This paper explores higher-resolution video outpainting with extensive content generation. We point out common issues faced by existing methods when attempting to largely outpaint videos: the generation of low-quality content and limitations imposed by GPU memory. To address these challenges, we propose a diffusion-based method called Infinite-Canvas. It builds upon two core designs. First, instead of employing the common practice of "single-shot" outpainting, we distribute the task across spatial windows and seamlessly merge them. It allows us to outpaint videos of any size and resolution without being constrained by GPU memory. Second, the source video and its relative positional relation are injected into the generation process of each window. It makes the generated spatial layout within each window harmonize with the source video. Coupling with these two designs enables us to generate higher-resolution outpainting videos with rich content while keeping spatial and temporal consistency. Infinite-Canvas excels in large-scale video outpainting, e.g., from 512 × 512 to 1152 × 2048 (9×), while producing high-quality and aesthetically pleasing results. It achieves the best quantitative results across various resolution and scale setups. The code is available at https://github.com/mayuelala/FollowYourCanvas.

In contemporary machine learning, enhancing model generalization through diversified datasets is essential. Yet, collecting additional data often faces prohibitive costs and privacy constraints, with no guarantee of improved diversity. In this paper, we propose Domain-Diff, featuring a pivotal Word-to-Image Mapping (WIM) mechanism. WIM constructs precise mapping between prompts and images, where the prompts only comprise style and class words. It generates intra-domain data by employing identical prompts to produce source-style images, preserving style and class consistency, thereby diversifying the dataset. Expanding on this innovation, we fuse multiple WIMs and use the prompts with multiple style words to create inter-domain data, which captures a fusion style of multiple source domains. Inter-domain data significantly widens the training data distribution, amplifying diversity. Experimental results demonstrate DomainDiff’s transformative potential, improving model performance on real-world data compared to using only real data. These findings highlight DomainDiff’s utility in enhancing generalization across diverse real-world scenarios.

Large-scale vision-language (V-L) models have demonstrated remarkable generalization capabilities for downstream tasks through prompt tuning. However, the mechanisms behind the learned text representations are unknown, limiting further generalization gains, and the limitations are more severe when faced with the prevalent class imbalances seen in web-sourced datasets. Recent advances in the neural collapse (NC) phenomenon of vision-only models suggest that the optimal representation structure is the simplex ETF, which paves the way to study representations in V-L models. In this paper, we make the first attempt to use NC for examining the representations in V-L models via prompt tuning. It is found that NC optimality of text-to-image representations shows a positive correlation with downstream generalizability, which is more severe under class imbalance settings. To improve the representations, we propose Neural-collapse-anchored Prompt Tuning (NPT), a novel method that learns prompts with text and image representations that satisfy the same simplex Equiangular Tight Frame (ETF). NPT incorporates two regularization terms: language-modality collapse and multi-modality isomorphism; and it is compatible with other prompt tuning methods. Extensive experiments show that NPT can consistently help to improve existing prompt tuning techniques across 11 datasets for both balanced and imbalanced settings.

We present Follow-Your-Emoji, a diffusion-based framework for portrait animation, which animates a reference portrait with target landmark sequences. The main challenge of portrait animation is to preserve the identity of the reference portrait and transfer the target expression to this portrait while maintaining temporal consistency and fidelity. To address these challenges, Follow-Your-Emoji equipped the powerful Stable Diffusion model with two well-designed technologies. Specifically, we first adopt a new explicit motion signal, namely expression-aware landmark, to guide the animation process. We discover this landmark can not only ensure the accurate motion alignment between the reference portrait and target motion during inference but also increase the ability to portray exaggerated expressions (i.e., large pupil movements) and avoid identity leakage. Then, we propose a facial fine-grained loss to improve the model’s ability of subtle expression perception and reference portrait appearance reconstruction by using both expression and facial masks. Accordingly, our method demonstrates significant performance in controlling the expression of freestyle portraits, including real humans, cartoons, sculptures, and even animals. By leveraging a simple and effective progressive generation strategy, we extend our model to stable long-term animation, thus increasing its potential application value. To address the lack of a benchmark for this field, we introduce EmojiBench, a comprehensive benchmark comprising diverse portrait images, driving videos, and landmarks. We show extensive evaluations on EmojiBench to verify the superiority of Follow-Your-Emoji. The code, training dataset and benchmark will be found in https://github.com/mayuelala/FollowYourEmoji.

