Text-to-image diffusion models have achieved remarkable success in generating photorealistic images. However, the inclusion of sensitive information during pre-training poses significant risks. Machine Unlearning (MU) offers a promising solution to eliminate sensitive concepts from these models. Despite its potential, existing MU methods face two main challenges: 1) limited generalization, where concept erasure is effective only within the unlearned set, failing to prevent sensitive concept generation from out-of-set prompts; and 2) utility degradation, where removing target concepts significantly impacts the model's overall performance. To address these issues, we propose a novel concept domain correction framework named \textbf{DoCo} (\textbf{Do}main \textbf{Co}rrection). By aligning the output domains of sensitive and anchor concepts through adversarial training, our approach ensures comprehensive unlearning of target concepts. Additionally, we introduce a concept-preserving gradient surgery technique that mitigates conflicting gradient components, thereby preserving the model's utility while unlearning specific concepts. Extensive experiments across various instances, styles, and offensive concepts demonstrate the effectiveness of our method in unlearning targeted concepts with minimal impact on related concepts, outperforming previous approaches even for out-of-distribution prompts.

In the dynamic realm of digital media, latent diffusion models (LDM) have revolutionized the generation of videos, surpassing the capabilities of traditional generative models. This paper presents Stable Video Signature, a pioneering watermarking framework for LDM in video generation. Addressing the pressing need for copyright and model protection, our approach is the first to implant watermarks directly into the generation process of LDM based on video through a novel two-stage process. We first encode watermarks into the video’s latent space embedding, ensuring a holistic temporal decoding mechanism of LDM. Then watermark is integrated into the LDM’s decoder. In this process, our method can maintain frame consistency, preserving the quality and robustness of generated videos during watermark implantation. We further show that the framework embeds watermarks seamlessly into LDM, maintaining the original functionality of the models and exhibiting resilience against a spectrum of watermark attacks. Our comprehensive experiments on both text-to-video and image-to-video generation tasks substantiate the efficacy of Stable Video Signature. This work not only pioneers watermarking in video generation LDM, but also sets a precedent for safeguarding intellectual property in the age of advanced media synthesis.

The integration of depth sensing and laser scanning technologies has propelled point cloud data to the forefront of 3D graphical modeling. This paper addresses a critical gap in the literature: the protection of intellectual property in generating point clouds using Diffusion Models (DMs). We introduce Dual-Process Watermarked Diffusion (DPWD), a pioneering watermarking framework for Diffusion Models (DMs) used in point cloud generation. DPWD embeds watermarks directly into DMs, ensuring strong protection against intellectual property theft. To do so, we introduce a two-stage watermark strategy: 1) watermark embedding using a permutation invariant module, and 2) watermark integration into the DM’s noise predictor. The framework is robust against common attacks and preserves the quality of generated point clouds. Empirical results on various point cloud tasks demonstrate DPWD’s effectiveness in safeguarding intellectual property rights without compromising model performance. DPWD sets a new standard for model protection in the GenAI era.

Machine unlearning (MU) aims to efficiently remove sensitive or harmful memory from a pre-trained model. The key challenge is to balance the potential tradeoff between unlearning efficacy and utility preservation, which involves forgetting undesirable information as defined while maintaining the model's original performance. One potential way to tackle this problem is to use multi-objective optimization to jointly optimize both the unlearning and utility preservation objectives. However, existing multi-objective methods only guarantee finding a Pareto-optimal solution without fine-grained control, which causes under-optimization of the unlearning objective. To this end, we first model MU as a constrained optimization problem, that is, optimizing the unlearning objective under the constraint of a bounded increase for utility loss. We then show that solving this optimization problem is equivalent to unilateral gradient surgery on the unlearning objective. To resolve the additional computational cost brought by gradient surgery, we propose an implicit gradient surgery method, which approximates the solution to the aforementioned constrained optimization problem via only one backpropagation, thereby achieving efficient utility-preserving MU. Theoretically, we provide a tight convergence analysis of the algorithm. Empirically, our extensive experiments show that the proposed algorithm achieves better tradeoff results than existing baselines. Codes are available at https://github.com/anseryuer/EUPMU-Efficient-Utility-Preserving-Machine-Unlearning.

