Modern power systems are becoming increasingly dynamic, with changing topologies and time-varying loads driven by renewable energy variability, electric vehicle adoption, and active grid reconfiguration. Despite these changes, publicly available test cases remain scarce, due to security concerns and the significant effort required to anonymize real systems. Such limitations call for generative tools that can jointly synthesize grid structure and nodal dynamics. However, modeling the joint distribution of network topology, branch attributes, bus properties, and dynamic load profiles remains a major challenge, while preserving physical feasibility and avoiding prohibitive computational costs. We present PowerGrow, a co-generative framework that significantly reduces computational overhead while maintaining operational validity. The core idea is dependence decomposition: the complex joint distribution is factorized into a chain of conditional distributions over feasible grid topologies, time-series bus loads, and other system attributes, leveraging their mutual dependencies. By constraining the generation process at each stage, we implement a hierarchical graph beta-diffusion process for structural synthesis, paired with a temporal autoencoder that embeds time-series data into a compact latent space, improving both training stability and sample fidelity. Experiments across benchmark settings show that PowerGrow not only outperforms prior diffusion models in fidelity and diversity but also achieves a 98.9% power flow convergence rate and improved N-1 contingency resilience. This demonstrates its ability to generate operationally valid and realistic power grid scenarios.

Optimizing recommender systems for objectives beyond accuracy, such as diversity, novelty, and personalization, is crucial for long-term user satisfaction. To this end, industrial practitioners have accumulated vast amounts of structured domain knowledge, which we term human priors (e.g., item taxonomies, temporal patterns). This knowledge is typically applied through post-hoc adjustments during ranking or post-ranking. However, this approach remains decoupled from the core model learning, which is particularly undesirable as the industry shifts to end-to-end generative recommendation foundation models. On the other hand, many methods targeting these beyond-accuracy objectives often require architecture-specific modifications and discard these valuable human priors by learning user intent in a fully unsupervised manner. Instead of discarding the human priors accumulated over years of practice, we introduce a backbone-agnostic framework that seamlessly integrates these human priors directly into the end-to-end training of generative recommenders. With lightweight, prior-conditioned adapter heads inspired by efficient LLM decoding strategies, our approach guides the model to disentangle user intent along human-understandable axes (e.g., interaction types, long- vs. short-term interests). We also introduce a hierarchical composition strategy for modeling complex interactions across different prior types. Extensive experiments on three large-scale datasets demonstrate that our method significantly enhances both accuracy and beyond-accuracy objectives. We also show that human priors allow the backbone model to more effectively leverage longer context lengths and larger model sizes.

Many modern recommender systems represent user and item attributes as embedding vectors, relying on them for accurate recommendations. However, entangled embeddings often capture not only intrinsic property factors (e.g., user interest in item property) but also popularity factors (e.g., user conformity to item popularity) indistinguishably. These embeddings, influenced by popularity distribution, may face challenges when the popularity distribution at test time differs from historical distribution. Existing remedies in the literature involve disentangled embedding learning, which aims to separately capture intrinsic and popularity factors, demonstrating plausible generalization during popularity distribution shifts. However, we highlight that these methods often overlook a crucial aspect of popularity shifts-their temporal nature-in both training and inference phases. To address this, we propose Temporal Popularity distribution shift generalizABle recommender system (TPAB), a novel disentanglement framework incorporating temporal popularity. TPAB introduce a new (1) temporal-aware embedding design for users and items. Within this design, (2) popularity coarsening and (3) popularity bootstrapping are proposed to enhance generalization further. We also provide theoretical analysis showing that the bootstrapping loss eliminates the effect of popularity on the learned model. During inference, we infer test-time popularity and corresponding embeddings, using them alongside property embeddings for prediction. Extensive experiments on real-world datasets validate TPAB, showcasing its outstanding generalization ability during temporal popularity distribution shifts.

In the era of big data and AI, recommender systems must adapt to evolving user preferences and new users/items to maintain high-quality recommendations. Fine-tuning, which updates model parameters using only new data, offers an efficient alternative to full retraining but struggles to balance stability (retaining past knowledge) and plasticity (adapting to new knowledge). While existing methods prioritize stability to address catastrophic forgetting, we argue that plasticity must also be explicitly strengthened, especially for users with rapidly changing preferences. In this work, we propose PlastIcity and StAbility balancing continual recommender systems (PISA), a novel framework that adaptively balances stability and plasticity based on user preference shifts. PISA quantifies preference shifts as changes in user distances to item clusters, and then guides user embeddings by prioritizing stability for stable users and plasticity for dynamic users. To achieve this, PISA leverages backward knowledge from the previous model and forward knowledge from fine-tuning on current data. During training, PISA maximizes mutual information between user-specific parameters and the relevant reference knowledge. Theoretically, we show that enhancing plasticity mitigates distribution shifts more effectively than fine-tuning alone. Empirically, extensive experiments on three real-world datasets validate PISA's superiority over existing methods and highlight the contributions of its components.

