東芝情報システム Official site

Next-generation cryptographic library compatible with quantum-resistant cryptography, preparing for the era of quantum computers: 'Quantum Safe Crypto Library'

■What is Post-Quantum Cryptography (PQC)? PQC is a public key cryptography method based on mathematical problems that are difficult to decipher even with the high-speed calculations of quantum computers. Traditional RSA and ECC are expected to have reduced security around 2030, and the National Institute of Standards and Technology (NIST) and the National Security Agency (NSA) recommend transitioning to PQC. ■What is the Quantum Safe Crypto Library? It is a cryptographic library for embedded systems that implements PQC-compatible cryptographic algorithms. It provides high security in anticipation of the quantum computer era. ■Main Features - Supports NIST standardized algorithms (ML-KEM, ML-DSA, LMS/XMSS) - Can be integrated with IPsec/TLS (IPsec uses the Quantum Safe IPsec Toolkit, TLS works with OpenSSL) - Plans to apply for FIPS 140-3 certification by the end of 2025 ■Future Developments We plan to gradually expand support for SLH-DSA (SPHINCS+) and FN-DSA (Falcon), among others.

Audio, only the necessary content for those who need it. A new audio sharing experience realized with Auracast.

■ Auracast is a new standard for audio broadcasting based on Bluetooth technology. It is a technology developed by the Bluetooth SIG (Special Interest Group) and has the following features and advantages: 1. Broadcasting Real-time delivery of designated audio in public spaces such as airports and museums. The LC3 codec allows for high-quality simultaneous delivery to multiple headphones. 2. Multi-channel Provides multiple audio channels, allowing users to select the audio they want to hear. Supports multilingual guidance and switching commentary for sports events. 3. Low latency Minimizes audio delay between devices, ensuring accurate synchronization.

We automatically create a large amount of high-precision training data in a short period using an annotation automation tool!

■What is Annotation in AI Development Annotation refers to the process of adding meaning (tags/labels) to the data that you want to train the AI on, thereby creating correct data. This serves as the foundational data for the AI model to learn from and is an essential step to ensure the model learns correctly through machine learning. The quality of the annotation significantly affects the performance of the AI model, making it an important task. ■Benefits of Automating Annotation <Cost Reduction> Creating a large amount of training data requires a lot of manpower. By automating this process, labor costs associated with manual work can be reduced. The more data you handle, the greater the cost-saving effect. <Efficiency in AI Development> By automating the creation of training data, you can prepare a large amount of high-quality learning data with consistent quality in a short period, ultimately contributing to the efficiency of AI development. ■Three Features of Automated Annotation Tools and Services - Automatically create high-precision training data - Support automation of semantic segmentation and bounding boxes - Significantly reduce the time for automatic generation, achieving world-class accuracy for training data