MaaS Effect: NTT Data launches Tsuzumi AI on Microsoft Azure

Cloud is still changing, profoundly in many areas.

Our initial (perhaps muted) reaction to the notion of Software-as-a-Service (SaaS) was quickly and exponentially expanded as we suddenly got used to wider virtualised offerings spanning DataBase-as-a-Service (DBaaS) and a whole universe that now runs through to everything-as-a-Service (XaaS) with Governance-as-a-Service (GaaS) to oversee our now increasingly abstracted approach to enterprise computing.

Naturally then, in a technology landscape now driven so adroitly by Artificial Intelligence (AI), we can think about the implementation of Models-as-a-Service (MaaS).

What are Models-as-a-Service?

The MaaS approach is predominantly cloud-native and is aligned to give software engineers a way of grasping pre-built, pre-trained, pre-tested, pre-configured pre-integrated Machine Learning (ML) models that can be deployed into systems designed to deliver AI functions from sentiment analysis to anomaly detection and natural language processing etc.

All solid enough so far then, but who is working at this level?

Digital business and IT services company NTT Data is now detailing the launch of Tsuzumi Large Language Model (LLM) through the Microsoft Azure AI Models-as-a-Service (MaaS) platform. 

As Shingo Kinoshita (pictured below), executive officer and general manager of the Research & Planning Division at NTT Data said at the time of its launch in 2023, noted, “As LLMs become larger in scale, they require a huge amount of energy for learning. Now, ‘tsuzumi’ leverages its ultra-lightweight nature to pursue reduced learning and inference costs necessary for LLM operation.”

Tsuzumi has capabilities in Japanese and English. It is designed to ease environmental and financial challenges typically associated with LLMs. By adjusting model size without compromising performance, Tsuzumi makes AI technologies, including generative AI, accessible to a range of users and applications.

Operational adaptability

A key feature of Tsuzumi is its operational adaptability. 

The model uses efficient tuning processes using industry adapters for customised knowledge learning. 

This allows Tsuzumi to be relevant and versatile, adjusting to specific use-case requirements and, thereby, reducing the costs associated with service provisioning.

“We are excited to continue our collaboration with Microsoft by integrating Tsuzumi into Microsoft Azure AI. This step not only strengthens our collaboration but also aligns with our commitment to creating a sustainable society,” Yutaka Sasaki, president and CEO at NTT Data Group Corporation.

Initially available in Japan on the Azure MaaS platform, plans are underway to expand Tsuzumi availability to other regions. 

Advancements also are planned in multimodality, which will further enhance Tsuzumi’s sophisticated capabilities and ensure it meets the evolving needs of businesses across the globe.

Data Center
Data Management