This competition captures

We develop and use intellectual systems that use artificial intelligence, capable to extract knowledge and forecast. It appears that current prototypes look much clever than some congressmen because prototypes can use basic knowledge and able to adapt. This competition captures.

Technology

Using artificial intelligence, that allows extract knowledge, gives possibility to process any data, extract important information, find logical and time relations and predict events and values using different methods including text analysis. Using natural language creates intuitively clear method of communication.

Usage

Taking into consideration that information exists and updated mostly in symbolic text format, mentioned above approach together with other methods of machine learning open possibilities to create and naturally inject digital assistants, advisors and intellectual systems in any business process. Examples of digital assistants: - intelligent news and events aggregation, search and following; - marketing assistant discovering groups, intents and feedbacks; - law assistants checking validity of contract and even preparation of legislation; - digital advisor helping find and explain financial instruments price dynamic in relation with events and other factors; - intellectual systems ensuring decision making in situation centers, emergency or commandment centers; - etc.

A winning gene template of a digital assistant

Input : Any data; Machine : An online semantic encoder, an online (adaptive) learning machine; Output : a natutal language inference, a summarization.

Hide less important details

AI in news and events processing is very productive. You will have enough time for a meditation. 1. Just use natural language for descriptions of any information you need. 2. Set up an accuracy of infomation selection or let assistant to adjust it adaptively.

Some projects state

We move prototypes to working services : intellectual news and events aggregator and financial decision-making assistant (no chart painting and technical analysis was injured) with training completed on 2 years news and events data.