AI Paper Atlas

Research map / Video / Video segmentation and tracking

Unbiased Gradient Estimation for Event Binning via Functional Backpropagation

Jinze Chen, Wei Zhai, Han Han, Tiankai Ma, Yang Cao, Bin Li, Zheng-Jun Zha

ICLR 2026 · Video › Video segmentation and tracking · approach: Event-based video reconstruction · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
synthesize weak-derivative gradients via functional backpropcontrast-maximization optimization · ECD angular velocity estimationRMS error: -10.3% RMS, 1.66x fasterImproved
synthesize weak-derivative gradients via functional backpropcontrast-maximization optimization · ECD linear velocity estimationRMS error: +3.9% RMS error, 1.48x fasterMixed
synthesize weak-derivative gradients via functional backpropcontrast-maximization optimization · ECD motion estimation overallRMS error: -3.2% RMS, 1.57x speedupImproved
use FBP instead of STE surrogate (ablation)FBP · shapes_rotation VarRMS error (deg/s): 96.89 -> 15.89Improved
use FBP instead of sigmoid surrogate (ablation)FBP · boxes_rotation VarRMS error (deg/s): 13.94 -> 12.44Improved
use linear instead of bicubic/Lanczos reconstruction kernel (ablation)FBP · ECD rotation, Var and LLRMS error and time: accuracy within ~0.5 deg/s, faster time e.g. 7.44 vs 8.56 msMixed
replace linear binning with rect kernel using FBP gradientsMotionPriorCMax · DSEC optical flowEPE: -9.4% EPEImproved
replace linear binning with rect kernel using FBP gradientsCMax-SLAM · ECD first 30sAbs RMS error: -5.1% Abs RMSImproved
apply FBP to log-likelihood score with linear binning kernelbias analysis · dynamic_rotation LL scoregradient biasHurt
apply FBP gradient computationJVP of binning · computation time, GPU Ne=20000time (us): Linear JVP 28.1 -> FuncLinear JVP 66.8Hurt

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