A recent press release reports, “Gopher Protocol Inc., a company specializing in the development of Internet of Things (IoT) and Artificial Intelligence enabled mobile technologies, including a global mesh network technology platform for both mobile and fixed solutions, announced today it released Avant! AI new version according to schedule. The new release includes user authentication, history saving and a new sentence embedding method which is a derivative of the InferSent method. The InferSent method is a sentence embedding technique that provides semantic sentence representations. It is trained on natural language inference (NLI) data, which means asking questions in natural language. Modern NLP (Natural Language Processing) systems rely on word embedding as base feature to capture relationships among multiple words and phrases in a single vector.”
The release goes on, “Gopher’s Avant! AI’s new sentence embedding system encapsulates advanced, proprietary methods in order to extract text and ‘understand’ its meaning. The new system includes general-purpose sentence embedding and captures generic information that is useful for a broad set of question types. Using this system, Avant! will be able to “understand” the meaning of general sentences (Questions) in natural language, faster and with higher accuracy. Equipped with this new semantic sentence encoder, Avant! will be able to search for related texts, perform faster identification of text chunks and about its meanings in order to provide the user faster and more accurate answers. Avant! will be able to check for sentences similarities under supervised training and skip through parts that are irrelevant. The end result is a robust NLI (Natural Language Interface) for question presentations and accurate/effective answers in virtually real-time.”
Read more at Nasdaq.
Image used under license from Shutterstock.com