Visible to the public Biblio

Filters: Author is Sakuma, Jun  [Clear All Filters]
2019-11-11
Kunihiro, Noboru, Lu, Wen-jie, Nishide, Takashi, Sakuma, Jun.  2018.  Outsourced Private Function Evaluation with Privacy Policy Enforcement. 2018 17th IEEE International Conference On Trust, Security And Privacy In Computing And Communications/ 12th IEEE International Conference On Big Data Science And Engineering (TrustCom/BigDataSE). :412–423.
We propose a novel framework for outsourced private function evaluation with privacy policy enforcement (OPFE-PPE). Suppose an evaluator evaluates a function with private data contributed by a data contributor, and a client obtains the result of the evaluation. OPFE-PPE enables a data contributor to enforce two different kinds of privacy policies to the process of function evaluation: evaluator policy and client policy. An evaluator policy restricts entities that can conduct function evaluation with the data. A client policy restricts entities that can obtain the result of function evaluation. We demonstrate our construction with three applications: personalized medication, genetic epidemiology, and prediction by machine learning. Experimental results show that the overhead caused by enforcing the two privacy policies is less than 10% compared to function evaluation by homomorphic encryption without any privacy policy enforcement.
2019-06-24
Yakura, Hiromu, Shinozaki, Shinnosuke, Nishimura, Reon, Oyama, Yoshihiro, Sakuma, Jun.  2018.  Malware Analysis of Imaged Binary Samples by Convolutional Neural Network with Attention Mechanism. Proceedings of the Eighth ACM Conference on Data and Application Security and Privacy. :127–134.
This paper presents a proposal of a method to extract important byte sequences in malware samples to reduce the workload of human analysts who investigate the functionalities of the samples. This method, by applying convolutional neural network (CNN) with a technique called attention mechanism to an image converted from binary data, enables calculation of an "attention map," which shows regions having higher importance for classification in the image. This distinction of regions enables extraction of characteristic byte sequences peculiar to the malware family from the binary data and can provide useful information for the human analysts without a priori knowledge. Furthermore, the proposed method calculates the attention map for all binary data including the data section. Thus, it can process packed malware that might contain obfuscated code in the data section. Results of our evaluation experiment using malware datasets show that the proposed method provides higher classification accuracy than conventional methods. Furthermore, analysis of malware samples based on the calculated attention maps confirmed that the extracted sequences provide useful information for manual analysis, even when samples are packed.
2019-03-22
Lu, Wen-jie, Sakuma, Jun.  2018.  More Practical Privacy-Preserving Machine Learning As A Service via Efficient Secure Matrix Multiplication. Proceedings of the 6th Workshop on Encrypted Computing & Applied Homomorphic Cryptography. :25-36.

An efficient secure two-party computation protocol of matrix multiplication allows privacy-preserving cloud-aid machine learning services such as face recognition and traffic-aware navigation. We use homomorphic encryption to construct a secure matrix multiplication protocol with a small communication overhead and computation overhead on the client's side, which works particularly well when a large number of clients access to the server simultaneously. The fastest secure matrix multiplication protocols have been constructed using tools such as oblivious transfer, but a potential limitation of these methods is the needs of using a wide network bandwidth between the client and the server, e.g., 10\textasciitildeGbps. This is of particular concern when thousands of clients interact with the server concurrently. Under this setting, the performance oblivious transfer-based methods will decrease significantly, since the server can only allocate a small ratio of its outgoing bandwidth for each client. With three proposed optimizations, our matrix multiplication protocol can run very fast even under the high concurrent setting. Our benchmarks show that it takes an Amazon instance (i.e., 72 CPUs and 25 Gbps outgoing bandwidth) less than 50 seconds to complete 1000 concurrent secure matrix multiplications with \$128\textbackslashtimes 128\$ entries. In addition, our method reduces more than \$74% - 97%\$ of the precomputation time of two privacy-preserving machine learning frameworks, SecureML (S&P'17) and MiniONN (CCS'17).

2018-06-07
Yakura, Hiromu, Shinozaki, Shinnosuke, Nishimura, Reon, Oyama, Yoshihiro, Sakuma, Jun.  2017.  Malware Analysis of Imaged Binary Samples by Convolutional Neural Network with Attention Mechanism. Proceedings of the 10th ACM Workshop on Artificial Intelligence and Security. :55–56.

This paper presents a method to extract important byte sequences in malware samples by application of convolutional neural network (CNN) to images converted from binary data. This method, by combining a technique called the attention mechanism into CNN, enables calculation of an "attention map," which shows regions having higher importance for classification in the image. The extracted region with higher importance can provide useful information for human analysts who investigate the functionalities of unknown malware samples. Results of our evaluation experiment using malware dataset show that the proposed method provides higher classification accuracy than a conventional method. Furthermore, analysis of malware samples based on the calculated attention map confirmed that the extracted sequences provide useful information for manual analysis.

2018-01-16
Emura, Keita, Hayashi, Takuya, Kunihiro, Noboru, Sakuma, Jun.  2017.  Mis-operation Resistant Searchable Homomorphic Encryption. Proceedings of the 2017 ACM on Asia Conference on Computer and Communications Security. :215–229.

Let us consider a scenario that a data holder (e.g., a hospital) encrypts a data (e.g., a medical record) which relates a keyword (e.g., a disease name), and sends its ciphertext to a server. We here suppose not only the data but also the keyword should be kept private. A receiver sends a query to the server (e.g., average of body weights of cancer patients). Then, the server performs the homomorphic operation to the ciphertexts of the corresponding medical records, and returns the resultant ciphertext. In this scenario, the server should NOT be allowed to perform the homomorphic operation against ciphertexts associated with different keywords. If such a mis-operation happens, then medical records of different diseases are unexpectedly mixed. However, in the conventional homomorphic encryption, there is no way to prevent such an unexpected homomorphic operation, and this fact may become visible after decrypting a ciphertext, or as the most serious case it might be never detected. To circumvent this problem, in this paper, we propose mis-operation resistant homomorphic encryption, where even if one performs the homomorphic operations against ciphertexts associated with keywords ω' and ω, where ω -ω', the evaluation algorithm detects this fact. Moreover, even if one (intentionally or accidentally) performs the homomorphic operations against such ciphertexts, a ciphertext associated with a random keyword is generated, and the decryption algorithm rejects it. So, the receiver can recognize such a mis-operation happens in the evaluation phase. In addition to mis-operation resistance, we additionally adopt secure search functionality for keywords since it is desirable when one would like to delegate homomorphic operations to a third party. So, we call the proposed primitive mis-operation resistant searchable homomorphic encryption (MR-SHE). We also give our implementation result of inner products of encrypted vectors. In the case when both vectors are encrypted, the running time of the receiver is millisecond order for relatively small-dimensional (e.g., 26) vectors. In the case when one vector is encrypted, the running time of the receiver is approximately 5 msec even for relatively high-dimensional (e.g., 213) vectors.