Concept Bottleneck Models (CBMs) map the black-box visual representations extracted by deep neural networks onto a set of interpretable concepts and use the concepts to make predictions, enhancing the transparency of the decision-making process. Multimodal pre-trained models can match visual representations with textual concept embeddings, allowing for obtaining the interpretable concept bottleneck without the expertise concept annotations. Recent research has focused on the concept bank establishment and the high-quality concept selection. However, it is challenging to construct a comprehensive concept bank through humans or large language models, which severely limits the performance of CBMs. In this work, we propose the Incremental Residual Concept Bottleneck Model (Res-CBM) to address the challenge of concept completeness. Specifically, the residual concept bottleneck model employs a set of optimizable vectors to complete missing concepts, then the incremental concept discovery module converts the complemented vectors with unclear meanings into potential concepts in the candidate concept bank. Our approach can be applied to any user-defined concept bank, as a post-hoc processing method to enhance the performance of any CBMs. Furthermore, to measure the descriptive efficiency of CBMs, the Concept Utilization Efficiency (CUE) metric is proposed. Experiments show that the Res-CBM outperforms the current state-of-the-art methods in terms of both accuracy and efficiency and achieves comparable performance to black-box models across multiple datasets.

With the growing size of pre-trained models, full fine-tuning and storing all the parameters for various down-stream tasks is costly and infeasible. In this paper, we propose a new parameter-efficient fine-tuning method, Gradient-based Parameter Selection (GPS), demonstrating that only tuning a few selected parameters from the pre-trained model while keeping the remainder of the model frozen can generate similar or better performance compared with the full model fine-tuning method. Different from the existing popular and state-of-the-art parameter-efficient fine-tuning approaches, our method does not in-troduce any additional parameters and computational costs during both the training and inference stages. Another ad-vantage is the model-agnostic and non-destructive property, which eliminates the need for any other design specific to a particular model. Compared with the full fine-tuning, GPS achieves 3.33% (91.78% vs. 88.45%, FGVC) and 9.61% (73.1% vs. 65.57%, VTAB) improvement of the accu-racy with tuning only 0.36% parameters of the pre-trained model on average over 24 image classification tasks; it also demonstrates a significant improvement of 17% and 16.8% in mDice and mIoU, respectively, on medical image segmentation task. Moreover, GPS achieves state-of-the-art performance compared with existing PEFT meth-ods. The code will be available in https://github.com/FightingFighting/GPS.git.

Multi-task learning (MTL) considers learning a joint model for multiple tasks by optimizing a convex combination of all task losses. To solve the optimization problem, existing methods use an adaptive weight updating scheme, where task weights are dynamically adjusted based on their respective losses to prioritize difficult tasks. However, these algorithms face a great challenge whenever label noise is present, in which case excessive weights tend to be assigned to noisy tasks that have relatively large Bayes optimal errors, thereby overshadowing other tasks and causing performance to drop across the board. To overcome this limitation, we propose Multi-Task Learning with Excess Risks (ExcessMTL), an excess risk-based task balancing method that updates the task weights by their distances to convergence instead. Intuitively, ExcessMTL assigns higher weights to worse-trained tasks that are further from convergence. To estimate the excess risks, we develop an efficient and accurate method with Taylor approximation. Theoretically, we show that our proposed algorithm achieves convergence guarantees and Pareto stationarity. Empirically, we evaluate our algorithm on various MTL benchmarks and demonstrate its superior performance over existing methods in the presence of label noise. Our code is available at https://github.com/yifei-he/ExcessMTL.

