Research map / Time series / Causal and generative time series
Temporal Latent Auto-Encoder: A Method for Probabilistic Multivariate Time Series Forecasting
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use nonlinear ReLU encoder/decoder instead of a linear FFN ([64, 16]) (ablation) | TLAE · Traffic | WAPE: 0.136 -> 0.107 | Helped |
| Use nonlinear ReLU encoder/decoder instead of a linear FFN ([64, 16]) (ablation) | TLAE · Traffic | MAPE: 0.157 -> 0.120 | Helped |
| Use nonlinear ReLU encoder/decoder instead of a linear FFN ([64, 16]) (ablation) | TLAE · Traffic | SMAPE: 0.130 -> 0.095 | Helped |
| Use nonlinear ReLU encoder/decoder instead of a linear FFN ([64, 32]) (ablation) | TLAE · Traffic | WAPE: 0.117 -> 0.106 | Helped |
| Use nonlinear ReLU encoder/decoder instead of a linear FFN ([64, 32]) (ablation) | TLAE · Traffic | MAPE: 0.134 -> 0.120 | Helped |
| Use nonlinear ReLU encoder/decoder instead of a linear FFN ([64, 32]) (ablation) | TLAE · Traffic | SMAPE: 0.109 -> 0.093 | Helped |
| Increase batch size from L+1 toward 3L (ablation) | TLAE probabilistic point prediction · Traffic | MAPE | Helped |
| Increase the latent regularization weight from a very small value into the useful range (ablation) | TLAE probabilistic point prediction · Traffic | MAPE | Helped |
| Increase latent embedding dimension across [2,4,8,16,32] (ablation) | TLAE probabilistic point prediction · Traffic | MAPE | Mixed |
| Use ℓ1 rather than ℓ2 reconstruction loss on the input series (ablation) | TLAE · Multivariate forecasting datasets | prediction error | Helped |
Closest papers by meaning
- Latent Diffusion Transformer for Probabilistic Time Series Forecasting (AAAI 2024 · similarity 0.91)
- TimePerceiver: An Encoder-Decoder Framework for Generalized Time-Series Forecasting (NeurIPS 2025 · similarity 0.90)
- Multivariate Probabilistic Time Series Forecasting with Correlated Errors (NeurIPS 2024 · similarity 0.89)
- OLinear: A Linear Model for Time Series Forecasting in Orthogonally Transformed Domain (NeurIPS 2025 · similarity 0.89)
- HDT: Hierarchical Discrete Transformer for Multivariate Time Series Forecasting (AAAI 2025 · similarity 0.89)
- TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables (NeurIPS 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.