AI Factory 2.0: Accelerated Productization of Text Analytics Engines with No-code Machine Learning

AI Factory 2.0

[Romania, 20 Sep 2023] Zetta Cloud (ZA Cloud SRL) is announcing a major release of its flagship “no code” Machine Learning platform AI Factory, version 2.0. The new release aims at bridging the gap from data to production for Text Analytics AI engines, with unprecedented ease of use designed for both machine learning professionals and non-technical business users.

AI Factory is an end-to-end platform for building and deploying AI Text Analytics engines, in a secure manner. The solution consists of “Factory No-Code” machine learning dashboard that allows business users to create or adapt NLP engines for Classification and Named Entity (NER) tasks and deploy them into production as Docker containers, and “IntelliDockers“, a suite of prebuilt and adaptable out-of-the-box Text Analytics Docker containers able to perform multiple tasks across more than 40 languages. The solution provides AI tasks such as language detection, sentiment analysis, classification, summarization, named entity recognition, OCR, semantic comparison, or topic clustering.

Major functionality improvements are notable both in the Machine Learning dashboard (Factory No-Code), with features designed for ease of use and productivity, and the production-ready text analytics engines (IntelliDockers).

What is new in AI Factory 2.0

On the “no-code” machine learning component, an emphasis was placed on making the entire process as easy as possible while increasing the quality output of the generated AI engines: hyperparameter optimization for automatically selecting the best settings for an ML process, adaptable classifiers, for fine-tuning pre-trained Transformer models, multi-label classification trainer capability, fine-tuning of cross-language (XLU) models by transferring knowledge from one language to others, and several UX improvements.

On the AI production engines side, the most notable improvement is the hardware optimization for commodity hardware, with engines now capable of reaching throughput of millions of words per minute on CPU-only machines. A new Machine Translation engine capable of processing more than 50 languages, in any combination, was added to the existing portfolio of 150+ NLP engines, and SSL support, API key authentication, web interfaces, and new supported input formats were added to the production Docker containers.

Read the full Product Release Notes: https://zettacloud.ai/ai-factory-2-0-accelerated-productization-of-text-analytics-engines-with-no-code-machine-learning/

Emil Ștețco, founder and CEO:
“With Factory 2.0 we continued to implement our Democratizing AI manifesto and significantly improve the capabilities and user-friendly approach of our no-code Machine Learning platform. Even though AI continues to become an important topic for most companies, adoption into production environments has not progressed according to expectations. The main reason for this is that the end-users of AI systems, the company business users, are still too disconnected from current technology offers that do not provide the productization levels, the security, or the ease of use required. We have embarked on a journey to make AI accessible to users on their own terms, in their infrastructure, for their own data and use cases, and most importantly, to make it accessible to non-technical users. With the latest AI Factory release we are closer to this goal, as creating machine learning engines for document-oriented process automation has never been easier.”

About Zetta Cloud

Zetta Cloud is an Artificial Intelligence R&D software company founded in 2013 in Cluj-Napoca, Romania specializing in end-to-end adaptable and secure Text Analytics solutions. The company is addressing use cases such as data triage, intelligent process automation, knowledge management, and open source intelligence, targeting mostly enterprise customers from regulated industries and the public sector. Its current technology stack consists of AI Factory, a leading no-code machine learning platform designed for secure private cloud and on-premise deployments, and TrustServista, a Google-funded solution for monitoring and detecting online disinformation with the use of AI.

For further information please contact
George Bara
Chief Strategist
george@zettacloud.ai
+40.752.20.0344

www.zettacloud.ai
ZA Cloud SRL