IKKA: Inversion Classification via Critical Anomalies for Robust Visual Servoing
2026-04-09 • Machine Learning
Machine Learning
AI summaryⓘ
The authors present IKKA, a new method that helps robots better interpret tricky visual information when conditions change, like in poor lighting or partial blockage. Instead of ignoring unusual data points as mistakes, IKKA treats them as important signals for making smarter decisions. This method uses a combination of mathematical ideas to weigh these points and guide control actions more accurately. Tested on a small computer, IKKA improved accuracy and speed in difficult scenarios compared to previous methods.
visual servoingdistribution shifttopological weightinganomaly detectionlocal extremalityboundary transversalitymulti-scale persistenceRaspberry Picontrol systemsnon-parametric analysis
Authors
Darya Pavlenko
Abstract
We introduce IKKA (Inversion Classification via Critical Anomalies), a topologically motivated weighting framework for robust visual servoing under distribution shift. Unlike conventional outlier handling, IKKA treats maverick points as structurally informative observations: points where small perturbations can induce qualitatively different control responses or class assignments. The method combines local extremality, boundary transversality, and multi-scale persistence into a single anomaly weight, W(x) = E(x) x T(x) x M(x), which modulates control updates near ambiguous decision regions. We instantiate IKKA in a CPU-only embedded visual-servoing pipeline on Raspberry Pi 4 and evaluate it across 230 reproducible runs under nominal and stress conditions. In stress scenarios involving dim illumination and transient occlusion, IKKA reduces the 95th-percentile lateral error by 24% relative to a hybrid baseline (0.124 to 0.094) while increasing throughput from 20.0 to 24.8 Hz. Non-parametric analysis confirms a large effect size (Cliff's delta = 0.79).