MOTIVATION: The rapid accumulation of multi-omics data presents a valuable opportunity to advance our understanding of complex diseases and biological systems, driving the development of integrative computational methods. However, the complexity of biological processes, spanning multiple molecular layers and involving intricate regulatory interactions, requires models that can capture both intra- and cross-omics relationships. Most existing integration methods primarily focus on sample-level similarities or intra-omics feature interactions, often neglecting the interactions across different omics layers. This limitation can result in the loss of critical biological information and suboptimal performance. To address this gap, we propose X-intNMF, a network-regularized non-negative matrix factorization (NMF) framework that simultaneously integrates intra- and cross-omics feature interactions into a shared low-dimensional representation (see Fig. 1). By modeling these multi-layered relationships, X-intNMF enhances the representation of biological interactions and improves integration quality and prediction accuracy. RESULTS: For evaluation, we applied X-intNMF to predict breast cancer phenotypes and classify clinical outcomes in lung and ovarian cancers using mRNA expression, microRNA expression, and DNA methylation data from TCGA. The results show that X-intNMF consistently outperforms state-of-the-art methods. Ablation studies confirm that incorporating both cross-omics and intra-omics interactions contributes significantly to the model's improved performance. Additionally, survival analysis on 25 TCGA cancer datasets demonstrates that the integrated multi-omics representation offers strong prognostic value for both overall survival and disease-free status. These findings highlight X-intNMF's ability to effectively model multi-layered molecular interactions while maintaining interpretability, robustness, and scalability within the NMF framework. AVAILABILITY AND IMPLEMENTATION: The source code and datasets supporting this study are publicly available at GitHub (https://github.com/compbiolabucf/X-intNMF) and archived on Zenodo (https://doi.org/10.5281/zenodo.18238385).

Modern data-driven applications generate vast, complex data that often contain irrelevant, redundant, or noisy features. The task involves selecting the optimal subset of features from a dataset by removing redundant and irrelevant ones to enhance downstream performance. Using reinforcement learning (RL) for feature selection is ideal for data-centric tasks due to its interactive nature, which dynamically adapts to evolving data environments. However, high initial exploration variability, reliance on downstream tasks, and the need to control the false discovery rate bring challenges. To address these issues, we introduce a knockoff-guided RL framework, which uses pseudo-features to control early-stage randomness and matrix reconstruction for unsupervised reward. Our approach is a single-agent RL approach with unsupervised rewards that utilize knockoff information and matrix reconstruction to enhance feature selection without relying on labeled data. The framework integrates three components: the knockoff information to control exploration variability, decision network pre-training to guide RL policy, and matrix reconstruction to guide unsupervised rewards. Extensive experiments are conducted on various datasets across different task types, including classification and regression, demonstrating the superiority of our framework. Codes are available here.

Utilizing offline reinforcement learning (RL) with real-world clinical data is getting increasing attention in AI for healthcare. However, implementation poses significant challenges. Defining direct rewards is difficult, and inverse RL struggles to infer accurate reward functions from expert behavior in complex environments. Offline RL also encounters distributional discrepancies between learned policies and observed human behavior, a critical issue in healthcare applications. To address challenges in applying offline RL to physical activity promotion for older adults at high risk of falls, based on wearable sensor activity monitoring, we introduce Kolmogorov-Arnold Networks and Diffusion Policies for Offline Inverse Reinforcement Learning (KANDI). Specifically, by leveraging the flexible function approximation in Kolmogorov-Arnold Networks, we estimate the reward function by learning free-living environment behavior from low-fall-risk older adults (experts). Additionally, diffusion-based policies within an Actor-Critic framework provide a generative approach for action refinement, enabling controlled exploration and mitigating distributional shift issues in offline RL. We evaluate KANDI using wearable activity monitoring data in a two-arm clinical trial from our Physio-feedback Exercise Program (PEER) study, emphasizing its practical application in a fall-risk intervention program to promote physical activity among older adults. Our analysis identifies the optimal timing for anti-sedentariness interventions tailored to varying levels of fall risk, thereby maximizing daily physical activity. Additionally, we evaluate KANDI on the D4RL benchmark, outperforming the state-of-the-art methods in each domain. These results underscore KANDI's potential to address key challenges in offline RL for healthcare applications, offering an effective solution for the optimal timing and policy for activity promotion intervention strategies.

