AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

Understanding and Enhancing Differentiable Architecture Search from Information Bottleneck Perspective

Haidong Kang, Lianbo Ma, Pengjun Chen, Qiang He, Bo Yi

AAAI 2026 · Architectures and efficiency › Neural architecture search · approach: Bi-level differentiable NAS · 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
add Batch Entropy-decay Regularization (BER) to DARTSDARTS(1st) · NAS-Bench-201 CIFAR-10test accuracy (%): 54.30 -> 94.36Improved
add BER to DARTSDARTS(1st) · NAS-Bench-201 CIFAR-100test accuracy (%): 15.61 -> 73.51Improved
add BER to DARTSDARTS(1st) · NAS-Bench-201 ImageNet16-120test accuracy (%): 16.32 -> 46.34Improved
add BER versus beta-DARTS regularisationbeta-DARTS · NAS-Bench-201search cost (GPU-hours): 3.20 -> 0.16Improved
add BER versus beta-DARTS regularisationbeta-DARTS · NAS-Bench-201 CIFAR-10test accuracy (%): 94.36 -> 94.37No clear effect
use lambda 0.75 with 5 epochs instead of lambda 0.50 with 15 epochs (ablation)r-DARTS · NAS-Bench-201 CIFAR-100test accuracy (%): 73.51 -> 73.09Mixed
search on CIFAR-100 instead of CIFAR-10 (ablation)r-DARTS (λ=0.75) · NAS-Bench-201 ImageNet16-120test accuracy (%): 46.33 -> 47.31Improved
search on CIFAR-100 instead of CIFAR-10 (ablation)r-DARTS (λ=0.75) · NAS-Bench-201 CIFAR-10test accuracy (%): 94.37 -> 93.99Hurt
add BER to DARTSDARTS · ImageNet1k (DARTS search space)top-1 accuracy (%): 73.3 -> 76.9Improved
add BER to DARTSDARTS · DARTS search space memoryGPU memory: 8554 -> 8554No clear effect
add BER to DARTS-based search on MobileNet-like space versus ZiCoZiCo · ImageNet1k (MobileNet-like)top-1 accuracy (%): 78.1 -> 78.9Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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