The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Dalam kertas kerja ini, kami mencadangkan skim penilaian pokok keputusan peribadi (PDTE) pertama yang sesuai untuk digunakan dalam senario Pembelajaran Mesin sebagai Perkhidmatan (MLaaS). Dalam skim kami, pengguna dan pemilik model menghantar teks sifir sampel dan model pepohon keputusan, masing-masing, dan pelayan tunggal mengelaskan sampel tanpa mengetahui sampel mahupun pepohon keputusan. Walaupun banyak skim PDTE telah dicadangkan setakat ini, kebanyakannya perlu mendedahkan pepohon keputusan kepada pelayan. Ini adalah tidak diingini kerana model klasifikasi ialah harta intelek pemilik model, dan/atau ia mungkin termasuk maklumat sensitif yang digunakan untuk melatih model, dan oleh itu model juga harus disembunyikan daripada pelayan. Dalam skim PDTE yang lain, berbilang pelayan bersama-sama menjalankan proses pengelasan dan pepohon keputusan dirahsiakan daripada pelayan dengan andaian mereka tidak bersekongkol. Malangnya, andaian ini mungkin tidak berlaku kerana MLaaS biasanya disediakan oleh satu syarikat. Sebaliknya, skim kami tidak mempunyai masalah sedemikian. Pada dasarnya, penyulitan homomorfik sepenuhnya membolehkan kami mengklasifikasikan sampel yang disulitkan berdasarkan pepohon keputusan yang disulitkan, dan sebenarnya, skema PDTE bukan interaktif sedia ada boleh diubah suai supaya pelayan mengelaskan hanya mengendalikan teks sifir. Walau bagaimanapun, skim yang terhasil adalah kurang cekap daripada kami. Kami juga menunjukkan keputusan percubaan untuk skim kami.
Yoshifumi SAITO
Tokyo Institute of Technology
Wakaha OGATA
Tokyo Institute of Technology
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Salinan
Yoshifumi SAITO, Wakaha OGATA, "Private Decision Tree Evaluation by a Single Untrusted Server for Machine Learnig as a Service" in IEICE TRANSACTIONS on Fundamentals,
vol. E105-A, no. 3, pp. 203-213, March 2022, doi: 10.1587/transfun.2021CIP0004.
Abstract: In this paper, we propose the first private decision tree evaluation (PDTE) schemes which are suitable for use in Machine Learning as a Service (MLaaS) scenarios. In our schemes, a user and a model owner send the ciphertexts of a sample and a decision tree model, respectively, and a single server classifies the sample without knowing the sample nor the decision tree. Although many PDTE schemes have been proposed so far, most of them require to reveal the decision tree to the server. This is undesirable because the classification model is the intellectual property of the model owner, and/or it may include sensitive information used to train the model, and therefore the model also should be hidden from the server. In other PDTE schemes, multiple servers jointly conduct the classification process and the decision tree is kept secret from the servers under the assumption they do not collude. Unfortunately, this assumption may not hold because MLaaS is usually provided by a single company. In contrast, our schemes do not have such problems. In principle, fully homomorphic encryption allows us to classify an encrypted sample based on an encrypted decision tree, and in fact, the existing non-interactive PDTE scheme can be modified so that the server classifies only handling ciphertexts. However, the resulting scheme is less efficient than ours. We also show the experimental results for our schemes.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2021CIP0004/_p
Salinan
@ARTICLE{e105-a_3_203,
author={Yoshifumi SAITO, Wakaha OGATA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Private Decision Tree Evaluation by a Single Untrusted Server for Machine Learnig as a Service},
year={2022},
volume={E105-A},
number={3},
pages={203-213},
abstract={In this paper, we propose the first private decision tree evaluation (PDTE) schemes which are suitable for use in Machine Learning as a Service (MLaaS) scenarios. In our schemes, a user and a model owner send the ciphertexts of a sample and a decision tree model, respectively, and a single server classifies the sample without knowing the sample nor the decision tree. Although many PDTE schemes have been proposed so far, most of them require to reveal the decision tree to the server. This is undesirable because the classification model is the intellectual property of the model owner, and/or it may include sensitive information used to train the model, and therefore the model also should be hidden from the server. In other PDTE schemes, multiple servers jointly conduct the classification process and the decision tree is kept secret from the servers under the assumption they do not collude. Unfortunately, this assumption may not hold because MLaaS is usually provided by a single company. In contrast, our schemes do not have such problems. In principle, fully homomorphic encryption allows us to classify an encrypted sample based on an encrypted decision tree, and in fact, the existing non-interactive PDTE scheme can be modified so that the server classifies only handling ciphertexts. However, the resulting scheme is less efficient than ours. We also show the experimental results for our schemes.},
keywords={},
doi={10.1587/transfun.2021CIP0004},
ISSN={1745-1337},
month={March},}
Salinan
TY - JOUR
TI - Private Decision Tree Evaluation by a Single Untrusted Server for Machine Learnig as a Service
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 203
EP - 213
AU - Yoshifumi SAITO
AU - Wakaha OGATA
PY - 2022
DO - 10.1587/transfun.2021CIP0004
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E105-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2022
AB - In this paper, we propose the first private decision tree evaluation (PDTE) schemes which are suitable for use in Machine Learning as a Service (MLaaS) scenarios. In our schemes, a user and a model owner send the ciphertexts of a sample and a decision tree model, respectively, and a single server classifies the sample without knowing the sample nor the decision tree. Although many PDTE schemes have been proposed so far, most of them require to reveal the decision tree to the server. This is undesirable because the classification model is the intellectual property of the model owner, and/or it may include sensitive information used to train the model, and therefore the model also should be hidden from the server. In other PDTE schemes, multiple servers jointly conduct the classification process and the decision tree is kept secret from the servers under the assumption they do not collude. Unfortunately, this assumption may not hold because MLaaS is usually provided by a single company. In contrast, our schemes do not have such problems. In principle, fully homomorphic encryption allows us to classify an encrypted sample based on an encrypted decision tree, and in fact, the existing non-interactive PDTE scheme can be modified so that the server classifies only handling ciphertexts. However, the resulting scheme is less efficient than ours. We also show the experimental results for our schemes.
ER -