Domain adaptation addresses the challenge where the distribution of target inference data differs from that of the source training data. Recently, data privacy has become a significant constraint, limiting access to the source domain. To mitigate this issue, Source-Free Domain Adaptation (SFDA) methods bypass source domain data by generating source-like data or pseudo-labeling the unlabeled target domain. However, these approaches often lack theoretical grounding. In this work, we provide a theoretical analysis of the SFDA problem, focusing on the general empirical risk of the unlabeled target domain. Our analysis offers a comprehensive understanding of how representativeness, generalization, and variety contribute to controlling the upper bound of target domain empirical risk in SFDA settings. We further explore how to balance this trade-off from three perspectives: sample selection, semantic domain alignment, and a progressive learning framework. These insights inform the design of novel algorithms. Experimental results demonstrate that our proposed method achieves state-of-the-art performance on three benchmark datasets—Office-Home, DomainNet, and VisDA-C—yielding relative improvements of 3.2%, 9.1%, and 7.5%, respectively, over the representative SFDA method, SHOT.

Recent studies have uncovered a troubling vulnerability in the fine-tuning stage of large language models (LLMs): even fine-tuning on entirely benign datasets can lead to a significant increase in the harmfulness of LLM outputs. Building on this finding, our red teaming study takes this threat one step further by developing a more effective attack. Specifically, we analyze and identify samples within benign datasets that contribute most to safety degradation, then fine-tune LLMs exclusively on these samples. We approach this problem from an outlier detection perspective and propose Self-Inf-N, to detect and extract outliers for fine-tuning. Our findings reveal that fine-tuning LLMs on 100 outlier samples selected by Self-Inf-N in the benign datasets severely compromises LLM safety alignment. Extensive experiments across seven mainstream LLMs demonstrate that our attack exhibits high transferability across different architectures and remains effective in practical scenarios. Alarmingly, our results indicate that most existing mitigation strategies fail to defend against this attack, underscoring the urgent need for more robust alignment safeguards. Codes are available at https://github.com/GuanZihan/Benign-Samples-Matter.

Domain adaptation has become an attractive learning paradigm, as it can leverage source domains with rich labels to deal with classification tasks in an unlabeled target domain. A few recent studies develop domain adaptation approaches for graph-structured data. In the case of node classification task, current domain adaptation methods only focus on the closed-set setting, where source and target domains share the same label space. A more practical assumption is that the target domain may contain new classes that are not included in the source domain. Therefore, in this paper, we introduce a novel and challenging problem for graphs, i.e., open-set domain adaptive node classification, and propose a new approach to solve it. Specifically, we develop an algorithm for efficient knowledge transfer from a labeled source graph to an unlabeled target graph under a separate domain alignment (SDA) strategy, in order to learn discriminative feature representations for the target graph. Our goal is to not only correctly classify target nodes into the known classes, but also classify unseen types of nodes into an unknown class. Experimental results on real-world datasets show that our method outperforms existing methods on graph domain adaptation.

Unsupervised domain adaptation (UDA) has become an appealing approach for knowledge transfer from a labeled source domain to an unlabeled target domain. However, when the classes in source and target domains are imbalanced, most existing UDA methods experience significant performance drop, as the decision boundary usually favors the majority classes. Some recent class-imbalanced domain adaptation (CDA) methods aim to tackle the challenge of biased label distribution by exploiting pseudo-labeled target samples during the training process. However, these methods suffer from the issues with unreliable pseudo labels and error accumulation during training. In this paper, we propose a pairwise adversarial training approach for class-imbalanced domain adaptation. Unlike conventional adversarial training in which the adversarial samples are obtained from the$\ell _{p}$ball of the original samples, we generate adversarial samples from the interpolated line of the aligned pairwise samples from source and target domains. The pairwise adversarial training (PAT) is a novel data-augmentation method which can be integrated into existing unsupervised domain adaptation (UDA) models to tackle the CDA problem. Inspired by the noise injection, we also extend the pairwise adversarial training to noisy pairwise adversarial training (nPAT), in which the random noise is injected into the generation of the adversarial samples. In our study, we evaluate our proposed methods as well as the baselines on three major benchmark datasets, namely Office-Home, DomainNet and Office-31. For Office-Home and Office-31, we sample the data according to the Reversely-unbalanced Source and Unbalanced Target (RS-UT) protocol so that the class distribution can be imbalanced. The extensive experimental results show that UDA models integrated with our proposed nPAT can achieve prominent improvements on most tasks compared to the baseline methods as well as the state-of-the-art CDA methods. The average accuracy of our nPAT can achieve 66.56% and 80.22% on Office-Home and DomainNet, respectively, which are higher than that of the second-best methods. Besides, Experiments also show that our method is robust to the unreliability of the pseudo labels.