Domain generalization (DG) aims to generalize the knowledge learned from multiple source domains to unseen target domains. Existing DG techniques can be subsumed under two broad categories, i.e., domain-invariant representation learning and domain manipulation. Nevertheless, it is extremely difficult to explicitly augment or generate the unseen target data. And when source domain variety increases, developing a domain-invariant model by simply aligning more domain-specific information becomes more challenging. In this paper, we propose a simple yet effective method for domain generalization, named Knowledge Distillation based Domain-invariant Representation Learning (KDDRL), that learns domain-invariant representation while encouraging the model to maintain domain-specific features, which recently turned out to be effective for domain generalization. To this end, our method incorporates multiple auxiliary student models and one student leader model to perform a two-stage distillation. In the first-stage distillation, each domain-specific auxiliary student treats the ensemble of other auxiliary students' predictions as a target, which helps to excavate the domain-invariant representation. Also, we present an error removal module to prevent the transfer of faulty information by eliminating incorrect predictions compared to the true labels. In the second-stage distillation, the student leader model with domain-specific features combines the domain-invariant representation learned from the group of auxiliary students to make the final prediction. Extensive experiments and in-depth analysis on popular DG benchmark datasets demonstrate that our KDDRL significantly outperforms the current state-of-the-art methods.

Prompt learning has become one of the most efficient paradigms for adapting large pre-trained vision-language models to downstream tasks. Current state-of-the-art methods, like CoOp and ProDA, tend to adopt soft prompts to learn an appropriate prompt for each specific task. Recent CoCoOp further boosts the base-to-new generalization performance via an image-conditional prompt. However, it directly fuses identical image semantics to prompts of different labels and significantly weakens the discrimination among different classes as shown in our experiments. Motivated by this observation, we first propose a class-aware text prompt (CTP) to enrich generated prompts with label-related image information. Unlike CoCoOp, CTP can effectively involve image semantics and avoid introducing extra ambiguities into different prompts. On the other hand, instead of reserving the complete image representations, we propose text-guided feature tuning (TFT) to make the image branch attend to class-related representation. A contrastive loss is employed to align such augmented text and image representations on downstream tasks. In this way, the image-to-text CTP and text-to-image TFT can be mutually promoted to enhance the adaptation of VLMs for downstream tasks. Extensive experiments demonstrate that our method outperforms the existing methods by a significant margin. Especially, compared to CoCoOp, we achieve an average improvement of 4.03% on new classes and 3.19% on harmonic-mean over eleven classification benchmarks.

Deep learning has achieved tremendous success in recent years, but most of these successes are built on an independent and identically distributed (IID) assumption. This somewhat hinders the application of deep learning to the more challenging out-of-distribution (OOD) scenarios. Although many OOD methods have been proposed to address this problem and have obtained good performance on testing data that is of major shifts with training distributions, interestingly, we experimentally find that these methods achieve excellent OOD performance by making a great sacrifice of the IID performance. We call this finding the IID-OOD dilemma. Clearly, in real-world applications, distribution shifts between training and testing data are often uncertain, where shifts could be minor, and even close to the IID scenario, and thus it is truly important to design a deep model with the balanced generalization ability between IID and OOD. To this end, in this paper, we investigate an intriguing problem of balancing IID and OOD generalizations and propose a novel Model Agnostic adaPters (MAP) method, which is more reliable and effective for distribution-shift-agnostic real-world data. Our key technical contribution is to use auxiliary adapter layers to incorporate the inductive bias of IID into OOD methods. To achieve this goal, we apply a bilevel optimization to explicitly model and optimize the coupling relationship between the OOD model and auxiliary adapter layers. We also theoretically give a first-order approximation to save computational time. Experimental results on six datasets successfully demonstrate that MAP can greatly improve the performance of IID while achieving good OOD performance.

