Time series data are becoming ubiquitous in numerous real world applications, e.g., IoT devices, healthcare, wearable devices, smart vehicles, financial markets, biological sciences, environmental sciences, etc. Given the availability of assive amounts of data, their complex underlying structures/distributions, together with the high-performance computing platforms, there is a great demand for developing new
theories and algorithms to tackle fundamental challenges(e.g., representation, classification, prediction, causal analysis, etc.) in various types of applications.
The goal of this workshop is to provide a platform for researchers and AI practitioners from both academia and industry to discuss potential research directions, key technical
issues, and present solutions to tackle related challenges in practical applications. The workshop will focus on both the
theoretical and practical aspects of time series data analysis and aims to trigger research innovations in theories, algorithms, and applications. We will invite researchers and AI
practitioners from the related areas of machine learning, data science, statistics, econometrics, and many others to contribute to this workshop.
This workshop encourages submissions of innovative solutions for a broad range of time series analysis problems. Topics of interest include but are not limited to the following:
- Time series forecasting and prediction
- Spatio-temporal forecasting and prediction
- Time series anomaly detection and diagnosis
- Time series change point detection
- Time series classification and clustering
- Time series similarity search
- Time series indexing
- Time series compression
- Time series pattern discovery
- Interpretation and explanation in time series
- Causal inference in time series
- Bias and fairness in time series
- Federated learning and security in time series
- Benchmarks, experimental evaluation, and comparison for time series analysis tasks
- Time series applications in various areas: E-commerce, Cloud computing, Transportation, Fintech, Healthcare, Internet of things, Wireless networks, Predictive maintenance, Energy, and Climate, etc.
Call for Papers
Contact with us: ai4ts.ijcai@gmail.com
Submissions should be 4-7 pages long, excluding references, and follow the AAAI2025 template. Submissions are double-blind and author identity should not be revealed to the reviewers. An optional appendix of arbitrary length is allowed and should be put at the end of the paper (after references).
Accepted papers will be presented as posters during the workshop and listed on the website (non-archival/without proceedings). A small number of accepted papers will be selected to be presented as contributed talks (15-minute oral presentations). We also welcome submissions of unpublished papers, including those submitted/accepted to other venues if that other venue allows.
Submission link: https://cmt3.research.microsoft.com/AI4TS2025
Note: Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.
Key Dates
Workshop Paper Submission Due Date: December 1st, 2024 December 8th, 2024 (23:59pm AoE)
Notification of Paper Acceptance: December 15th, 2024 December 18th, 2024 (23:59pm AoE)
AAAI AI4TS Workshop Date: March 4th, 2025
Location: Pennsylvania Convention Center, Room 121C
Detailed Workshop Schedule:
| Time (EST) | Speaker | Title |
|---|---|---|
| 9:00 am - 9:05 am | Dr. Dongjin Song | Opening Remarks |
| 9:05 am - 9:50 am | Prof. Chang-Tien Lu | Keynote Talk 1: AI-Driven Forecasting: Harnessing Social Media for Event and Epidemic Prediction |
| 9:50 am - 10:35 am | Dr. Qingsong Wen | Keynote Talk 2: AI for Time Series Analysis: from Transformer to LLM and Foundation Models |
| 10:35 am – 11:00 am | Coffee Break | |
| 11:00 am – 11:45 am | Prof. Anindya Roy | Keynote Talk 3: Utility Preserving Privacy Mechanisms for Time Series Data Release |
| 11:45 am - 12:45 pm | Oral Paper Presentations |
Paper 3: Neural-HATS: Neural Hybrid Approach for Time Series Causal Discovery Paper 4: MOAT: Motif-guided Debiasing Framework for Time Series Forecasting Paper 6: StiefelGen: Time Series Data Augmentation Over the Stiefel Manifold Learning |
| 12:45 pm - 2:00 pm | Lunch Break | |
| 2:00 pm - 2: 45 pm | Dr. David Hallac | Keynote Talk 4: Inferring Structure from Multivariate Time Series Sensor Data |
| 2:45 pm - 3: 30 pm | Dr. Zhangyang (Atlas) Wang | Keynote Talk 5: Cracking the Market Code: Building Large Foundation Models for High-Frequency Trading |
| 3:30 pm - 4:00 pm | Coffee Break | |
| 3:30 pm - 5:30 pm | Poster Session |
Speakers
Title: AI-Driven Forecasting: Harnessing Social Media for Event and Epidemic Prediction
Social media data offers a powerful tool for predicting and responding to global challenges. This talk explores AI-driven techniques for forecasting societal events and epidemics. I will introduce an automated system that leverages time-dependent patterns in open-source data from platforms like Twitter, blogs, and news articles to predict societal disruptions such as natural disasters and public demonstrations. Additionally, I will present SimNest, a deep learning framework that integrates computational epidemiology with social media data for real-time outbreak monitoring. Lastly, I will discuss a multi-task learning model that enhances event forecasting across regions, addressing data imbalances and geographic variations. These approaches demonstrate how AI and social media analytics enhance prediction accuracy, support timely interventions, and strengthen crisis preparedness.