As AI systems have obtained significant performance to be deployed widely in our daily lives and human society, people both enjoy the benefits brought by these technologies and suffer many social issues induced by these systems. To make AI systems good enough and trustworthy, plenty of researches have been done to build guidelines for trustworthy AI systems. Machine learning is one of the most important parts of AI systems, and representation learning is the fundamental technology in machine learning. How to make representation learning trustworthy in real-world application, e.g., cross domain scenarios, is very valuable and necessary for both machine learning and AI system fields. Inspired by the concepts in trustworthy AI, we proposed the first trustworthy representation learning across domains framework, which includes four concepts, i.e., robustness, privacy, fairness, and explainability, to give a comprehensive literature review on this research direction. Specifically, we first introduce the details of the proposed trustworthy framework for representation learning across domains. Second, we provide basic notions and comprehensively summarize existing methods for the trustworthy framework from four concepts. Finally, we conclude this survey with insights and discussions on future research directions.

Short text clustering is a challenging unsupervised learning task which requires a complex representation of each document to effectively model the semantics and syntactic structure of the text. Existing works have attempted to tackle this challenging task by incorporating additional information to the model, such as number of clusters, number of datapoints in each clusters, the distribution of the input data, and more. Unlike previous approaches, we propose to exploit an auxiliary dataset that is fully labeled to augment the quality of the learned representations. We also define the problem as cross domain clustering (XDC), which leverages adversarial learning to train an adaptive clustering model across text domains. Specifically, XDC jointly exploits a labeled source domain and an unlabeled target domain during model training. Owing to domain adversarial learning, the distribution shift across source and target domains could be mitigated. Moreover, XDC is implemented as a linkage-based clustering approach using graphs, which is agnostic of the number of clusters. We evaluate our XDC framework on three text datasets, and results show that it outperforms the state-of-the-art text clustering methods in most cases. Ablation studies and qualitative analysis also demonstrate the effectiveness of our framework.

Graph, as an important data representation, is ubiquitous in many real world applications ranging from social network analysis to biology. How to correctly and effectively learn and extract information from graph is essential for a large number of machine learning tasks. Graph embedding is a way to transform and encode the data structure in high dimensional and non-euclidean feature space to a low dimensional and structural space, which is easily exploited by other machine learning algorithms. We have witnessed a huge surge of such embedding methods, from statistical approaches to recent deep learning methods such as the graph convolutional networks (GCN). Deep learning approaches usually outperform the traditional methods in most graph learning benchmarks by building an end-to-end learning framework to optimize the loss function directly. However, most of the existing GCN methods can only perform convolution operations with node features, while ignoring the handy information in edge features, such as relations in knowledge graphs. To address this problem, we present CensNet, Convolution with Edge-Node Switching graph neural network, for learning tasks in graph-structured data with both node and edge features. CensNet is a general graph embedding framework, which embeds both nodes and edges to a latent feature space. By using line graph of the original undirected graph, the role of nodes and edges are switched, and two novel graph convolution operations are proposed for feature propagation. Experimental results on real-world academic citation networks and quantum chemistry graphs show that our approach achieves or matches the state-of-the-art performance in four graph learning tasks, including semi-supervised node classification, multi-task graph classification, graph regression, and link prediction.

Unsupervised domain adaptation (UDA) has attracted increasing attention in recent years, which adapts classifiers to an unlabeled target domain by exploiting a labeled source domain. To reduce the discrepancy between source and target domains, adversarial learning methods are typically selected to seek domain-invariant representations by confusing the domain discriminator. However, classifiers may not be well adapted to such a domain-invariant representation space, as the sample- and class-level data structures could be distorted during adversarial learning. In this article, we propose a novel transferable feature learning approach on graphs (TFLG) for unsupervised adversarial domain adaptation (DA), which jointly incorporates sample- and class-level structure information across two domains. TFLG first constructs graphs for minibatch samples and identifies the classwise correspondence across domains. A novel cross-domain graph convolutional operation is designed to jointly align the sample- and class-level structures in two domains. Moreover, a memory bank is designed to further exploit the class-level information. Extensive experiments on benchmark datasets demonstrate the effectiveness of our approach compared to the state-of-the-art UDA methods.

Unsupervised domain adaptation (UDA) has become an appealing approach for knowledge transfer from a labeled source domain to an unlabeled target domain. However, when the classes in source and target domains are imbalanced, most existing UDA methods experience significant performance drop, as the decision boundary usually favors the majority classes. Some recent class-imbalanced domain adaptation (CDA) methods aim to tackle the challenge of biased label distribution by exploiting pseudo-labeled target samples during training process. However, these methods suffer from the issues with unreliable pseudo labels and error accumulation during training. In this paper, we propose a pairwise adversarial training approach for class-imbalanced domain adaptation. Unlike conventional adversarial training in which the adversarial samples are obtained from the lp ball of the original samples, we generate adversarial samples from the interpolated line of the aligned pairwise samples from source and target domains. The pairwise adversarial training (PAT) is a novel data-augmentation method which can be integrated into existing UDA models to tackle with the CDA problem. Experimental results and ablation studies show that the UDA models integrated with our method achieve considerable improvements on benchmarks compared with the original models as well as the state-of-the-art CDA methods. Our source code is available at: https://github.com/DamoSWL/Pairwise-Adversarial-Training