Algorithmic fairness has been receiving increasing attention in recent years. Among others, individual fairness, with its root in the dictionary definition of fairness, offers a fine-grained fairness notion. At the algorithmic level, individual fairness can often be operationalized as a convex regularization term with respect to a similarity matrix. Appealing as it might be, a notorious challenge of individual fairness lies in how to find appropriate distance or similarity measure, which largely remains open to date. Consequently, the similarity or distance measure used in almost any individually fair algorithm is likely to be imperfect due to various reasons such as imprecise prior/domain knowledge, noise, or even adversaries. In this paper, we take an important step towards resolving this fundamental challenge and ask: how sensitive is the individually fair learning algorithm with respect to the given similarities? How can we make the learning results robust with respect to the imperfection of the given similarity measure? First (Soul-M), we develop a sensitivity measure to characterize how the learning outcomes of an individually fair learning algorithm change in response to the change of the given similarity measure. Second (Soul-A ), based on the proposed sensitive measure, we further develop a robust individually fair algorithm by adversarial learning that optimizes the similarity matrix to defend against L_∞ attack. A unique advantage of our sensitivity measure and robust algorithm lies in that they are applicable to a broad range of learning models as long as the objective function is twice differentiable. We conduct extensive experiments to demonstrate the efficacy of our methods.

Data collected in the real world often encapsulates historical discrimination against disadvantaged groups and individuals. Existing fair machine learning (FairML) research has predominantly focused on mitigating discriminative bias in the model prediction, with far less effort dedicated towards exploring how to trace biases present in the data, despite its importance for the transparency and interpretability of FairML. To fill this gap, we investigate a novel research problem: discovering samples that reflect biases/prejudices from the training data. Grounding on the existing fairness notions, we lay out a sample bias criterion and propose practical algorithms for measuring and countering sample bias. The derived bias score provides intuitive sample-level attribution and explanation of historical bias in data. On this basis, we further design two FairML strategies via sample-bias-informed minimal data editing. They can mitigate both group and individual unfairness at the cost of minimal or zero predictive utility loss. Extensive experiments and analyses on multiple real-world datasets demonstrate the effectiveness of our methods in explaining and mitigating unfairness. Code is available at https://github.com/ZhiningLiu1998/AIM.

User-side group fairness is crucial for modern recommender systems, alleviating performance disparities among user groups defined by sensitive attributes like gender, race, or age. In the everevolving landscape of user-item interactions, continual adaptation to newly collected data is crucial for recommender systems to stay aligned with the latest user preferences. However, we observe that such continual adaptation often worsen performance disparities. This necessitates a thorough investigation into user-side fairness in dynamic recommender systems. This problem is challenging due to distribution shifts, frequent model updates, and nondifferentiability of ranking metrics. To our knowledge, this paper presents the first principled study on ensuring user-side fairness in dynamic recommender systems. We start with theoretical analyses on fine-tuning v.s. retraining, showing that the best practice is incremental fine-tuning with restart. Guided by our theoretical analyses, we propose FAir Dynamic rEcommender (FADE), an end-to-end fine-tuning framework to dynamically ensure user-side fairness over time. To overcome the non-differentiability of recommendation metrics in the fairness loss, we further introduce Differentiable Hit (DH) as an improvement over the recent NeuralNDCG method, not only alleviating its gradient vanishing issue but also achieving higher efficiency. Besides that, we also address the instability issue of the fairness loss by leveraging the competing nature between the recommendation loss and the fairness loss. Through extensive experiments on real-world datasets, we demonstrate that FADE effectively and efficiently reduces performance disparities with little sacrifice in the overall recommendation performance.

