JEPA · 2026-10-28

JEPA-Inspired Latent Field Prediction for Sparse-Pilot MIMO-OFDM Channel Reconstruction

Yuzhi Yang, Brahim Mefgouda, Hang Zou, Lina Bariah, Anis Bara, Yuhuan Lu, Hao Zhang, Merouane Debbah

arXiv:2610.37391PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

The authors propose a JEPA-inspired field world model (FWM) for sparse-pilot MIMO-OFDM channel reconstruction. CSI is tokenized into latent time-frequency tokens; the model predicts masked/future latent field states from historical multi-band context, then fuses the predicted latent prior with sparse noisy pilots to reconstruct full CSI. Ray-tracing experiments show improved NMSE over pilot-only and direct reconstruction baselines.

Why it matters

Wireless channel prediction can be fragile when operating directly on raw, phase-sensitive CSI coefficients. By predicting in a latent propagation-field domain and using the prediction as a structured prior for pilot-assisted reconstruction, the method targets more stable propagation structure while leaving fine details to be recovered from sparse current pilots, reducing pilot overhead.

Method

  • Tokenize CSI cells into K latent tokens (time-frequency field) with 2D positional encoding and learnable within-cell token embeddings.
  • Train a JEPA-style two-stage ViT predictor: context encoder uses only visible cells; predictor uses projected context plus learnable query tokens to predict masked/future latent tokens.
  • For reconstruction, fuse predicted latent prior with sparse noisy pilots (zero-filled pilot pattern) via a latent-conditioned regressor, using teacher-forced latent conditions during downstream training.

Limitation

The authors note that the simple prediction baseline “fails to converge due to the random disturbance in the locations.”

Abstract (from arXiv)

Accurate acquisition of full channel state information (CSI) is essential for multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) wireless systems, while dense pilot-based estimation incurs substantial overhead. Directly minimizing the prediction error over raw CSI coefficients is fragile in realistic propagation environments, since small-scale fading, mobility, local scattering, and small spatial perturbations can induce large phase-sensitive variations. Nevertheless, wireless channels contain strong spatial, temporal, and frequency-domain correlations that can be more predictably represented in a latent propagation-field domain. Motivated by this observation, we propose a joint-embedding predictive architecture (JEPA)-inspired field world model, defined as a latent predictive model of the wireless propagation field, for sparse-pilot channel reconstruction. The model maps CSI observations into latent time-frequency tokens, predicts masked or future latent field states from historical multi-band context, and fuses the predicted latent prior with sparse noisy current pilots to reconstruct the full MIMO-OFDM channel. Ray-tracing experiments show that the proposed method improves normalized mean square error (NMSE) over pilot-only and direct reconstruction baselines, supporting latent field prediction as an effective prior for sparse-pilot channel acquisition.

Related papers

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First public release: 2026-10-27 · Indexed by Pub-AI: 2026-10-28 UTC

preprint

JEPA-Inspired Latent Field Prediction for Sparse-Pilot MIMO-OFDM Channel Reconstruction

Yuzhi Yang, Brahim Mefgouda, Hang Zou, Lina Bariah, Anis Bara, Yuhuan Lu, Hao Zhang, Merouane Debbah

Preprint posted: 2026-10-27 · Original record ↗

arxiv
2610.37391
Explicit source versions (1)
  • v1 · 2026-10-27
  1. Preprint available

    Source date: 2026-10-27 · First detected: 2026-10-28 UTC · Historical baseline

    Source evidence (2)
Abstract · arxiv

Accurate acquisition of full channel state information (CSI) is essential for multiple-input multiple-output orthogonal frequency-division multiplexing (MIMO-OFDM) wireless systems, while dense pilot-based estimation incurs substantial overhead. Directly minimizing the prediction error over raw CSI coefficients is fragile in realistic propagation environments, since small-scale fading, mobility, local scattering, and small spatial perturbations can induce large phase-sensitive variations. Nevertheless, wireless channels contain strong spatial, temporal, and frequency-domain correlations that can be more predictably represented in a latent propagation-field domain. Motivated by this observation, we propose a joint-embedding predictive architecture (JEPA)-inspired field world model, defined as a latent predictive model of the wireless propagation field, for sparse-pilot channel reconstruction. The model maps CSI observations into latent time-frequency tokens, predicts masked or future latent field states from historical multi-band context, and fuses the predicted latent prior with sparse noisy current pilots to reconstruct the full MIMO-OFDM channel. Ray-tracing experiments show that the proposed method improves normalized mean square error (NMSE) over pilot-only and direct reconstruction baselines, supporting latent field prediction as an effective prior for sparse-pilot channel acquisition.