AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

The Optimal Control Foundation of Early Exits – Turnpikes and ResNets

Jens Püttschneider, Simon Heilig, Asja Fischer, Timm Faulwasser

NeurIPS 2026 · Architectures and efficiency › RNNs and normalization · approach: Theory of depth and width

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.

Analyzes joint optimization of output-layer and intermediate-layer losses (early exits) in residual networks through an optimal control lens, linking it to turnpike properties.

Problem

Early-exit training jointly optimizes losses at the output and intermediate layers to trade off efficiency and accuracy, but the theoretical behavior of this joint optimization is not well understood.

Approach

Frames deep residual networks with intermediate losses as an optimal control problem and analyzes the joint optimization of intermediate and output losses using turnpike theory.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase stage-cost weight gamma to >=0.2 (ablation)Early-exit ResNet-110 joint training · CIFAR-10 (N=54, M=1)final-block performanceHurt
use very small gamma (<=0.002) (ablation)Early-exit ResNet-110 joint training · CIFAR-10 (N=54, M=1)early-layer convergenceHurt
share early-exit parameters with the backbone (ablation)Early-exit ResNet · CIFAR-10test accuracyHurt
place exits every 6 blocks instead of every block (ablation)Early-exit ResNet · CIFAR-10test accuracyImproved
place exits every 6 blocks instead of every block (ablation)Early-exit ResNet · CIFAR-10average exit depth (compute)Hurt
use homogeneous (constant 64-filter) ResNet instead of standard (ablation)Early-exit ResNet · CIFAR-10/CIFAR-100test accuracyHurt
add joint early-exit training (M=1) to ResNet-110ResNet-110 · TinyImageNet testtest accuracy (%): 62.4 -> 58.3Hurt
increase backbone depth (N=18,36,54) under joint training (ablation)Early-exit ResNet · CIFAR-10 early-exit accuracy vs thresholdearly-exit accuracyNo clear effect

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