Given an evolving tensor time series and multiple time ranges, how can we compute Tucker decomposition for each time range efficiently and accurately? Tucker decomposition has been widely used in a variety of applications to obtain latent factors of tensor data. For example, Tucker decomposition on air pollution data allows us to analyze and compare air pollution patterns between different locations during different periods of time. In these applications, a common need is to compute Tucker decomposition for a given time range. Furthermore, real-world tensor time series are typically evolving in the time dimension. Such needs call for a data structure that can efficiently and accurately support range queries of Tucker decomposition and stream updates. Unfortunately, existing methods do not support either range queries or stream updates. For methods that do not support range queries, they have to re-compute from scratch for each query. Not until 2021 has a data structure called Zoom-Tucker been proposed to support range queries via block-wise preprocessing. However, Zoom-Tucker does not support stream updates and, more critically, suffers from a reluctant efficiency-accuracy tradeoff --- a large block size causes inaccuracy, while a small block size leads to inefficiency. This challenging problem has remained open for years prior to our work. To solve this challenging problem, we propose TUCKET, a data structure that can efficiently and accurately handle both range queries and stream updates. Our key idea is to design a new data structure that we call a stream segment tree by generalizing the segment tree , a data structure that was originally invented for computational geometry. For a range query of length L , our TUCKET can find O (log L ) nodes (called the hit set ) from the tree and efficiently stitch their preprocessed decompositions to answer the range query. We also propose an algorithm to optimally prune the hit set via an approximation of subtensor decomposition. For the T -th stream update, our TUCKET modifies only amortized O (1) nodes and only O (log T ) nodes in the worst case. Extensive evaluation demonstrates that our TUCKET consistently achieves the highest efficiency and accuracy across four large-scale datasets. Our TUCKET achieves at least 3 times lower latency and at least 1.4 times smaller reconstruction error than Zoom-Tucker on all datasets. The full version can be found at https://github.com/q-rz/TUCKET/blob/main/TUCKET-Full.pdf.

Diffusion on graphs is ubiquitous with numerous high-impact applications, ranging from the study of residential segregation in socioeconomics and activation cascading in neuroscience, to the modeling of disease contagion in epidemiology and malware spreading in cybersecurity. In these applications, complete diffusion histories play an essential role in terms of identifying dynamical patterns, reflecting on precaution actions, and forecasting intervention effects. Despite their importance, complete diffusion histories are rarely available and are highly challenging to reconstruct due to ill-posedness, explosive search space, and scarcity of training data. To date, few methods exist for diffusion history reconstruction. They are exclusively based on the maximum likelihood estimation (MLE) formulation and require to know true diffusion parameters. In this paper, we study an even harder problem, namely reconstructing Diffusion history from A single SnapsHot (DASH), where we seek to reconstruct the history from only the final snapshot without knowing true diffusion parameters. We start with theoretical analyses that reveal a fundamental limitation of the MLE formulation. We prove: (a) estimation error of diffusion parameters is unavoidable due to NP-hardness of diffusion parameter estimation, and (b) the MLE formulation is sensitive to estimation error of diffusion parameters. To overcome the inherent limitation of the MLE formulation, we propose a novel barycenter formulation: finding the barycenter of the posterior distribution of histories, which is provably stable against the estimation error of diffusion parameters. We further develop an effective solver named DIffusion hiTting Times with Optimal proposal (DITTO) by reducing the problem to estimating posterior expected hitting times via the Metropolis-Hastings Markov chain Monte Carlo method (M-H MCMC) and employing an unsupervised graph neural network to learn an optimal proposal to accelerate the convergence of M-H MCMC. We conduct extensive experiments to demonstrate the efficacy of the proposed method. Our code is available at https://github.com/q-rz/KDD23-DITTO. The appendix can be found at https://arxiv.org/abs/2306.00488.

Multivariate time series (MTS) imputation is a widely studied problem in recent years. Existing methods can be divided into two main groups, including (1) deep recurrent or generative models that primarily focus on time series features, and (2) graph neural networks (GNNs) based models that utilize the topological information from the inherent graph structure of MTS as relational inductive bias for imputation. Nevertheless, these methods either neglect topological information or assume the graph structure is fixed and accurately known. Thus, they fail to fully utilize the graph dynamics for precise imputation in more challenging MTS data such as networked time series (NTS), where the underlying graph is constantly changing and might have missing edges. In this paper, we propose a novel approach to overcome these limitations. First, we define the problem of imputation over NTS which contains missing values in both node time series features and graph structures. Then, we design a new model named PoGeVon which leverages variational autoencoder (VAE) to predict missing values over both node time series features and graph structures. In particular, we propose a new node position embedding based on random walk with restart (RWR) in the encoder with provable higher expressive power compared with message-passing based graph neural networks (GNNs). We further design a decoder with 3-stage predictions from the perspective of multi-task learning to impute missing values in both time series and graph structures reciprocally. Experiment results demonstrate the effectiveness of our model over baselines.