XPRESSO

Enabling next generation cognitive bots using AI,
NLP and deep learning

The Platform Converting Text to Context

Our core platform XPRESSO is a natural language processing and understanding engine that processes unstructured textual information available in any digital format and extracts knowledge relevant for enterprise by converting “text to context”. Context-based sentiment analysis, statement type determination, emotion detection are some of the key capabilities of the engine. XPRESSO has unmatched ability to understand the nuances of human expressions and discover the context from the unstructured text, and the sentiment,associated with the different contexts. XPRESSO combines breakthrough innovation and state-of- the-art technologies and tools to provide accuracy of over 85% in production environment.

XPRESSO uses a combination of lexical parsers and machine learning techniques, leverages domain modelling /ontologies, parts of speech tagging and lemmatisation, etc. to make unstructured data actionable. The key innovation is to represent words and phrases as multi-dimensional (over 300) vectors of numbers. XPRESSO can process such unstructured data at significant scale – we have built state-of-the-art cloud infrastructure that can scale up to meet the requirements.

What it Does

  • Convert Text to Context

    Associate texts to business contexts, from natural language, to make it actionable. Whenever one encounters a text/review, the obvious questions comes to mind – What is the context behind this review? The obvious limitations of manual analysis led to the development of machine learning based contextual understanding from natural conversation

  • Sentiment Detection

    Perceive the essence of a sentence – positive, negative or neutral. Is the text positive or negative or does it have no sentiment at all? Does the text talk something good or does it point out something bad? How will you know if you don't look into the conversation? How can you look by hand if it's millions of comments per hour? Xpresso is equipped to understand all these “Voice of Consumer” deeply.

  • Emotion Detection

    Detect emotional state of the writer – Anger, Disgust, Fear, Joy, Sadness, and Surprise. Xpresso identifies the emotive phrases within a post and aspects with which these emotive phrases are associated. Recognizing the emotion of the text plays a key role in the human-computer interaction.

  • Intent Classification

    Understand latent need of a customer from natural conversation. This will help to create a social lead generation solution to identify new prospect, understand on-demand product features and design personalized marketing communication.

  • Expression Detection

    Capture different ways of human expression like Advocacy, Complaint, Suggestion and Opinion from text. A feedback management system can be built on top of this to identify customer expression in almost real time basis, address different complaints in more organized way and take suggestions/opinion to customize products according to market need

  • Named Entity Recognition

    Extract the name of people, place, company, food etc from text. This helps to understand the primary stakeholder (Brand, Product or any person) of a particular conversation.

Key Features

Not just Key Words

  • Not a key word spotting engine.
  • Understand natural language contextually – "chair" and "furniture" are related
  • "This washer uses a lot of power" (negative), "and has a lot of power" (positive) – contextual distillation

Unsupervised Domain Knowledge

  • Semantic similarity using word vectors (trained using deep learning) for suggesting new contexts/ topic for unclassified data "The miasma was nauseating" gets correctly picked up as a meteorological event.
  • Bathroom, Washroom, Restroom etc have semantic similarity and all of these can be clubbed under Facility, without any prior trained/supervised classification

Flexible

  • Domain and platform agnostic

Rich Domain Concept

  • Domain Modelling is an additional layer of customization on top of generic NLP classification eg: For the sales domain "velocity of sales" is an attribute for sales performance and nowhere related to "velocity" in Physics.
  • Convert entities/ text to domain-specific aspects/ context
  • More than 50 domain knowledge already built in
  • Chicken is related to "Food" for hospitality domain but related to "In-flight services" for airline
  • This is incrementally refined and rebuilt by high precision human annotators.

Continuous Learning System

  • Patented in-house technology to harness user feedback in an automated and a scalable way and become a continuous learner for concepts
  • Unclassified or wrongly classified data in 1st run, can be properly classified in consecutive run, through CLS system
Thank You for connecting with us. We will get back to you within next 24 hours.
  • Follow us on

© 2010 - 2016 Abzooba Inc. |