AI Paper Atlas

Research map / Time series / Long-term forecasting

Rethinking Time Series Tokenization from a Frequency Perspective

Zelong Tian, Ming Jin, Bo Du, Shirui Pan

NeurIPS 2026 · Time series › Long-term forecasting · approach: Spatiotemporal forecasting models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Argues that common subsequence-based tokenization of time series introduces spectral distortions that push learned representations away from true trend, periodicity, and seasonality, and studies this from a frequency perspective.

Problem

Tokenizing time series by partitioning them into subsequences (patches) is standard in time series modeling, but the authors claim current tokenization strategies inherently distort the spectrum, hurting representations of trend, periodicity, and seasonality.

Approach

Analyzes subsequence tokenization theoretically through a frequency lens, showing it induces spectral distortion; the abstract does not specify the proposed fix.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace patch tokenizer with TimeFT in PatchTSTPatchTST · ETTm1 (H=96)MSE: 0.321 -> 0.290Improved
replace phase tokenizer with TimeFT in PhaseFormerPhaseFormer · ETTm1 (H=96)MSE: 0.312 -> 0.293Improved
replace patch tokenizer with TimeFT in PatchTSTPatchTST · Traffic (H=96)MSE: 0.403 -> 0.395Improved
add Stage 1 frequency-domain decomposition (ablation)TimeFT · ETTm1 (H=96)MSE: 0.328 -> 0.290Improved
add Stage 2 spectral shifting (ablation)TimeFT · ETTm2 (H=96)MSE: 0.194 -> 0.162Improved
add Stage 3 information-preserving token extraction (ablation)TimeFT · ETTm1 (H=96)MSE: 0.375 -> 0.290Improved
add Stage 3 token extraction (ablation)TimeFT · Traffic (H=96)MSE: 0.475 -> 0.395Improved
use TimeFT tokens instead of patch tokens in PatchTST backbonePatchTST · ETTh1 (avg over H)MSE: 0.417 vs TimeMosaic 0.412Mixed
use TimeFT instead of baselines on WeatherPatchTST · Weather (avg over H)MSE: 0.232 vs xPatch 0.224No clear effect
increase tokenization stride for longer inputs (ablation)TimeFT · ETTm1 (N=96 vs N=720)MSE/MAEMixed
use TimeFT tokens for classificationPatchTST · UEA (10 datasets)accuracy (%): 69.13 -> 71.31Improved
use TimeFT tokens for anomaly detectionPatchTST · 5 anomaly datasetsF1 (%): 82.79 -> 86.12Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most 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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy