AIREPAIR: A Repair Platform for Neural Networks

Xidan Song; Youcheng Sun; Mustafa A. Mustafa; Lucas C. Cordeiro

AIREPAIR: A Repair Platform for Neural Networks

Xidan Song, Youcheng Sun, Mustafa A. Mustafa, Lucas C. Cordeiro

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

112 Downloads (Pure)

Abstract

We present AIREPAIR, a platform for repairing neural networks. It features the integration of existing network repair tools. Based on AIREPAIR, one can run different repair methods on the same model, thus enabling the fair comparison of different repair techniques. In this paper, we evaluate AIREPAIR with five recent repair methods on popular deeplearning datasets and models. Our evaluation confirms the utility of AIREPAIR, by comparing and analyzing the results from
different repair techniques. A demonstration is available at https://youtu.be/UkKw5neeWhw.

Original language	English
Title of host publication	45th IEEE/ACM International Conference on Software Engineering
Publication status	Accepted/In press - 20 Jan 2023

Access to Document

icse2023Accepted author manuscript, 410 KB

EnnCore: End-to-End Conceptual Guarding of Neural Architectures
Cordeiro, L. (PI), Brown, G. (CoI), Freitas, A. (CoI), Luján, M. (CoI) & Mustafa, M. (CoI)
1/02/21 → 31/12/24
Project: Research

Cite this

@inproceedings{402d15fa64f54fa9ab84f9ee3da9a550,

title = "AIREPAIR: A Repair Platform for Neural Networks",

abstract = "We present AIREPAIR, a platform for repairing neural networks. It features the integration of existing network repair tools. Based on AIREPAIR, one can run different repair methods on the same model, thus enabling the fair comparison of different repair techniques. In this paper, we evaluate AIREPAIR with five recent repair methods on popular deeplearning datasets and models. Our evaluation confirms the utility of AIREPAIR, by comparing and analyzing the results fromdifferent repair techniques. A demonstration is available at https://youtu.be/UkKw5neeWhw.",

author = "Xidan Song and Youcheng Sun and Mustafa, {Mustafa A.} and Cordeiro, {Lucas C.}",

year = "2023",

month = jan,

day = "20",

language = "English",

booktitle = "45th IEEE/ACM International Conference on Software Engineering",

}

TY - GEN

T1 - AIREPAIR: A Repair Platform for Neural Networks

AU - Song, Xidan

AU - Sun, Youcheng

AU - Mustafa, Mustafa A.

AU - Cordeiro, Lucas C.

PY - 2023/1/20

Y1 - 2023/1/20

N2 - We present AIREPAIR, a platform for repairing neural networks. It features the integration of existing network repair tools. Based on AIREPAIR, one can run different repair methods on the same model, thus enabling the fair comparison of different repair techniques. In this paper, we evaluate AIREPAIR with five recent repair methods on popular deeplearning datasets and models. Our evaluation confirms the utility of AIREPAIR, by comparing and analyzing the results fromdifferent repair techniques. A demonstration is available at https://youtu.be/UkKw5neeWhw.

AB - We present AIREPAIR, a platform for repairing neural networks. It features the integration of existing network repair tools. Based on AIREPAIR, one can run different repair methods on the same model, thus enabling the fair comparison of different repair techniques. In this paper, we evaluate AIREPAIR with five recent repair methods on popular deeplearning datasets and models. Our evaluation confirms the utility of AIREPAIR, by comparing and analyzing the results fromdifferent repair techniques. A demonstration is available at https://youtu.be/UkKw5neeWhw.

M3 - Conference contribution

BT - 45th IEEE/ACM International Conference on Software Engineering

ER -

AIREPAIR: A Repair Platform for Neural Networks

Abstract

Access to Document

Fingerprint

Projects

EnnCore: End-to-End Conceptual Guarding of Neural Architectures

Cite this