Due to the non-stationarity of time series, the distribution shift problem largely hinders the performance of time series forecasting. Existing solutions either rely on using certain statistics to specify the shift, or developing specific mechanisms for certain network architectures. However, the former would fail for the unknown shift beyond simple statistics, while the latter has limited compatibility on different forecasting models. To overcome these problems, we first propose a decoupled formulation for time series forecasting, with no reliance on fixed statistics and no restriction on forecasting architectures. This formulation regards the removing-shift procedure as a special transformation between a raw distribution and a desired target distribution and separates it from the forecasting. Such a formulation is further formalized into a bi-level optimization problem, to enable the joint learning of the transformation (outer loop) and forecasting (inner loop). Moreover, the special requirements of expressiveness and bi-direction for the transformation motivate us to propose instance normalization flow (IN-Flow), a novel invertible network for time series transformation. Different from the classic ''normalizing flow'' models, IN-Flow does not aim for normalizing input to the prior distribution (e.g., Gaussian distribution) for generation, but creatively transforms time series distribution by stacking normalization layers and flow-based invertible networks, which is thus named ''normalization'' flow. Finally, we have conducted extensive experiments on both synthetic data and real-world data, which demonstrate the superiority of our method.

Feature selection removes redundant features to enhance both performance and computational efficiency in downstream tasks. Existing methods often struggle to capture complex feature interactions and adapt to diverse scenarios. Recent advances in this domain have incorporated generative intelligence to address these drawbacks by uncovering intricate relationships between features. However, two key limitations remain: 1) embedding feature subsets in a continuous space is challenging due to permutation sensitivity, as changes in feature order can introduce biases and weaken the embedding learning process; 2) gradient-based search in the embedding space assumes convexity, which is rarely guaranteed, leading to reduced search effectiveness and suboptimal subsets. To address these limitations, we propose a new framework that can: 1) preserve feature subset knowledge in a continuous embedding space while ensuring permutation invariance; 2) effectively explore the embedding space without relying on strong convex assumptions. For the first objective, we develop an encoder-decoder paradigm to preserve feature selection knowledge into a continuous embedding space. This paradigm captures feature interactions through pairwise relationships within the subset, removing the influence of feature order on the embedding. Moreover, an inducing point mechanism is introduced to accelerate pairwise relationship computations. For the second objective, we employ a policy-based reinforcement learning (RL) approach to guide the exploration of the embedding space. The RL agent effectively navigates the space by balancing multiple objectives. By prioritizing high-potential regions adaptively and eliminating the reliance on convexity assumptions, this search strategy effectively reduces the risk of converging to local optima. Finally, we conduct extensive experiments to demonstrate the effectiveness, efficiency, robustness and explicitness of our model. Our code and dataset are publicly accessible on GitHub. https://github.com/RayLiu1103/CAPS.

We propose a novel seismic signal processing approach to efficiently and effectively attenuate seismic random noises. The proposed approach is a generalized seismic noise attenuation solution that can be applied to typical denoising operators. Our work has two main contributions. First, conventional filtering operators “regularize” the denoised results through the operator design. However, as seismic data have strong nonstationarity, it is inevitable to remove certain signal components. The resampling mechanism alleviates the signal loss. Second, the resampling operation does not require a lot of parameter tuning, which improves the denoising efficiency. Using the proposed approach, compared with existing denoising operators, the intrinsic seismic signal components are better recovered since random noise has been suppressed. Synthetic example and field data applications quantitatively and qualitatively demonstrate excellent performances of the proposed approach.