Universal domain adaptation (UniDA) aims to transfer knowledge from the source domain to the target domain without any prior knowledge about the label set. The challenge lies in how to determine whether the target samples belong to common categories. The mainstream methods make judgments based on the sample features, which overemphasizes global information while ignoring the most crucial local objects in the image, resulting in limited accuracy. To address this issue, we propose a Universal Attention Matching (UniAM) framework by exploiting the self-attention mechanism in vision transformer to capture the crucial object information. The proposed framework introduces a novel Compressive Attention Matching (CAM) approach to explore the core information by compressively representing attentions. Furthermore, CAM incorporates a residual-based measurement to determine the sample commonness. By utilizing the measurement, UniAM achieves domain-wise and category-wise Common Feature Alignment (CFA) and Target Class Separation (TCS). Notably, UniAM is the first method utilizing the attention in vision transformer directly to perform classification tasks. Extensive experiments show that UniAM outperforms the current state-of-the-art methods on various benchmark datasets.

Domain generalization (DG) is a prevalent problem in real-world applications, which aims to train well-generalized models for unseen target domains by utilizing several source domains. Since domain labels, i.e., which domain each data point is sampled from, naturally exist, most DG algorithms treat them as a kind of supervision information to improve the generalization performance. However, the original domain labels may not be the optimal supervision signal due to the lack of domain heterogeneity, i.e., the diversity among domains. For example, a sample in one domain may be closer to another domain, its original label thus can be the noise to disturb the generalization learning. Although some methods try to solve it by re-dividing domains and applying the newly generated dividing pattern, the pattern they choose may not be the most heterogeneous due to the lack of the metric for heterogeneity. In this paper, we point out that domain heterogeneity mainly lies in variant features under the invariant learning framework. With contrastive learning, we propose a learning potential-guided metric for domain heterogeneity by promoting learning variant features. Then we notice the differences between seeking variance-based heterogeneity and training invariance-based generalizable model. We thus propose a novel method called H eterogeneity-based Two-stage Contrastive Learning (HTCL) for the DG task. In the first stage, we generate the most heterogeneous dividing pattern with our contrastive metric. In the second stage, we employ an invariance-aimed contrastive learning by re-building pairs with the stable relation hinted by domains and classes, which better utilizes generated domain labels for generalization learning. Extensive experiments show HTCL better digs heterogeneity and yields great generalization performance.

Model pre-training is essential in human-centric perception. In this paper, we first introduce masked image modeling (MIM) as a pre-training approach for this task. Upon revisiting the MIM training strategy, we reveal that human structure priors offer significant potential. Motivated by this insight, we further incorporate an intuitive human structure prior - human parts - into pre-training. Specifically, we employ this prior to guide the mask sampling process. Image patches, corresponding to human part regions, have high priority to be masked out. This encourages the model to concentrate more on body structure information during pre-training, yielding substantial benefits across a range of human-centric perception tasks. To further capture human characteristics, we propose a structure-invariant alignment loss that enforces different masked views, guided by the human part prior, to be closely aligned for the same image. We term the entire method as HAP. HAP simply uses a plain ViT as the encoder yet establishes new state-of-the-art performance on 11 human-centric benchmarks, and on-par result on one dataset. For example, HAP achieves 78.1% mAP on MSMT17 for person re-identification, 86.54% mA on PA-100K for pedestrian attribute recognition, 78.2% AP on MS COCO for 2D pose estimation, and 56.0 PA-MPJPE on 3DPW for 3D pose and shape estimation.

Domain generalization (DG) aims to learn from multiple source domains a model that can generalize well on unseen target domains. Existing DG methods mainly learn the representations with invariant marginal distribution of the input features, however, the invariance of the conditional distribution of the labels given the input features is more essential for unknown domain prediction. Meanwhile, the existing of unobserved confounders which affect the input features and labels simultaneously cause spurious correlation and hinder the learning of the invariant relationship contained in the conditional distribution. Interestingly, with a causal view on the data generating process, we find that the input features of one domain are valid instrumental variables for other domains. Inspired by this finding, we propose an instrumental variable-driven DG method (IV-DG) by removing the bias of the unobserved confounders with two-stage learning. In the first stage, it learns the conditional distribution of the input features of one domain given input features of another domain. In the second stage, it estimates the relationship by predicting labels with the learned conditional distribution. Theoretical analyses and simulation experiments show that it accurately captures the invariant relationship. Extensive experiments on real-world datasets demonstrate that IV-DG method yields state-of-the-art results.