Prof. Anindya Roy
Research Staff, United States Census Bureau
Professor, University of Maryland Baltimore County
Baltimore
Title: Utility Preserving Privacy Mechanisms for Time Series Data Release
Private data release mechanisms, particularly those satisfying differential privacy, have undergone evolutionary changes in recent years. However, there are no widely accepted privacy mechanisms for releasing dependent data, such as time series data. We will begin with a short survey of the existing privacy mechanisms for time series and discuss the need to go beyond what is available. Based on a desideratum that balances the nuances of traditional statistical time series analysis and the demands of modern data privacy mechanisms, a formal privacy-utility framework will be presented for time series data. Theoretical and operational characteristics of the proposed framework will be studied and possible future directions will be discussed.
Title: AI for Time Series Analysis: from Transformer to LLM and Foundation Models
Time series analysis is ubiquitous, serving as a cornerstone for extracting valuable insights across a myriad of real-world applications. Recent advancements in Transformer, Large Language Models (LLMs), and Foundation Models (FMs) have fundamentally reshaped the paradigm of model design for time series analysis, significantly boosting performance in various downstream tasks. In this talk, I will first provide an up-to-date overview of this exciting area, highlighting how Transformer, LLMs, and FMs are being leveraged to address the unique challenges of time series data. Then, I will present our recent research on these models to highlight their roles in advancing time series modeling. By consolidating the latest AI advancements for time series analysis, this talk aims to illuminate the transformative potential of these models and identify promising avenues for future research exploration.
Title: Inferring Structure from Multivariate Time Series Sensor Data
Real world systems, ranging from airplanes to data centers, generate large amounts of time series data. In most cases, this data is multivariate and heterogeneous, where the readings come from various types of entities, or sensors. These time series datasets are often sparse, unlabeled, dynamic, and difficult to interpret. Therefore, there is a need for AI methods that learn interpretable structure from such data, especially ones that can apply across many different domains. In this talk, I will discuss several approaches for pattern detection on time series data, how to scale these algorithms to terabyte and petabyte-scale, and applications of this approach to various real-world problems, both in and beyond time series domains, including the work we are currently doing at Viaduct.
Title: Cracking the Market Code: Building Large Foundation Models for High-Frequency Trading
In this talk, I will delve into the exciting research opportunities and unique challenges of building large foundation models for high-frequency trading. Our focus centers on the high-stakes problem of alpha signal extraction from massive streams of order book and transaction data. Key themes include why high-frequency markets remain predictably exploitable despite their complexity, how data-driven approaches are outperforming traditional economic models, and the potential of foundation models trained on semi-structured data such as time series, tables, and graphs. I will share insights on scaling model sizes in extremely noisy data settings, the surprising robustness of learned features compared to hand-crafted ones, and the critical role of strict causality — offering a rare opportunity to measure true “intelligence” progress by avoiding the train-test contamination often seen in modern benchmarks. I aim to shed light on the intersections between academia and industry, highlighting opportunities for cross-disciplinary innovation, collaboration, and the creation of scalable AI systems for high-impact applications.