The fairness issue is very important in deploying machine learning models as algorithms widely used in human society can be easily in discrimination. Researchers have studied disparity on tabular data a lot and proposed many methods to relieve bias. However, studies towards unfairness in graph are still at early stage while graph data that often represent connections among people in real-world applications can easily give rise to fairness issues and thus should be attached to great importance. Fair representation learning is one of the most effective methods to relieve bias, which aims to generate hidden representations of input data while obfuscating sensitive information. In graph setting, learning fair representations of graph (also called fair graph embeddings) is effective to solve graph unfairness problems. However, most existing works of fair graph embeddings only study fairness in a coarse granularity (i.e., group fairness), but overlook individual fairness. In this paper, we study fair graph representations from different levels. Specifically, we consider both group fairness and individual fairness on graph. To debias graph embeddings, we propose FairGAE, a fair graph auto-encoder model, to derive unbiased graph embeddings based on the tailor-designed fair Graph Convolution Network (GCN) layers. Then, to achieve multi-level fairness, we design a Wasserstein distance based regularizer to learn the optimal transport for fairer embeddings. To overcome the efficiency concern, we further bring up Sinkhorn divergence as the approximations of Wasserstein cost for computation. Finally, we apply the learned unbiased embeddings into the node classification task and conduct extensive experiments on two real-world graph datasets to demonstrate the improved performances of our approach.

Internet of Things (IoT) enables extensive connections between cyber and physical "things". Nevertheless, the streaming data among IoT sensors bring "big data" issues, for example, large data volumes, data redundancy, lack of scalability and so on. Under "big data" circumstances, IoT system monitoring becomes a challenge. Furthermore, cyberattacks which threaten IoT security are hard to be detected. In this paper, we propose an online distributed IoT security monitoring algorithm (ODIS). An advanced influential point selection operation extracts important information from multidimensional time series data across distributed sensor nodes based on the spatial and temporal data dependence structure. Then, an accurate data structure model is constructed to capture the IoT system behaviors. Next, hypothesis testing is carried out to quantify the uncertainty of the monitoring tasks. Besides, the distributed system architecture solves the scalability issue. Using a real sensor network testbed, we commit cyberattacks to an IoT system with different patterns and strengths. The proposed ODIS algorithm demonstrates promising detection and monitoring performances.

Estimating the dependence structure of multidimensional time series data in real-time is challenging. With large volumes of streaming data, the problem becomes more difficult when the multidimensional data are collected asynchronously across distributed nodes, which motivates us to sample representative data points from streams. We propose a leverage score sampling (LSS) method for efficient online inference of the streaming vector autoregressive (VAR) model. We define the leverage score for the streaming VAR model so that the LSS method selects informative data points in real-time with statistical guarantees of parameter estimation efficiency. Moreover, our LSS method can be directly deployed in an asynchronous decentralized environment, e.g., a sensor network without a fusion center, and produce asynchronous consensus online parameter estimation over time. By exploiting the temporal dependence structure of the VAR model, the LSS method selects samples independently on each dimension and thus is able to update the estimation asynchronously. We illustrate the effectiveness of the LSS method in synthetic, gas sensor and seismic datasets.

Seismic reflectivity inversion is widely applied to improve the seismic resolution to obtain detailed underground understandings. Based on the convolution model, seismic inversion removes the wavelet effect by solving an optimization problem. Taking advantage of the sparsity property, the ℓ1-norm is commonly adopted in the regularization terms to overcome the noise/interference vulnerability observed in the lp-losses minimization. However, no one has provided a deterministic conclusion that ℓ1-norm regularization is the best choice for seismic reflectivity inversion. Instead of using an unproved fixed regularization norm, we propose an optimal seismic reflectivity inversion approach. Our method adaptively adopts an ℓp-loss-ℓq-regularization (i.e., ℓp,q-regularization) for p = 2, 0q-norm regularization. Then, the majorization-minimization and CV algorithms are briefly described. The performance of the proposed seismic inversion approach is evaluated through synthetic examples and a field example from the Bohai Bay Basin, China.