In observational studies, confounder separation and balancing are the fundamental problems of treatment effect estimation. Most of the previous methods focused on addressing the problem of confounder balancing by treating all observed pre-treatment variables as confounders, ignoring confounder separation. In general, not all the observed pre-treatment variables are confounders that refer to the common causes of the treatment and the outcome, some variables only contribute to the treatment (i.e., instrumental variables) and some only contribute to the outcome (i.e., adjustment variables). Balancing those non-confounders, including instrumental variables and adjustment variables, would generate additional bias for treatment effect estimation. By modeling the different causal relations among observed pre-treatment variables, treatment variables and outcome variables, we propose a synergistic learning framework to i) separate confounders by learning decomposed representations of both confounders and non-confounders, ii) balance confounder with sample re-weighting technique, and simultaneously iii) estimate the treatment effect in observational studies via counterfactual inference. Empirical results on synthetic and real-world datasets demonstrate that the proposed method can precisely decompose confounders and achieve a more precise estimation of treatment effect than baselines.

Domain generalization (DG) aims to learn from multiple known source domains a model that can generalize well to unknown target domains. The existing DG methods usually exploit the fusion of shared multi-source data to train a generalizable model. However, tremendous data is distributed across lots of places nowadays that can not be shared due to privacy policies. In this paper, we tackle the problem of federated domain generalization where the source datasets can only be accessed and learned locally for privacy protection. We propose a novel framework called Collaborative Semantic Aggregation and Calibration (CSAC) to enable this challenging problem. To fully absorb multi-source semantic information while avoiding unsafe data fusion, we conduct data-free semantic aggregation by fusing the models trained on the separated domains layer-by-layer. To address the semantic dislocation problem caused by domain shift, we further design cross-layer semantic calibration with an attention mechanism to align each semantic level and enhance domain invariance. We unify multi-source semantic learning and alignment in a collaborative way by repeating the semantic aggregation and calibration alternately, keeping each dataset localized, and the data privacy is carefully protected. Extensive experiments show the significant performance of our method in addressing this challenging problem.

Considerable progress has been made in domain generalization (DG) which aims to learn a generalizable model from multiple well-annotated source domains to unknown target domains. However, it can be prohibitively expensive to obtain sufficient annotation for source datasets in many real scenarios. To escape from the dilemma between domain generalization and annotation costs, in this paper, we introduce a novel task named label-efficient domain generalization (LEDG) to enable model generalization with label-limited source domains. To address this challenging task, we propose a novel framework called Collaborative Exploration and Generalization (CEG) which jointly optimizes active exploration and semi-supervised generalization. Specifically, in active exploration, to explore class and domain discriminability while avoiding information divergence and redundancy, we query the labels of the samples with the highest overall ranking of class uncertainty, domain representativeness, and information diversity. In semi-supervised generalization, we design MixUp-based intra- and inter-domain knowledge augmentation to expand domain knowledge and generalize domain invariance. We unify active exploration and semi-supervised generalization in a collaborative way and promote mutual enhancement between them, boosting model generalization with limited annotation. Extensive experiments show that CEG yields superior generalization performance. In particular, CEG can even use only 5% data annotation budget to achieve competitive results compared to the previous DG methods with fully labeled data on PACS dataset.

