AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

An In-Depth Study on Deep Learning Model Cloning

Bin Hu, Xiancong Pan, Dongjin Yu, Tianyi Hu

ICML 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Parameter-efficient transfer analysis

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 structural similarity alone (drop weight similarity) (ablation)MCDetector · ModelCloneBench Type-1 and Type-2detection P/R/F1: 1.00 -> N/AHurt
use weight similarity alone (drop structural similarity) (ablation)MCDetector · ModelCloneBench Type-1 to Type-3detection P/R/F1: 1.00 -> N/AHurt
add weight similarity on top of structural similarity for Type-3 clones (ablation)MCDetector · ModelCloneBench Type-3F1: 0.96 -> 0.96No clear effect
set structural similarity threshold to 0.7MCDetector · ModelCloneBench Type-3recall: 1.0000 precision, 0.9146 recallMixed
combine structural and weight similarity for retrieval of derived modelsMCDetector · HuggingFace derivation retrievalHit@k: Hit@1 0.58 -> Hit@3 1.00Improved

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