We are the world leader in machine reading solutions
that provide users with action-ready insights from unstructured text in real-time.
How did we arrive here?
Our journey on artificial intelligence (AI) started in 2007 with multi-lingual text. At that time we were extracting meaning out of legal text written in 24 European Languages. We worked on very challenging tasks such as argumentation mining, question answering, natural language inference, legal text summarization, ontology-based information extraction. These challenges enabled us to build scalable solutions that we have today.
Beginning 2019 we were working with a social media agency that wasn’t happy with the social media analytics offerings. After analyzing these offerings, we found that the accuracy provided by these platforms was quite low. We first started by analyzing our own social media accounts by using our AI Platform; the results were impressive. Finally, we started to build up a product from which any user could benefit from this platform.
How does our AI platform work?
Previous offerings in the market perform machine translation into English and they use the English-only pipeline which result in a loss of accuracy due to the machine translation. Unlike this approach, we are using native language models which excels in accuracy. In Sentiment Analysis, we used manually annotated large training sets which yielded more than 92% F1 per language.
As an infrastructure, we benefit from the full power of the underlying Socialays Knowledge Graph, which automatically enriches social comments based on rich in-depth, domain-specific background knowledge and segments conversations into meaningful semantic categories. Our artificial intelligence (AI) driven intelligence platform extracts and transparently summarizes trends in value perception with human accuracy at machine scale.
At Socialays, we are using the cutting-edge tech stack from Natural Language Processing (NLP) to create and define our models. Especially the field of Natural Language Understanding (NLU) is rapidly developing in recent years due to breakthroughs in the design of task specific deep neural networks. It is a rapidly developing field and we keep our team constantly up-to-date.