Time series data is now pervasive across domains such as healthcare, finance, entertainment, and transportation, driven by advances in sensing technologies that enable continuous data collection. The resulting increase in data volume and complexity poses significant challenges to traditional analysis methods, calling for the development of advanced, interdisciplinary approaches to temporal data mining. This workshop aims to: (1) identify key challenges in learning from time series data, including irregular sampling, spatiotemporal dependencies, and uncertainty quantification; (2) explore recent advances in algorithmic, statistical, theoretical, and systems-based solutions-ranging from classical methods to emerging techniques involving large language models (LLMs); and (3) foster collaboration by highlighting open problems and novel research directions in time series analysis. Bridging theory and practice, the workshop provides a platform for researchers and practitioners from academia, industry, and government to exchange ideas, discuss technical challenges, and showcase practical applications. Contributions from related areas such as AI, machine learning, data science, and statistics are strongly encouraged.

Audio spoofing has surged with the rise of generative artificial intelligence, posing a serious threat to online communication. Recent studies have shown promising avenues in detecting spoofed audio specifically those that use human expert knowledge in representation learning, but more work is needed to evaluate performance across various realistic scenarios that tend to pose challenges in spoofed audio detection. In this paper, we introduce a comprehensive framework for expert-in-the-loop representation learning for audio anti-spoofing that is robust enough to address four specific challenging scenarios. Multimodal, Multilingual, Multi-speaker, and Multi-attack (4M). Preliminary results demonstrate the framework’s potential effectiveness in audio anti-spoofing.

Time series data has become ubiquitous across various fields such as healthcare, finance, entertainment, and transportation, driven by advancements in sensing technologies that enable continuous monitoring and recording. This growth in data size and complexity presents new challenges for traditional analysis techniques, necessitating the development of advanced, interdisciplinary temporal mining algorithms. The goals of this workshop are to: (1) highlight significant challenges in learning and mining from time series data, such as irregular sampling, spatiotemporal structures, and uncertainty quantification; (2) discuss recent developments in algorithmic, theoretical, statistical, and systems-based approaches for addressing these challenges, including both classical methods and large language models (LLMs); and (3) synergize research efforts by exploring both new and open problems in time series analysis and mining. This workshop will focus on both the theoretical and practical aspects of time series data analysis, providing a platform for researchers and practitioners from academia, government, and industry to discuss potential research directions, critical technical issues, and present solutions for practical applications. Contributions from related fields such as AI, machine learning, data science, and statistics are also included.

Collocation of measurements from active and passive satellite sensors refers to the combination of data from two sensors that observe the same geographic area at nearly the same time but with differing spatial resolutions and viewing angles. This collocated data, often known as on-track data, comes with precise product labels from the active sensor but comprises only the pixels located directly on the path of an active satellite’s orbit. As a result, its spatial coverage is quite limited, especially when compared to the vast quantities of off-track data. Handling the abundant and information-dense off-track data is crucial for training machine learning models that can effectively integrate the unique features of this data along with on-track data. However, the sheer volume of off-track data presents significant challenges for these models. To address the challenges of large amounts of unlabeled off-track data in remote sensing applications, we introduce a self-supervised representation learning model with VAE and domain adaptation methods to learn a domain invariant classifier for the on-track and off-track data. The model’s performance is enhanced by pre-training off-track data with VAE generative model using off-track data, to learn a good representation that can be transferred to the down-streaming domain adaptation and classification tasks. The classifier is built on these representations to classify different cloud types in passive sensing data, with the goal of achieving higher accuracy in cloud property retrieval. Extensive quantitative and qualitative evaluation demonstrate our method achieves higher accuracy in cloud property retrieval for off-track remote sensing data.

