F. Ebihara arxiv

Federated Learning (FL) offers a framework for training models collaboratively while preserving data privacy of each client. Recently, research has focused on Federated Source-Free Domain Adaptation (FFREEDA), a more rea…

Source-Free Domain Adaptation Federated Learning