Symilarity is a software platform that uses unsupervised deep learning algorithms to iteratively learn language and concepts directly from source data i.e. without being told what to look for, or where. Symilarity transforms unstructured and semi structured data, images, sentiment and moods into structured concepts without human intervention.
Symilarity performs real-time simultaneous complex queries over multiple corpora all containing unstructured or a mix of disparate data. This data can be unstructured text, incoming high velocity IoT, images, conversational communication (Chatbots) or any other source of data.
Other than ease of use what sets Symilarity apart from other products includes:
• Configuration. No configuration is required but the platform retains the capability for bespoke configuration
• Languages, Symilarity is multi-lingual
• Hosting: Symilarity was designed to be hosted either in the cloud or on-premises
• Data Ingestion: Symilarity was designed ingest data in a variety of simple combinations including TXT, PDF, WORD and XML:
• Auto-upload: Symilarity will load documents from a shared folder, on a scheduled basis in near real-time and batch
• Use Cases: Symilarity was designed to support three types of use case:
Matching: Symilarity uses machine learning techniques to provide improved search results over traditional free text search. Indexes can be built on a variety of fields including fields extracted through NLP.
Discovery: Symilarity provides a range of tools such as clustering, word frequency analysis, automated classification, alongside machine learning, to enable users to find hidden content within large corpora.
Search: Symilarity enables simultaneous search over multiple corpora. It provides synonym/hyponym/hypernym search features and NLP functionality to extract references to location, enabling corpora to be searched spatially.