Research map / Time series / Long-term forecasting
Learning Deep Time-index Models for Time Series Forecasting
Gerald Woo, Chenghao Liu, Doyen Sahoo, Akshat Kumar, Steven Hoi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 DeepTime instead of N-HiTS | multivariate time-series forecasting · ETTm2, horizon 96 | MSE: 0.176 -> 0.166 | Improved |
| Use DeepTime instead of N-HiTS | multivariate time-series forecasting · ECL, horizon 96 | MSE: 0.147 -> 0.137 | Improved |
| Use DeepTime instead of ETSformer | multivariate time-series forecasting · Exchange, horizon 96 | MSE: 0.085 -> 0.081 | Improved |
| Use DeepTime instead of N-HiTS | multivariate time-series forecasting · Traffic, horizon 96 | MSE: 0.402 -> 0.390 | Improved |
| Use DeepTime instead of N-HiTS | multivariate time-series forecasting · Weather, horizon 96 | MSE: 0.158 -> 0.166 | Hurt |
| Use DeepTime instead of N-HiTS | multivariate time-series forecasting · ILI, horizon 24 | MSE: 1.862 -> 2.425 | Hurt |
| Replace random Fourier features with a linear input map (ablation) | DeepTime · ETTm2 multivariate, horizon 96 | MSE: 0.166 -> 0.186 | Hurt |
| Use periodic activations instead of DeepTime's random Fourier feature backbone (ablation) | DeepTime · ETTm2 multivariate, horizon 96 | MSE: 0.166 -> 0.236 | Hurt |
| Use an autoregressive RNN backbone (ablation) | DeepTime · ETTm2 multivariate, horizon 96 | MSE: 0.166 -> 0.233 | Hurt |
| Add datetime features (ablation) | DeepTime · ETTm2 multivariate, horizon 96 | MSE: 0.166 -> 0.226 | Hurt |
| Replace the closed-form ridge regressor with a gradient-trained linear layer (ablation) | DeepTime · ETTm2 multivariate, horizon 96 | MSE: 0.166 -> 3.072 | Hurt |
| Add training an INR from scratch for each lookback window (ablation) | DeepTime · ETTm2 multivariate, horizon 96 | MSE: 0.166 -> 0.251 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- N-HiTS — NHITS: Neural Hierarchical Interpolation for Time Series Forecasting (AAAI 2023) · better on 3 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts (ICLR 2025 · similarity 0.89)
- LETS Forecast: Learning Embedology for Time Series Forecasting (ICML 2025 · similarity 0.89)
- Selective Learning for Deep Time Series Forecasting (NeurIPS 2025 · similarity 0.89)
- SynTSBench: Rethinking Temporal Pattern Learning in Deep Learning Models for Time Series (NeurIPS 2025 · similarity 0.88)
- Time-LLM: Time Series Forecasting by Reprogramming Large Language Models (ICLR 2024 · similarity 0.88)
- TEMPO: Prompt-based Generative Pre-trained Transformer for Time Series Forecasting (ICLR 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.