Graph neural networks (GNNs) have achieved tremendous success in the task of graph classification and its diverse downstream real-world applications. Despite the huge success in learning graph representations, current GNN models have demonstrated their vulnerability to potentially existent adversarial examples on graph-structured data. Existing approaches are either limited to structure attacks or restricted to local informatio, urging for the design of a more general attack framework on graph classification, which faces significant challenges due to the complexity of generatinglocal-node-leveladversarial examples using theglobal-graph-levelinformation. To address this ”global-to-local” attack challenge, we present a novel and general frameworkCAMAto generate adversarial examples via manipulating graph structure and node features. Specifically, we make use of Graph Class Activation Mapping and its variant to produce node-level importance corresponding to the graph classification task. Then through a heuristic design of algorithms, we can perform both feature and structure attacks under unnoticeable perturbation budgets with the help of both node-level and subgraph-level importance. Experiments towards attacking four state-of-the-art graph classification models on six real-world benchmarks verify the flexibility and effectiveness of our framework.

Despite the broad interest in meta-learning, the generalization problem remains one of the significant challenges in this field. Existing works focus on meta-generalization to unseen tasks at the meta-level by regularizing the meta-loss, while ignoring that adapted models may not generalize to the task domains at the adaptation level. In this paper, we propose a new regularization mechanism for meta-learning - Minimax-Meta Regularization, which employs inverted regularization at the inner loop and ordinary regularization at the outer loop during training. In particular, the inner inverted regularization makes the adapted model more difficult to generalize to task domains; thus, optimizing the outer-loop loss forces the meta-model to learn meta-knowledge with better generalization. Theoretically, we prove that inverted regularization improves the meta-testing performance by reducing generalization errors. We conduct extensive experiments on the representative scenarios, and the results show that our method consistently improves the performance of meta-learning algorithms.

In code vulnerability detection tasks, a detector trained on a label-rich source domain fails to provide accurate prediction on new or unseen target domains due to the lack of labeled training data on target domains.Previous studies mainly utilize domain adaptation to perform cross-domain vulnerability detection.But they ignore the negative effect of private semantic characteristics of the target domain for domain alignment, which easily causes the problem of negative transfer.In addition, these methods forcibly reduce the distribution discrepancy between domains and do not take into account the interference of irrelevant target instances for distributional domain alignment, which leads to the problem of excessive alignment.To address the above issues, we propose a novel cross-domain code vulnerability detection framework named MN-CRI.Specifically, we introduce mutual nearest neighbor contrastive learning to align the source domain and target domain geometrically, which could align the common semantic characteristics of two domains and separate out the private semantic characteristics of each domain.Furthermore, we introduce an instance re-weighting scheme to alleviate the problem of excessive alignment.This scheme dynamically assign different weights to instances, reducing the contribution of irrelevant instances so as to achieve better domain alignment.Finally, extensive experiments demonstrate that MNCRI significantly outperforms state-of-the-art cross-domain code vulnerability detection methods by a large margin.

Current learning-based edge caching schemes usually suffer from dynamic content popularity, e.g., in the emerging short video platforms, users’ request patterns shift significantly over time and across different edges. An intuitive solution for a specific local edge cache is to collect more request histories from other edge caches. However, uniformly merging these request histories may not perform satisfactorily due to heterogeneous content distributions on different edges. To solve this problem, we propose a collaborative edge caching framework. First, we design a meta-learning-based collaborative strategy to guarantee that the local model can timely meet the continually changing content popularity. Then, we design an edge sampling method to select more "valuable" neighbor edges to participate in the local training. To evaluate the proposed framework, we conduct trace-driven experiments to demonstrate the effectiveness of our design: it improves the average cache hit rate by up to 10.12% (normalized) compared with other baselines.

The content request patterns perceived by edge devices are becoming highly dynamic, especially for emerging short video platforms compared to traditional video platforms. This calls for caching policies that can continuously adapt to dynamic environments, challenging previously popular reinforcement learning (RL)-based policies. A straightforward solution, i.e., repeatedly restarting and training RL agents, would fail to converge timely while meeting the observed adaptation process. Offering transferable knowledge is considered a possible method to speed up the adaptation process. Unfortunately, it fails to outperform the RL-based approach as an alternative solution in these scenarios. To alleviate this drawback, we 1) design a sequential-pair meta-learning for edge caching that captures the meta-knowledge of dynamic changes from sequential-pair-wise intervals, which are segmentations from the whole dynamic episode, and 2) develop an online meta-RL-based solution called Online Meta Actor-Critic (OMAC), which updates the meta-knowledge in an online manner. To evaluate the proposed framework, we conduct trace-driven experiments to demonstrate the effectiveness of our design: it improves the average cache hit rate by up to 37.4% (normalized) compared with other baselines.

