Pith. sign in

REVIEW 11 cited by

TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2402.19072 v4 pith:R2HM75JQ submitted 2024-02-29 cs.LG cs.AI

TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables

classification cs.LG cs.AI
keywords variablesforecastingendogenousexogenousinformationtimexerseriestime
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Deep models have demonstrated remarkable performance in time series forecasting. However, due to the partially-observed nature of real-world applications, solely focusing on the target of interest, so-called endogenous variables, is usually insufficient to guarantee accurate forecasting. Notably, a system is often recorded into multiple variables, where the exogenous variables can provide valuable external information for endogenous variables. Thus, unlike well-established multivariate or univariate forecasting paradigms that either treat all the variables equally or ignore exogenous information, this paper focuses on a more practical setting: time series forecasting with exogenous variables. We propose a novel approach, TimeXer, to ingest external information to enhance the forecasting of endogenous variables. With deftly designed embedding layers, TimeXer empowers the canonical Transformer with the ability to reconcile endogenous and exogenous information, where patch-wise self-attention and variate-wise cross-attention are used simultaneously. Moreover, global endogenous tokens are learned to effectively bridge the causal information underlying exogenous series into endogenous temporal patches. Experimentally, TimeXer achieves consistent state-of-the-art performance on twelve real-world forecasting benchmarks and exhibits notable generality and scalability. Code is available at this repository: https://github.com/thuml/TimeXer.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 11 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. $\text{DT}^2$: Decision-Targeted Digital Twins

    cs.LG 2026-06 unverdicted novelty 7.0

    DT² trains digital twins to preserve pairwise policy rankings from fitted Q-evaluation on offline data rather than minimizing one-step transition errors, improving policy ranking and reducing decision regret.

  2. Deep Time Series Models: A Comprehensive Survey and Benchmark

    cs.LG 2024-07 unverdicted novelty 7.0

    This survey and benchmark of deep time series models using the released TSLib library finds that models with specific structures perform well only on distinct analysis tasks.

  3. MIDiff: Tackling Sparsity and Imbalance in Mobile Usage Generation via Multivariate-Imaging Diffusion

    cs.LG 2026-07 conditional novelty 6.0

    A diffusion model trained on C-GASF images of mobile usage generates synthetic user traces that match real trace statistics far better than prior time-series generative baselines.

  4. Physics-Guided Spatiotemporal State Space Modeling for Lookahead Molten Pool Segmentation in Laser Wire-Feed Welding

    cs.CV 2026-06 unverdicted novelty 6.0

    WeldMamba achieves 74.63% mIoU for 500 ms lookahead segmentation of keyhole, wire, and molten pool using spatiotemporal state space modeling conditioned on welding signals and physics-based losses on a 43-sequence dataset.

  5. Masked Training for Robust Arrhythmia Detection from Digitalized Multiple Layout ECG Images

    cs.LG 2025-08 conditional novelty 6.0

    PatchECG applies masked patch training and disordered attention to handle asynchronous and partially missing ECG signals from varied layouts, reaching average AUROC 0.835 on simulated conditions and 0.778 on real hosp...

  6. Time Series Forecasting as Reasoning: A Slow-Thinking Approach with Reinforced LLMs

    cs.LG 2025-06 unverdicted novelty 6.0

    Time-R1 trains LLMs via supervised fine-tuning followed by reinforcement learning with a time-series-specific reward and non-uniform GRIP sampling to enable multi-step reasoning that improves forecasting accuracy.

  7. Hopformer: Homogeneity-Pursuit Transformer for Time Series Forecasting

    stat.ML 2026-07 reject novelty 5.0

    A two-stage forecaster (SPA trend extraction + LoRA-fine-tuned residual Transformer) that the paper claims beats prior models by 6.56% MASE, though the claim is not robust to its own extended baseline tables.

  8. Exogenous Dropout: A Simple, Strong Baseline for Corruption-Robust Time Series Forecasting with Covariates

    cs.LG 2026-07 conditional novelty 5.0

    Exogenous dropout, a model-agnostic whole-channel training mask, restores corruption robustness better than a deliberately bounded gated-FiLM architecture across three forecasting domains.

  9. Zeus: Towards Tuning-Free Foundation Model for Time Series Analysis

    cs.LG 2026-07 unverdicted novelty 5.0

    Zeus proposes a multi-scale Transformer with point-wise tokenization and Multi-Objective Temporal Masking to enable tuning-free performance on forecasting, interpolation, and other time series tasks.

  10. Ister: Linear Transformer for Efficient Multivariate Time Series Forecasting

    cs.LG 2024-12 unverdicted novelty 5.0

    Ister is a linear-complexity transformer using Dot-attention and inverted seasonal-trend decomposition for multivariate time series forecasting that reports state-of-the-art benchmark performance.

  11. One Step Closer to Ground Truth: A Multi-Scale Residual-Aware Representation Learning Pipeline for Predicting Time Series Data

    cs.LG 2026-06 unverdicted novelty 4.0

    A two-stage residual-aware framework adds a meta-corrector after a base transformer to model structured errors and reports state-of-the-art results on eight time-series benchmarks.