Text

Language Identification

Detect one or multiple of over 25 different natural languages in text.

Language Identification engines in the Veritone cognitive engine ecosystem analyze text to pinpoint the language(s) present.

Translation 

Translate text in over 110 different languages and dialects.

Translation engines in the Veritone cognitive engine ecosystem transform written and transcribed text from one language to another.

With specialized machine-learning algorithms for not only languages but dialects and slang, you can dramatically increase the scope of niche translation use cases.

Entity Extraction

Classify entities in text into categories such as people or places.

Entity Extraction — often known as named-entity recognition — engines in the Veritone cognitive engine ecosystem classify words and phrases found in unstructured text into predefined categories, such as people, organizations, and locations.

Entity extraction cognitive engines can also denote the time or date which assists during pre-processing large volumes of data.

Content Classification

Categorize text files and images according to their content.

Content Classification engines in the Veritone cognitive engine ecosystem categorize one or more documents into predefined categories based on what words the text contains.

Keyword Extraction

Identify specific terms and/or phrases in text.

Keyword Extraction engines in the Veritone cognitive engine ecosystem identify the most significant keywords and expressions that appear in one or more documents, based on parts of speech, salience, or other criteria.

Summarization

Generate a summary of all or a selected portion of a text file.

Summarization engines in the Veritone cognitive engine ecosystem generate a condensed version of a text whose length is based on a preset percentage of the original text.

Text Extraction  

Extract unstructured text and express it in a structured format.

Text Extraction engines in the Veritone cognitive engine ecosystem extract text from semi-structured, text-based data such as PDFs or Microsoft Word documents and express that recognized text in a structured format. 

Sentiment Analysis

Discern tone in text to classify it by emotion.

Sentiment Analysis engines in the Veritone cognitive engine ecosystem discern the tone behind a series of written words, which helps you gain an understanding of the attitudes, opinions, and emotions expressed.

These engines classify text according to sentiment or emotion, which depending on the engine, can be a score representing the overall negative, positive, or neutral sentiment for the file and by sentence, or can include a wider breadth of tags such as “happy” or “excited.”

Anomaly Detection 

Locate potential abnormalities in a time series.

Anomaly Detection engines in the Veritone cognitive engine ecosystem assign a confidence value to specific entries in time-series data sets with the goal of predicting which events are anomalous.

Simply ingest time-series data that gives the engine a list of numbers representing certain events and corresponding timestamps of when they occurred. The engine will process this and return the likelihood that any given event is anomalous.