| Cryptography |
|
Generic Construction of Key-Policy Attribute-Based Encryption from Public-Key Searchable Encryption
|
|
It is known that public-key encryption with keyword search (PEKS) can be constructed from attribute-based encryption (ABE), but the opposite is unproved. In this paper, we propose a generic construction of key-policy ABE from PEKS.
|
|
Continual Auxiliary Leakage Resilient Signatures
|
|
Leakage-resilient cryptography is practically desirable since it guarantees that even if the secret-keys are partially leaked, the security remains. We propose the first secure constructions of digital signature, identity-based signature, and attribute-based signature, resilient to the continual auxiliary leakage.
|
| Machine Learning |
|
Adversarial examples generation for black-box classifier with less number of queries
|
|
Deep learning has achieved high accuracy in many kind of tasks e.g. image recognition. At the same time, it is possible to cheat a classifier only by adding slight perturbations to an input data. This is called adversarial examples. In this research, we propose a method to generate adversarial examples for black-box classifier with less number of queries.
|
| Distributed ledger |
|
Security proof of POW-type blockchain
|
|
Blockchain is a distributed ledger technology, which has a potential to construct multiparty consensus protocols without a trusted third party. This technology is spreading to apply such as cryptocurrencies and others. Bitcoin adopts a blockchain with proof-of-work (POW). We introduce
a security proof of such blockchain, and resulting implications.
|
| Information Lifecycle Control Employing Blockchain and Secret Sharing |
|
Many of our communication tools are not resilient to censorship: when a server becomes a target of censorship, it can be an application-level privacy threat to users. In this research, combining public blockchain and secret sharing, we propose an auto deleting application that is resilient to censorship.
|
| Control system security |
|
Improvement of Anomaly Detection Performance using Packet Flow Regularity in Industrial Control Network
|
|
Cyber attacks against Industrial Control Systems have increased. This research demonstrates the improvement of detection performance using packet flow regularity in industrial control networks. In particular, we propose the anomaly detection method with word2vec utilized in natural language processing.
|
|