On the 7th, Datamaker announced that it will officially showcase its MLOps platform, 'Datamaker Synapse' (datamaker synapse), at the 'AI SUMMIT 2022' held at COEX in Seoul.
According to the company, the annotator of 'Datamaker Synapse' is labeling software equipped with an AI-based smart labeling UI (user interface) designed to improve work efficiency.
Most existing MLOps platforms are developer-only tools, making accessibility very low for non-developers who lack development knowledge, while MLOps platforms for non-developers often provide only limited functions.
MLOps platforms typically only have machine learning management functions and do not include an annotator, so users must use multiple services in combination, making it difficult to build automated learning pipelines. Therefore, Datamaker Synapse provides systems in No, Low, and High Code formats that can be professionally used by both developers and non-developers, allowing all necessary functions for AI development through dataset management, annotator, and workspace.
A company representative said, "By utilizing the Datamaker SDK (software development kit), users can easily integrate Synapse's functions into applications they develop," adding, "Our own storage and GPU servers can also be linked with Synapse, enabling custom development, excellent scalability, and solution use in closed network environments."
Datamaker also stated that, based on its accumulated project operation know-how, it has equipped a workspace optimized for data labeling.
The workspace of 'Datamaker Synapse' is advantageous for building high-quality data by enabling automatic data validation for data quality management and detailed inspection by area. The tutorial function, which automatically scores workers' outputs by registering labeling answer templates, is also useful in the process of selecting multiple workers.
One of the strengths of 'Datamaker Synapse' is that all areas can be easily customized based on the core annotator provided by Datamaker through the use of plugins. Functions required by client companies can be added, and beyond annotator customization, plugins can also be used for artificial neural networks and custom scripts.
Enok, CEO of Datamaker, said, "The reason we started the data labeling service was due to the difficulties AI developers face in acquiring data," adding, "Now, we have launched a world-leading level solution to also resolve the difficulties in training and managing models."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
