Dynamic Long-Term Time-Series Forecasting via Meta Transformer Networks

Muhammad Anwar Ma'sum, Rasel Sarkar, Mahardhika Pratama, Savitha Ramasamy, Sreenatha Anavatti, Lin Liu, Habibullah, Ryszard Kowalczyk

Research output: Contribution to journalArticlepeer-review


A reliable long-term time-series forecaster is highly demanded in practice but comes across many challenges such as low computational and memory footprints as well as robustness against dynamic learning environments. This paper proposes Meta-Transformer Networks (MANTRA) to deal with the dynamic long-term time-series forecasting tasks. MANTRA relies on the concept of fast and slow learners where a collection of fast learners learns different aspects of data distributions while adapting quickly to changes. A slow learner tailors suitable representations to fast learners. Fast adaptations to dynamic environments are achieved using the universal representation transformer layers producing task-adapted representations with a small number of parameters. Our experiments using four datasets with different prediction lengths demonstrate the advantage of our approach with at least 3% improvements over the baseline algorithms for both multivariate and univariate settings. Source codes of MANTRA are publicly available in <uri>https://github.com/anwarmaxsum/MANTRA</uri>.

Original languageEnglish
Pages (from-to)1-11
Number of pages11
JournalIEEE Transactions on Artificial Intelligence
Publication statusAccepted/In press - 2024


  • Artificial intelligence
  • Australia
  • concept drifts
  • Deep learning
  • deep learning
  • Forecasting
  • Self-supervised learning
  • Task analysis
  • time-series forecasting
  • Transformers
  • transformers


Dive into the research topics of 'Dynamic Long-Term Time-Series Forecasting via Meta Transformer Networks'. Together they form a unique fingerprint.

Cite this