[Quebec, Canada & Cluj, Romania] [December 2019] Artificial intelligence innovators Epsilonsys (Canada) and Zetta Cloud (Romania) join forces to launch state-of-the-art multilingual conversational AI solution “Question Answering Machine (QAM)”.
QAM is built using state-of-the-art Artificial Intelligence technology and can answer a question based on a given context and answering with a segment from the context. It operates on more than 90 languages and could be used for answering random questions against any of your documentation or knowledge base, including: English, German, Romanian, Italian, Polish, Hungarian, Arabic, Russian, Chinese, Persian and other.
Epsilonsys (http://www.epsilonsys.com/) is a Quebec-based AI company that builds real-world applications of AI technology and has over 10 years of experience in development and integration of Intelligent Systems, IoT, Security, ERP and applied Customer Insight solutions, backed up by an experienced team of Natural Language Processing research team. The collaboration with Zetta Cloud (www.zettacloud.ro), based in Romania and already having a mature solution portfolio of in-house built Natural Language Understanding solutions, is geared toward the productizing and distributing QAM to a wider EMEA market.
“The agility and market insight that Zetta Cloud has in the European market has compelled Epsilonsys to partner for the launch of the Question Answering Machine (QAM) outside of the NASA region. The goal of this partnership is to deliver an easy-to-use, easy-to-integrate and value-added AI solution with the help of Zetta Cloud. The research and development investment from both companies resulted in a unique solution that makes use of Google BERT approach to provide a high accuracy question and answer solution.”Cristian Liciu, Epsilonsys CEO
Targeted Markets and Use-Cases
Question Answering Machine (QAM) is designed to be deployed on private cloud or on-premise environments and support intelligent discovery and customer support projects for virtually any organization that wants to shorten analysis processes, make full use of available information and communicate with any database or data repository in a natural way.
The main use-cases that QAM will affect the most are:
- Enterprise Search: QAM can be connected to virtually any data source and replace the classic search processes, which can find relevant information for a search query but still rely on the user to select the correct response and select the actual useful information, with a more natural “question and answer” approach that provides precise responses to a specific question.
- Conversational chatbots: Even the most advanced chatbots still rely on predefined business rules and processes to perform certain tasks or fetch required information from a user inquiry. QAM can eliminate the need of strongly structured knowledge bases and business rules that define every single potential answer to a question, virtually allowing customer support systems to support any chat scenario.
How does QM work?
Question Answering Machine (QAM) implements the Google BERT (Bidirectional Encoder Representations from Transformers) approach, the complex technology used by Google Search to improve the results of complicated queries that require context. QAM builds on BERT with a focus on question and answering capabilities and delivered as ready-to-use software suitable for integration into any enterprise IT ecosystem.
QAM will be made available alongside Zetta Cloud’s IntelliDockers Natural Language Understanding (NLU) product suite (www.intellidockers.com) that already cover a number of capabilities delivered as Docker containers covering a range of languages: automatic summarization, language detection, named entity extraction, automated classification, sentiment analysis and semantic comparison.
Question Answering Machine (QAM) can be tested at https://www.intellidockers.com/#answers
Examples of QAM in action
For further information please contact:
Emil Stetco (CEO, Zetta Cloud)
Mobile: + 40.723.175.635