Our Process

EAGLYS is a strategic partner who turns your data assets into value. Our Company's role is to maintain the data quality data in high resolution, and provide a secure processing environment that can be used safely.
We support solving your limitations around data utilization, for example if you are not allowed to leverage data in an AI project, not allowed to share data across departments or organization, not allowed to manage/analyze data on the cloud, and/or you want to streamline your existing data security policies.

Want to analyze data and introduce AI

Introducing AI would increase sales or improve productivity of business operation, but proceeding along with the data handling limitations is difficult and lose extendability and flexibility in the system…
Currently, each company is putting some efforts on AI projects, but it is said that only a few percent of the cases go to accutual implementation of AI sytem so far. This is largely due to issues such as data sharing, intellectual property protection, cybersecurity and data security (Accenture search, below).
Even if the data could be used in the PoC projects, it is hard to use all of variables or data in the production due to the data security limitations and governance. In our service, we design both the safe use of future data and safe operation of AI, and then proceed as follows.
(Reference: https://www.accenture.com/us-en/insights/industry-x-0/ai-transforms-products)

definition of the purpose

We start by organizing the purpose of AI utilization and the premise required for business utilization (Explanability, accuracy, etc.) . Then, we identify candidates of sensitive data that is expected to be used.

Data assessment and algorithm selection

We do data assesment: analyzing the data space to understand the characteristics of each data. Each variable is evaluated from multiple points of view to determine the best algorithm that best represents the desired AI (oftenl be expected to be highly accurate).

AI Model Design(Secure AI™ Model Design)

Design your model with specific granularity that matches the characteristics of your data. When sensitive data such as personal data or customer information is included, the SecureAI model (AI with secure comupting ) needs to be designed.

PoC Development (demonstration experiment)

Following the blueprint, we implement and validate the model as PoC using DataArmror modules. If the standards for this development are clarified and achieved, we will proceed to this development.

System design and development for actual operation

We design and develop a secure computing system to protect sensitive data and AI model parameters appropriately in order to put the verified model into practical operation. Using secure computing platform DataArmor, we propose efficient ways to manage infrastructure or operation for utilizing data securely.

Maintaining Learning AI Status (tuning and maintenance)

The accuracy of AI changes as the characteristics of data spaces change during operations. To ensure optimal output at all times, we will visualize model learning and update AI parameters and algorithms.

I want to operate the existing AI safely.

We are using sensitive data for AI, but there is a problem in using cloud to protect information, so we are considering how to operate it…
We will proceed with the construction of an environment that enables “AI operation in a secure environment” in the following steps.

definition of the purpose

Identify the functions required and the data to be protected for proper operation of the constructed AI.

Diagnosing learned AI models

In order to properly put the constructed AI into the operational system, we will determine how the existing learned AI model will be secured and how much processing will be generated.

Confidential computation adaptation verification, SecureAI™ (demonstration experiment)

Apply the DataArmror™module to an existing learned AI model so that it can process the encrypted data (SecureAI™ conversion).

System design and development for actual operation

We design and develop a system that can be used and operated while protecting sensitive data and AI model parameters appropriately in order to put verified models into practical operation. If necessary, we will use the secret computing platform DataArmorTM to study and propose ways to reduce the environmental construction and operation costs that were originally required for building a security system.

Effective use of data assets and AI operations

In order to make continuous use of data assets, we support always-accurate and secure AI operation (Visualizing model learning, etc.) while storing and linking multiple data (Integration).

To secure existing business systems and business databases

I want to move to the cloud for a good price/performance, but I am worried about the management involved in the data asset migration and the continuous operation in the cloud…
We will proceed with the construction of an environment that enables “Continuous data asset management in a secure environment” in the following steps.

definition of the purpose

Identify the necessary functions and data to be protected in order to properly manage and operate corporate data assets.

diagnosis of existing systems

We plan to secure the current system in the form of add-on drivers (API) so that data assets can be appropriately accumulated, linked, and managed.

Confidential computation adaptation verification, SecureAI™ (demonstration experiment)

Apply DataArmor™ modules to your current system, build pipelines that can process sensitive data encrypted, and perform test validation.

System design and development for actual operation

In order to put the verified system into practical operation, driver design and application with the current system, and tuning such as speedup is performed as necessary. In some cases, we use the secret computing platform DataArmor™ to study and propose ways to reduce the environmental construction and operation costs that were originally required for building a security system.

Continuous use and scalable operation of data assets

Provide and support an environment that enables continuous, secure, and scalable sharing and collaboration of data across departments and organizations.

If you have anymore questions
just let us know

お問い合わせはこちら