Model-free RL-based recommender systems have recently received increasing research attention due to their capability to handle partial feedback and long-term rewards. However, most existing research has ignored a critical feature in recommender systems: one user's feedback on the same item at different times is random. The stochastic rewards property essentially differs from that in classic RL scenarios with deterministic rewards, which makes RL-based recommender systems much more challenging. In this paper, we first demonstrate in a simulator environment where using direct stochastic feedback results in a significant drop in performance. Then to handle the stochastic feedback more efficiently, we design two stochastic reward stabilization frameworks that replace the direct stochastic feedback with that learned by a supervised model. Both frameworks are model-agnostic, i.e., they can effectively utilize various supervised models. We demonstrate the superiority of the proposed frameworks over different RL-based recommendation baselines with extensive experiments on a recommendation simulator as well as an industrial-level recommender system.

We study the budget allocation problem in online marketing campaigns that utilize previously collected offline data. We first discuss the long-term effect of optimizing marketing budget allocation decisions in the offline setting. To overcome the challenge, we propose a novel game-theoretic offline value-based reinforcement learning method using mixed policies. The proposed method reduces the need to store infinitely many policies in previous methods to only constantly many policies, which achieves nearly optimal policy efficiency, making it practical and favorable for industrial usage. We further show that this method is guaranteed to converge to the optimal policy, which cannot be achieved by previous value-based reinforcement learning methods for marketing budget allocation. Our experiments on a large-scale marketing campaign with tens-of-millions users and more than one billion budget verify the theoretical results and show that the proposed method outperforms various baseline methods. The proposed method has been successfully deployed to serve all the traffic of this marketing campaign.

The rapid growth of rich multimedia data in today’s Internet, especially video traffic, has challenged the content delivery networks (CDNs). Caching serves as an important means to reduce user access latency so as to enable faster content downloads. Motivated by the dynamic nature of the real-world edge traces, this paper introduces aprovably wellonline caching policy in dynamic environments where: 1) the popularity is highly dynamic; 2) no regular stochastic pattern can model this dynamic evaluation process. First, we design an online optimization framework, which aims to minimize thedynamic regretthat finds the distance between an online caching policy and the best dynamic policy in hindsight. Second, we propose a dynamic online learning method to solve the non-stationary caching problem formulated in the previous framework. Compared to the linear dynamic regret of previous methods, our proposal is proved to achieve asublinear dynamic regret, from which it is guaranteed to be nearly optimal. We verify the design using both synthetic and real-world traces: the proposed policy achieves the best performance in the synthetic traces with different levels of dynamicity, which verifies the dynamic adaptation; our proposal consistently achieves at least 9.4% improvement than the baselines, including LRU, LFU, Static Online Learning based replacement, and Deep Reinforcement Learning based replacement, in random edge areas from real-world traces (from iQIYI), further verifying the effectiveness and robustness on the edge.

Models trained with offline data often suffer from continual distribution shifts and expensive labeling in changing environments. This calls for a new online learning paradigm where the learner can continually adapt to changing environments with limited labels. In this paper, we propose a new online setting – Online Active Continual Adaptation, where the learner aims to continually adapt to changing distributions using both unlabeled samples and active queries of limited labels. To this end, we propose Online Self-Adaptive Mirror Descent (OSAMD), which adopts an online teacher-student structure to enable online self-training from unlabeled data, and a margin-based criterion that decides whether to query the labels to track changing distributions. Theoretically, we show that, in the separable case, OSAMD has an $O({T}^{2/3})$ dynamic regret bound under mild assumptions, which is aligned with the $\Omega(T^{2/3})$ lower bound of online learning algorithms with full labels. In the general case, we show a regret bound of $O({T}^{2/3} + \alpha^* T)$, where $\alpha^*$ denotes the separability of domains and is usually small. Our theoretical results show that OSAMD can fast adapt to changing environments with active queries. Empirically, we demonstrate that OSAMD achieves favorable regrets under changing environments with limited labels on both simulated and real-world data, which corroborates our theoretical findings.