Conducting survival analysis on distributed healthcare data is an important research problem, as privacy laws and emerging data-sharing regulations prohibit the sharing of sensitive patient data across multiple institutions. The distributed healthcare survival data often exhibit heterogeneity, non-uniform censoring and involve patients with multiple health conditions (competing risks), which can result in biased and unreliable risk predictions. To address these challenges, we propose employing federated learning (FL) for survival analysis with competing risks. In this work, we present two main contributions. Firstly, we propose a simple algorithm for estimating consistent federated pseudo values (FPV) for survival analysis with competing risks and censoring. Secondly, we introduce a novel and flexible FPV-based deep learning framework named Fedora, which jointly trains our proposed transformer-based model, TransPseudo, specific to the participating institutions (clients) within the Fedora framework without accessing clients' data, thus, preserving data privacy. We conducted extensive experiments on both real-world distributed healthcare datasets characterized by non-IID and non-uniform censoring properties, as well as synthetic data with various censoring settings. Our results demonstrate that our Fedora framework with the TransPseudo model performs better than the federated learning frameworks employing state-of-the-art survival models for competing risk analysis.

Time series data has become pervasive across domains such as finance, transportation, retail, entertainment, and healthcare. This shift towards continuous monitoring and recording, fueled by advancements in sensing technologies, necessitates the development of new tools and solutions. Despite extensive study, the importance of time series analysis continues to increase. However, modern time series data present challenges to existing techniques, including irregular sampling and spatiotemporal structures. Time series mining research is both challenging and rewarding as it connects diverse disciplines and requires interdisciplinary solutions. The goals of this workshop are to (1) highlight the significant challenges that underpin learning and mining from time series data (e.g., irregular sampling, spatiotemporal structure, uncertainty quantification), (2) discuss recent algorithmic, theoretical, statistical, or systems-based developments for tackling these problems, and (3) to synergize the research activities and discuss both new and open problems in time series analysis and mining. In summary, our workshop will focus on both the theoretical and practical aspects of time series data analysis and will provide a platform for researchers and practitioners from academia and industry to discuss potential research directions and critical technical issues and present solutions to tackle related issues in practical applications. We will invite researchers and practitioners from the related areas of AI, machine learning, data science, statistics, and many others to contribute to this workshop.

Survival analysis, aka time-to-event analysis, has a wide-ranging impact on patient care. Federated Survival Analysis (FSA) is an emerging Federated Learning (FL) paradigm for performing survival analysis on distributed decentralized data available at multiple medical institutions. FSA enables individual medical institutions, referred to as clients, to improve their survival predictions while ensuring privacy. However, FSA faces challenges due to non-linear and non-IID data distributions among clients, as well as bias caused by censoring. Although recent studies have adapted Cox Proportional Hazards (CoxPH) survival models for FSA, a systematic exploration of these challenges is currently lacking. In this paper, we address these critical challenges by introducing FedPseudo, a pseudo value-based deep learning framework for FSA. FedPseudo uses deep learning models to learn robust representations from non-linear survival data, leverages the power of pseudo values to handle non-uniform censoring, and employs FL algorithms such as FedAvg to learn model parameters. We propose a novel and simple approach for estimating pseudo values for FSA. We provide theoretical proof that the estimated pseudo values, referred to as Federated Pseudo Values, are consistent. Moreover, our empirical results demonstrate that they can be computed faster than traditional methods of deriving pseudo values. To ensure and enhance the privacy of both the estimated pseudo values and the shared model parameters, we systematically investigate the application of differential privacy (DP) on both the federated pseudo values and local model updates. Furthermore, we adapt V -Usable Information metric to quantify the informativeness of a client's data for training a survival model and utilize this metric to show the advantages of participating in FSA. We conducted extensive experiments on synthetic and real-world survival datasets to demonstrate that our FedPseudo framework achieves better performance than other FSA approaches and performs similarly to the best centrally trained deep survival model. Moreover, FedPseudo consistently achieves superior results across different censoring settings.

