AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Memory Savings at What Cost? A Study of Alternatives to Backpropagation

Kunjal Panchal, Sunav Choudhary, Yuriy Brun, Hui Guan

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Neural scaling laws · 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
use BP with checkpointing instead of ZO-Vanillafine-tuning · BoolQ (Llama 3.1 8B)test accuracy (%): 57.1 -> 88.3Improved
use BP with checkpointing instead of FMAD-Vanillafine-tuning · AGNews (Llama 3.1 8B)test accuracy (%): 80.5 -> 94.2Improved
use FMAD instead of ZO (analytic jvp vs finite difference)gradient estimation · AGNews (Llama 3.1 8B)test accuracy (%): 73.6 -> 80.5Improved
add gradient accumulation (ablation)FMAD-Vanilla · BoolQ (Llama 3.1 8B)test accuracy (%): 60.7 -> 70.3Improved
add gradient accumulation (ablation)FMAD-Vanilla · GSM8K (Llama 3.1 8B)test accuracy (%): 37.7 -> 30.8Hurt
average over multiple perturbations per iteration (n=10, 50) (ablation)ZO and FMAD · AGNews (Llama 3.1 8B)test accuracy: +13.2-14.0% (ZO), +5.7-7.7% (FMAD)Improved
increase perturbations to 10 per iteration in ZO (ablation)ZO-Vanilla · AGNews (Llama 3.1 8B)compute cost: 9.7x computeHurt
accumulation window of 200 (ablation)ZO and FMAD · AGNews (Llama 3.1 8B)convergence time: 14.8x (FMAD), 10.9x (ZO) slowerHurt
add adaptive perturbation sampling (ablation)FMAD-Vanilla · BoolQ and MultiRC (Llama 3.1 8B)test accuracy (%): -4.3% BoolQ, -3.2% MultiRCHurt
add sparse parameter updates (top-1%) (ablation)FMAD and ZO · Llama 3.1 8B taskstest accuracy: -1.2 to -10.1% (FMAD), -1.9 to -9.0% (ZO)Hurt
add SVRG variance reduction (ablation)FMAD and ZO · classification tasks (Llama 3.1 8B)test accuracy: +4.0 to +11.1% classification, -4.2% GSM8KMixed
use BP-Checkpointing instead of ZO-Vanilla in total computetraining cost · AGNews (Llama 3.1 8B)TFLOPs until convergence: 251.2 -> 65.2 (x10^4)Improved

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