TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
We introduce TiRex-2, a recurrent xLSTM-based time series foundation model that generalizes the univariate TiRex to multivariate forecasting with both past and future covariates. Real-world forecasting is inherently sequential: observations arrive continuously, variables evolve jointly, and a subset of covariates is known ahead of time. Existing Transformer-based time series foundation models capture cross-variate dependencies but incur quadratic complexity in context length and require full-history recomputation as new observations arrive. TiRex-2 addresses these limitations through a memory-
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
- Linked via arxiv authorPatrick Podest →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorMarco Pichler →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorElias Bürger →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorLevente Zólyomi →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorBernhard Voggenberger →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorWilhelm Berghammer →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorDaniel Klotz →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorSebastian Böck →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorGünter Klambauer →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming
- Linked via arxiv authorSepp Hochreiter →
TiRex-2: Generalizing TiRex to Multivariate Data and Streaming