Multi-state survival analysis (MSA) uses multi-state models for analyzing time-to-event data collected from subjects who may transition to different states before experiencing the final event of interest over time. A key challenge in MSA is the accurate subject-specific prediction of multi-state model quantities such as transition probability and state occupation probability in the presence of censoring. Censoring is another crucial challenge in MSA, leading to the overestimation of multi-state model quantities. The traditional statistical multi-state models typically do not use covariates, which renders them infeasible for making subject-specific predictions. Moreover, they assume a strict Markov stochastic process while modeling transition probabilities along with proportional hazard or linear covariate effect assumptions - that may not hold in real-world data. The current MSA methods have not investigated the impact of different types of censoring on the multi-state model quantities estimation. Recently proposed state-of-the-art neural ordinary differential equation (ODE) models for MSA relax statistical assumptions, but they do not handle the censoring mechanism well. To fill the gap in the MSA literature, we propose a new class of pseudo-value based deep learning models for MSA, where we show that pseudo values - designed to handle censoring - can be a natural replacement for estimating the subject-specific multi-state model quantities when derived from Aalen-Johansen (AJ) or Landmark AJ consistent estimators. We systematically study our proposed models’ performance under different censoring settings and when Markovianity or linearity assumptions get violated. Empirical results on both the simulated and real-world MSA datasets show that our proposed models perform better or comparably to existing MSA methods under various censoring settings.

Atmospheric gravity waves are produced when gravity attempts to restore disturbances through stable layers in the atmosphere. They have a visible effect on many atmospheric phenomena such as global circulation and air turbulence. Despite their importance, however, little research has been conducted on how to detect gravity waves using machine learning algorithms. We faced two major challenges in our research: our raw data had a lot of noise and the labeled dataset was extremely small. In this study, we explored various methods of preprocessing and transfer learning in order to address those challenges. We pre-trained an autoencoder on unlabeled data before training it to classify labeled data. We also created a custom CNN by combining certain pre-trained layers from the InceptionV3 Model trained on ImageNet with custom layers and a custom learning rate scheduler. Experiments show that our best model outperformed the best performing baseline model by 6.36% in terms of test accuracy.

Domain adaptation techniques using deep neural networks have been mainly used to solve the distribution shift problem in homogeneous domains where data usually share similar feature spaces and have the same dimensionalities. Nevertheless, real world applications often deal with heterogeneous domains that come from completely different feature spaces with different dimensionalities. In our remote sensing application, two remote sensing datasets collected by an active sensor and a passive one are heterogeneous. In particular, CALIOP actively measures each atmospheric column. In this study, 25 measured variables/features that are sensitive to cloud phase are used and they are fully labeled. VIIRS is an imaging radiometer, which collects radiometric measurements of the surface and atmosphere in the visible and infrared bands. Recent studies have shown that passive sensors may have difficulties in prediction cloud/aerosol types in complicated atmospheres (e.g., overlapping cloud and aerosol layers, cloud over snow/ice surface, etc.). To overcome the challenge of the cloud property retrieval in passive sensor, we develop a novel VAE based approach to learn domain invariant representation that capture the spatial pattern from multiple satellite remote sensing data (VDAM), to build a domain invariant cloud property retrieval method to accurately classify different cloud types (labels) in the passive sensing dataset. We further exploit the weight based alignment method on the label space to learn a powerful domain adaptation technique that is pertinent to the remote sensing application. Experiments demonstrate our method outperforms other state-of-the-art machine learning methods and achieves higher accuracy in cloud property retrieval in the passive satellite dataset.

Time series data are ubiquitous, and is one of the fastest growing and richest types of data. Recent advances in sensing technologies has resulted in a rapid growth in the size and complexity of time series archives. This demands development of new tools and solutions. The goals of this workshop are to: (1) highlight the significant challenges that underpin learning and mining from time series data (e.g. irregular sampling, spatiotemporal structure, uncertainty quantification), (2) discuss recent algorithmic, theoretical, statistical, or systems-based developments for tackling these problems, and (3) exploring new frontiers in time series analysis and their connections with important topics such as knowledge representation, reasoning, control, and business intelligence. In summary, our workshop will focus on both the theoretical and practical aspects of time series data analysis and will provide a platform for researchers and practitioners from both academia and industry to discuss potential research directions, key technical issues, and present solutions to tackle related issues in practical applications. We will invite researchers and practitioners from the related areas of AI, machine learning, data science, statistics, and many others to contribute to this workshop.

