Launching the "AI BANTO" service, a solution utilizing large language models (LLM).

Our company, in collaboration with JDSC, has developed a new solution called "AI BANTO," utilizing large language models (LLM) (*1), to streamline interactions with clients, facilitate contract searches, and simplify regulations and research tasks. The service will launch on October 1, 2024. "AI BANTO" leverages AWS generative AI services, enabling cross-referencing and responses to a wide range of documents, such as maritime industry-specific contracts, technical information, regulatory details, FAQs, and emails. This significantly reduces the time employees spend handling inquiries and allows for responses to specialized and multilingual inquiries.

(*1) A large language model is an AI system capable of understanding and generating human language by processing vast amounts of text data.

"AI BANTO" to be a new solution of future of maritime industries.

■ Features of "AI BANTO"

 "AI BANTO" addresses the various challenges that arise in ship operation management by investigating issues based on internal documents and emails, using Retrieval-Augmented Generation (RAG) (*2) technology. It provides functions such as summarization, responses, and email draft creation through LLM.

In addition, it enables less experienced employees to handle highly specialized tasks and multilingual responses that previously required the expertise of experienced professionals. This contributes to the standardization of operations, broadens the range of personnel utilization, facilitates the transfer of operational methods to successors, and promotes knowledge sharing, thereby improving the work environment.

■ Mechanism of Large Language Models (LLM)

 Shipowners manage and operate their vessels based on general knowledge of ship operation management.

 However, there are times when specific expertise is required, such as dealing with ship issues, legal procedures, or charter party disputes.

 In such cases, shipowners typically consult with specialists, investigate past cases or legal documents, and find solutions based on the knowledge and information obtained.

 As a substitute for these specialists, LLMs can answer various questions posed by shipowners.

 However, to provide reliable answers with citations, the model also requires assistance in conducting research.

 This assistance comes in the form of a process known as Retrieval-Augmented Generation (RAG).

 Using the information obtained through RAG, the LLM generates human-like, fluent conversations in natural language.

 (*2) Retrieval-Augmented Generation (RAG) is a technology that enhances the accuracy and reliability of generative AI models by incorporating information from external sources.

■ Problem Solving with "AI BANTO"

Challenges Facing the Maritime Industry:

1. The complexity of obtaining information due to labor shortages

2. Knowledge management and transfer

3. Dealing with complex tasks (understanding not only international conventions but also national regulations)

Information Management and AI Chat for Problem Solving

1. AI-driven information search and organization contribute to improving work efficiency.

2. The RAG process allows for unified storage of internal and external information, uncovering hidden knowledge and data.

3. LLM provides rapid information access and multilingual support.

Benefits of Introducing "AI BANTO"

- Increased Work Efficiency : Minimize the time spent on research.

- Accessibility : Easily connect from anywhere, even while on the go.

- Up-to-date Information : Constant updates on the latest general maritime information, turning it into a data source.

==For Those Considering the Service==

If you wish to try the service or are considering immediate use, please feel free to contact us at the address below.

AI BANTO Support: [e-mail]

==For Inquiries About This Article==

Orient Marine Co., Ltd. 

DX Strategy Planning Division / Marine Engineering Division 

Contact: https://www.toyosenpaku.com/en/contact/