Accepted Oral Papers
* TS-OOD: An Evaluation Framework for Time-Series Out-of-Distribution Detection and Prospective Directions for Progress Onat Gungor, Amanda Rios, Nilesh Ahuja, Tajana Rosing
* WinTSR: A Windowed Temporal Saliency Rescaling Method for Interpreting Time Series Deep Learning Models MD Khairul Islam, Judy Fox
* Neural-HATS: Neural Hybrid Approach for Time Series Causal Discovery Saima Absar, Wen Huang, Yongkai Wu, Lu Zhang
* MOAT: Motif-guided Debiasing Framework for Time Series Forecasting Li Zhang, Yifeng Gao, Mucun Sun, Shuochao Yao, Ashley Gomez, Jessica Lin
* SST: Multi-Scale Hybrid Mamba-Transformer Experts for Long-Short Range Time Series Forecasting Xiongxiao Xu, Canyu Chen, Yueqing Liang, Baixiang Huang, Guangji Bai, Liang Zhao, Kai Shu
* StiefelGen: Time Series Data Augmentation Over the Stiefel Manifold Prasad Cheema, Mahito Sugiyama
Accepted Posters
* Embedding Periodic Patterns into Tokens:
Masked Autoencoder for Household Electricity Demand Forecasting
Yuugou Ohno, Tomonori Honda, Masaki Onishi, Norihiro Itsubo
* TSKANMixer: Kolmogorov–Arnold Networks with MLP-Mixer Model for Time
Series Forecasting
Young-Chae Hong, Bei Xiao, Yangho Chen
* Retrieval Augmented Time Series Forecasting
Kutay Tire1, Ege Onur Taga, M. Emrullah Ildiz, Samet Oymak
* FrAug: Enhanced Fraud Detection in Interbank Transfers via Augmented Account Features
Seonkyu Lim, Jeongwhan Choi, Jaehoon Lee, Noseong Park
* IMA: An Imputation-based Mixup Augmentation Using Self-Supervised Learning
for Time Series Data
Nha Dang Nguyen, Dang Hai Nguyen, Khoa Tho Anh Nguyen
* Targeted Adversarial Denoising Autoencoders (TADA)
for Neural Time Series Filtration
Benjamin J. Choi, Griffin Milsap, Clara A. Scholl, Francesco Tenore, Mattson Ogg
* Arbitrage energy and wind power forecasting in the presence of diverse actors
Forecasting
Damien Fay, Soumyendu Sarkar
* COVID-19 Prediction with Doubly Multi-task Gaussian Process
Sooyon Kim, Yongtaek Lim, Sungjun Lim, Gyeongdeok Seo, Jihee Kim,
Hojun Park, Jaehun Jung, Kyungwoo Song
* Anomalous Agreement: How to find the Ideal Number of Anomaly Classes in
Correlated, Multivariate Time Series Data
Ferdinand Rewicki, Joachim Denzler, Julia Niebling
* LLM-driven Knowledge Distillation for Dynamic Text-Attributed Graphs
Amit Roy , Ning Yan, Masood Mortazavi
* Evaluation of AI-based Methods for Time-Series Modeling of Measurement Data
of a Complex Cryogenic System
Bryan P. Maldonado, Frank Liu, Pradeep Ramuhalli
* PromptTSS: A Unified Model for Time Series Segmentation with
Multi-Granularity States
Ching Chang, Ming-Chih Lo, Wen-Chih Peng, Tien-Fu Chen
* Evaluating Zero-Shot Foundation Models for Time Series Forecasting in Clinical
Settings: A Simulation Study with Electronic Health Records
Gernot Pucher, Amin Dada, Felix Nensa, Martin Schuler,
Christian Reinhardt, Jens Kleesiek, Christopher M. Sauer,
* ReLATE: Resilient Learner Selection for Multivariate Time-Series Classification
Against Adversarial Attacks
Cagla Ipek Kocal, Onat Gungor, Aaron Tartz, Tajana Rosing, Baris Aksanli
* Integrating Macroeconomic Indicators and U.S. Commodity Futures Prices for
Pair Trading on Chinese Commodity Futures
Xinyu Lin, Guanrou Deng, Li Zhang
* GNN-Based Candidate Node Predictor for Influence Maximization in Temporal
Graphs
Priyanka Gautam, Balasubramaniam Natarajan, Sai Munikoti, S M Ferdous†, Mahantesh Halappanavar
* From Stocks to Sustainability: Predicting Carbon Emissions with Machine
Learning Models
Divya Chaudhary, Saanidhya Vats, Anjali Haryani, Siva Sai Gopaal Praturi
* Patient Stratification with Temporal Self-Supervised Learning
Dimitrios Proios, Alban Bornet, Anthony Yazdani, Douglas Teodoro,
* Self-explainable Reasoning over Temporal Knowledge Graph with Adaptive
Logical Rules
Qing Li, GuanzhongWu, Jingjing Song, Chenglie Du
*WormKAN: Are KAN Effective for Identifying and Tracking Concept Drift in Time
Series?
Kunpeng Xu, Lifei Chen, ShengruiWang
* Enhance Robustness of Deep-Hashing Model in Streaming Time-Series via
Ambiguous-Sample Awareness Training Framework
Richard Tapia, Yifeng Gao
Organizers