Survival analysis aims to predict the risk of an event, such as death due to cancer, in the presence of censoring. Recent research has shown that existing survival techniques are prone to unintentional biases towards protected attributes such as age, race, and/or gender. For example, censoring assumed to be unrelated to the prognosis and covariates (typically violated in real data) often leads to overestimation and biased survival predictions for different protected groups. In order to attenuate harmful bias and ensure fair survival predictions, we introduce fairness definitions based on survival functions and censoring. We propose novel fair and interpretable survival models which use pseudo valued-based objective functions with fairness definitions as constraints for predicting subject-specific survival probabilities. Experiments on three real-world survival datasets demonstrate that our proposed fair survival models show significant improvement over existing survival techniques in terms of accuracy and fairness measures. We show that our proposed models provide fair predictions for protected attributes under different types and amounts of censoring. Furthermore, we study the interplay between interpretability and fairness; and investigate how fairness and censoring impact survival predictions for different protected attributes.

Time series data are ubiquitous. Rapid advances in diverse sensing technologies, ranging from remote sensors to wearables and social sensing, are generating a rapid growth in the size and complexity of time series archives. This has resulted in a fundamental shift away from parsimonious, infrequent measurement to nearly continuous monitoring and recording. This demands development of new tools and solutions. The goals of this workshop are to: (1) highlight the significant challenges that underpin learning and mining from time series data (e.g. irregular sampling, spatiotemporal structure, and uncertainty quantification), (2) discuss recent algorithmic, theoretical, statistical, or systems-based developments for tackling these problems, and (3) synergize the research activities and discuss both new and open problems in time series analysis and mining.

Domain adaptation techniques have been developed to handle data from multiple sources or domains. Most existing domain adaptation models assume that source and target domains are homogeneous, i.e., they have the same feature space. Nevertheless, many real world applications often deal with data from heterogeneous domains that come from completely different feature spaces. In our remote sensing application, data in source domain (from an active spaceborne Lidar sensor CALIOP onboard CALIPSO satellite) contain 25 attributes, while data in target domain (from a passive spectroradiometer sensor VIIRS onboard Suomi-NPP satellite) contain 20 different attributes. CALIOP has better representation capability and sensitivity to aerosol types and cloud phase, while VIIRS has wide swaths and better spatial coverage but has inherent weakness in differentiating atmospheric objects on different vertical levels. To address this mismatch of features across the domains/sensors, we propose a novel end-to-end deep domain adaptation with domain mapping and correlation alignment (DAMA) to align the heterogeneous source and target domains in active and passive satellite remote sensing data. It can learn domain invariant representation from source and target domains by transferring knowledge across these domains, and achieve additional performance improvement by incorporating weak label information into the model (DAMA-WL). Our experiments on a collocated CALIOP and VIIRS dataset show that DAMA and DAMA-WL can achieve higher classification accuracy in predicting cloud types.

Identifying interpretable discriminative high-order feature interactions given limited training data in high dimensions is challenging in both machine learning and data mining. In this paper, we propose a factorization based sparse learning framework termed FHIM for identifying high-order feature interactions in linear and logistic regression models, and study several optimization methods for solving them. Unlike previous sparse learning methods, our model FHIM recovers both the main effects and the interaction terms accurately without imposing tree-structured hierarchical constraints. Furthermore, we show that FHIM has oracle properties when extended to generalized linear regression models with pairwise interactions. Experiments on simulated data show that FHIM outperforms the state-of-the-art sparse lear-ning techniques. Further experiments on our experimentally generated data from patient blood samples using a novel SOMAmer (Slow Off-rate Modified Aptamer) technology show that, FHIM performs blood-based cancer diagnosis and bio-marker discovery for Renal Cell Carcinoma much better than other competing methods, and it identifies interpretable block-wise high-order gene interactions predictive of cancer stages of samples. A literature survey shows that the interactions identified by FHIM play important roles in cancer development.