[CVPR 2026] Elucidating the SNR-t Bias of Diffusion Probabilistic Models
-
Updated
May 11, 2026 - Python
#
exposure-bias
Here are 5 public repositories matching this topic...
Reference implementation for the paper "Rollout-Decoded Reconstruction for Latent World Models": one loss term that trains a latent world model's decoder on its own free-running rollout, at zero added parameters.
reinforcement-learning chaotic-systems model-based-rl state-space-models time-series-forecasting dreamer world-models scientific-machine-learning neural-operators exposure-bias
-
Updated
Aug 19, 2026 - Python
Reciprocal recommendation and two-sided ranking with joint match prediction, exposure allocation, and leakage-safe evaluation.
learning-to-rank fairness counterfactual-simulation offline-evaluation exposure-bias reciprocal-recommendation
-
Updated
Sep 10, 2026 - Python
Unbiased offline evaluation on real Kuaishou logs: ranking survives randomized exposure, calibration does not, and no single calibrator fixes both.
python machine-learning pytorch calibration lightgbm recommender-systems off-policy-evaluation duckdb two-tower exposure-bias counterfactual-evaluation kuairand
-
Updated
Sep 14, 2026 - HTML
Two-stage short-video recommender with a pre-registered, exposure-unbiased evaluation — built, evaluated, and deployed on Cloud Run.
-
Updated
Sep 23, 2026 - Python
Add this topic to your repo
To associate your repository with the exposure-bias topic, visit your repo's landing page and select "manage topics."