Unsupervised domain adaptation aims to learn domain-invariant features across domains to transfer knowledge from a well-labeled source domain to an unlabeled target domain. Recently, some unsupervised domain adaptation methods focus on semantically aligning data distributions with pseudo-labels of the target domain. However, semantic alignment based on pseudo-labels has potential risks, e.g., inaccurate pseudo-labeling from classifier, and error accumulation from pseudo-label bias. To alleviate these risks, we propose a novel self-supervision based semantic alignment (S3A) approach for unsupervised domain adaptation, which can jointly incorporate the source alignment and cross-domain target alignment for better semantic alignment across domains. S3A consists of a two-stage semantic alignment procedure with self-supervision. One is to capture the discriminative structure of source domain by aligning source data to source class prototypes, and the other is to match each target data to its neighbor in source domain with self-supervision. Extensive experiments on three benchmark datasets demonstrate the effectiveness of our proposed method, compared with the representative adversarial learning and self-supervised learning based unsupervised domain adaptation methods.

Universal domain adaptation (UniDA) aims to transfer knowledge from a labeled source domain to an unlabeled target domain where both domains share a common label space and hold a private label space respectively. One of the most challenging goals in UniDA is to separate target samples from common classes and these from private classes without any prior knowledge on the target label space. In this paper, we propose a novel self-supervised adaptive memory network with consistency regularization for UniDA. The adaptive memory includes all target samples and source class centers, which dynamically divides target samples into common area, uncertain area, and unknown area based on the entropy. Our proposed framework jointly assigns a specific neighborhood to each target sample and clusters the target sample to its neighbor from the neighborhood. Most importantly, the proposed framework adopts consistency regularization that gradually makes the output of the classifier more reliable. This simple strategy is proved to be very effective for UniDA problem. Experimental results on two UniDA benchmarks demonstrate the effectiveness of our method.

Unsupervised domain adaptation adapts classifiers to an un-labeled target domain by exploiting a labeled source domain. To reduce discrepancy between source and target domains, adversarial learning methods are typically selected to seek domain-invariant representations by confusing the domain discriminator. However, classifiers may not be well adapted to such a domain-invariant representation space, as the sample-level and class-level data structures could be distorted during adversarial learning. In this paper, we propose a novel Transferable Feature Learning approach on Graphs (TFLG) for unsupervised adversarial domain adaptation, which jointly incorporates sample-level and class-level structure information across two domains. TFLG first constructs graphs and identifies the class-wise correspondence across domains. A novel cross-domain graph convolutional operation is designed to jointly align the sample-level and class-level structures in two domains. Extensive experiments on benchmark datasets demonstrate the effectiveness of our approach, compared to the representative unsupervised domain adaptation methods.

Owing to the remarkable capability of extracting effective graph embeddings, graph convolutional network (GCN) and its variants have been successfully applied to a broad range of tasks, such as node classification, link prediction, and graph classification. Traditional GCN models suffer from the issues of overfitting and oversmoothing, while some recent techniques like DropEdge could alleviate these issues and thus enable the development of deep GCN. However, training GCN models is non-trivial, as it is sensitive to the choice of hyperparameters such as dropout rate and learning weight decay, especially for deep GCN models. In this paper, we aim to automate the training of GCN models through hyperparameter optimization. To be specific, we propose a self-tuning GCN approach with an alternate training algorithm, and further extend our approach by incorporating the population based training scheme. Experimental results on three benchmark datasets demonstrate the effectiveness of our approaches on optimizing multi-layer GCN, compared with several representative baselines.

This paper proposes an encoder-decoder network to disentangle shape features during 3D face reconstruction from single 2D images, such that the tasks of reconstructing accurate 3D face shapes and learning discriminative shape features for face recognition can be accomplished simultaneously. Unlike existing 3D face reconstruction methods, our proposed method directly regresses dense 3D face shapes from single 2D images, and tackles identity and residual (i.e., non-identity) components in 3D face shapes explicitly and separately based on a composite 3D face shape model with latent representations. We devise a training process for the proposed network with a joint loss measuring both face identification error and 3D face shape reconstruction error. To construct training data we develop a method for fitting 3D morphable model (3DMM) to multiple 2D images of a subject. Comprehensive experiments have been done on MICC, BU3DFE, LFW and YTF databases. The results show that our method expands the capacity of 3DMM for capturing discriminative shape features and facial detail, and thus outperforms existing methods both in 3D face reconstruction accuracy and in face recognition accuracy.