The conflicting gradients problem is one of the major bottlenecks for the effective training of machine learning models that deal with multiple objectives. To resolve this problem, various gradient manipulation techniques, such as PCGrad, MGDA, and CAGrad, have been developed, which directly alter the conflicting gradients to refined ones with alleviated or even no conflicts. However, the existing design and analysis of these techniques are mainly conducted under the full-batch gradient setting, ignoring the fact that they are primarily applied with stochastic mini-batch gradients. In this paper, we illustrate that the stochastic gradient manipulation algorithms may fail to converge to Pareto optimal solutions. Firstly, we show that these different algorithms can be summarized into a unified algorithmic framework, where the descent direction is given by the composition of the gradients of the multiple objectives. Then we provide an explicit two-objective convex optimization instance to explicate the non-convergence issue under the unified framework, which suggests that the non-convergence results from the determination of the composite weights solely by the instantaneous stochastic gradients. To fix the non-convergence issue, we propose a novel composite weights determination scheme that exponentially averages the past calculated weights. Finally, we show the resulting new variant of stochastic gradient manipulation converges to Pareto optimal or critical solutions and yield comparable or improved empirical performance.

Adapting deep neural networks to the changing environments is critical in practical utility, especially for online web applications, where the data distribution changes gradually due to the evolving environments. For instance, the web photos of cellphones change gradually over years due to appearance changes. This paper deals with such a problem via active gradual domain adaptation, where the learner continually and actively selects the most informative labels from the target to enhance labeling efficiency and utilizes both labeled and unlabeled samples to improve the model adaptation under gradual domain drift. We propose the active gradual self-training (AGST) algorithm with novel designs of active pseudolabeling and gradual semi-supervised domain adaptation. Specifically, AGST pseudolabels the samples with high confidence, and selects the most informative labels from the unconfident samples based on both uncertainty and diversity, and then gradually self-trains itself by confident pseudolabels and queried labels. To study the gradual domain shift problem in the web data and verify the proposed algorithm, we create a new dataset -- Evolving-Image-Search (EVIS), collected from the web search engine and covers a 12-years range. Since the appearance of the products evolves over these years, such dataset naturally contains gradual domain drift. We extensively evaluate AGST on the synthetic dataset, real-world dataset, and EVIS dataset. AGST achieves up to 62% accuracy improvement (absolute value) against unsupervised gradual self-training with only 5% additional labels, and 19% accuracy improvement against directly applying CLUE, demonstrating the effectiveness of the designs of active pseudolabel and gradual semi-supervised domain adaptation.

Graph Convolutional Networks (GCNs) are promising deep learning approaches in learning representations for graph-structured data. Despite the proliferation of such methods, it is well known that they are vulnerable to carefully crafted adversarial attacks on the graph structure. In this paper, we first conduct an adversarial vulnerability analysis based on matrix perturbation theory. We prove that the low- frequency components of the symmetric normalized Laplacian, which is usually used as the convolutional filter in GCNs, could be more robust against structural perturbations when their eigenvalues fall into a certain robust interval. Our results indicate that not all low-frequency components are robust to adversarial attacks and provide a deeper understanding of the relationship between graph spectrum and robustness of GCNs. Motivated by the theory, we present GCN-LFR, a general robust co-training paradigm for GCN-based models, that encourages transferring the robustness of low-frequency components with an auxiliary neural network. To this end, GCN-LFR could enhance the robustness of various kinds of GCN-based models against poisoning structural attacks in a plug-and-play manner. Extensive experiments across five benchmark datasets and five GCN-based models also confirm that GCN-LFR is resistant to the adversarial attacks without compromising on performance in the benign situation.