Instrumental variables (IVs), sources of treatment randomization that are conditionally independent of the outcome, play an important role in causal inference with unobserved confounders. However, the existing IV-based counterfactual prediction methods need well-predefined IVs, while it’s an art rather than science to find valid IVs in many real-world scenes. Moreover, the predefined hand-made IVs could be weak or erroneous by violating the conditions of valid IVs. These thorny facts hinder the application of the IV-based counterfactual prediction methods. In this article, we propose a novel Automatic Instrumental Variable decomposition (AutoIV) algorithm to automatically generate representations serving the role of IVs from observed variables (IV candidates). Specifically, we let the learned IV representations satisfy the relevance condition with the treatment and exclusion condition with the outcome via mutual information maximization and minimization constraints, respectively. We also learn confounder representations by encouraging them to be relevant to both the treatment and the outcome. The IV and confounder representations compete for the information with their constraints in an adversarial game, which allows us to get valid IV representations for IV-based counterfactual prediction. Extensive experiments demonstrate that our method generates valid IV representations for accurate IV-based counterfactual prediction.

Real-world networks exhibit prominent hierarchical and modular structures, with various subgraphs as building blocks. Most existing studies simply consider distinct subgraphs as motifs and use only their numbers to characterize the underlying network. Although such statistics can be used to describe a network model, or even to design some network algorithms, the role of subgraphs in such applications can be further explored so as to improve the results. In this article, the concept of subgraph network (SGN) is introduced and then applied to network models, with algorithms designed for constructing the 1st-order and 2nd-order SGNs, which can be easily extended to build higher-order ones. Furthermore, these SGNs are used to expand the structural feature space of the underlying network, beneficial for network classification. Numerical experiments demonstrate that the network classification model based on the structural features of the original network together with the 1st-order and 2nd-order SGNs always performs the best as compared to the models based only on one or two of such networks. In other words, the structural features of SGNs can complement that of the original network for better network classification, regardless of the feature extraction method used, such as the handcrafted, network embedding and kernel-based methods.

For efficient malware detection, there are more and more deep learning methods based on raw software binaries. Recent studies show that deep learning models can easily be fooled to make a wrong decision by introducing subtle perturbations to inputs, which attracts a large influx of work in adversarial attacks. However, most of the existing attack methods are based on manual features (e.g., API calls) or in the white-box setting, making the attacks impractical in current real-world scenarios. In this work, we propose a novel attack framework called GAPGAN, which generates adversarial payloads (padding bytes) with generative adversarial networks (GANs). To the best of our knowledge, it is the first work that performs end-to-end black-box attacks at the byte-level against deep learning based malware binaries detection. In our attack framework, we map input discrete malware binaries to continuous space, then feed it to the generator of GAPGAN to generate adversarial payloads. We append payloads to the original binaries to craft an adversarial sample while preserving its functionality. We propose to use a dynamic threshold for reducing the loss of the effectiveness of the payloads when mapping it from continuous format back to the original discrete format. For balancing the attention of the generator to the payloads and the adversarial samples, we use an automatic weight tuning strategy. We train GAPGAN with both malicious and benign software. Once the training is finished, the generator can generate an adversarial sample with only the input malware in less than twenty milliseconds. We apply GAPGAN to attack the state-of-the-art detector MalConv and achieve 100% attack success rate with only appending payloads of 2.5% of the total length of the data for detection. We also attack deep learning models with different structures under different defense methods. The experiments show that GAPGAN outperforms other state-of-the-art attack models in efficiency and effectiveness.

Attackers often use domain generation algorithms (DGAs) to create various kinds of pseudorandom domains dynamically and select a part of them to connect with command and control servers, therefore it is important to automatically detect the algorithmically generated domains (AGDs). AGDs can be broken down into two categories: character-based domains and wordlist-based domains. Recently, methods based on machine learning and deep learning have been widely explored. However, much of the previous work perform well in detecting one kind of DGA families but poorly in classifying another kind. A general detection system which is applicable to both kinds of domains still remains a challenge. To address this problem, we propose a novel real-time detection method with high accuracy as well as high coverage. We first convey a domain name into a sequence of word-level or character-level components, then design a deep neural network based on temporal convolutional network to extract the implicit pattern and classify the domain into two or more categories. Our experimental results demonstrate that our model outperforms state-of-the-art approaches in both binary classification and multi-class classification, and shows a good performance in detecting different kinds of DGAs. Besides, the high training efficiency of our model makes it adjust to new malicious domains quickly.