Hume Platform- Graph-powered Insights Engine | NLP

The Hume platform is an NLP-focused, Graph-Powered Insights Engine.

It creates a digital twin of your business in the form of a Collaborative Knowledge Graph which surfaces critical but previously buried and undetected relevance in your organization.

This newly accessible relevance can be surfaced and used in a variety of ways as shown below.

Business User Capabilities

• Conceptual / Predictive Search
• -Combine Hume’s Collaborative Knowledge Graph with Elasticsearch to search your data more effectively.
• -Search API, Elasticsearch Integration plugins, Auto-suggest/Autocomplete
• Dynamic Chat and Conversation
• -Dynamic translation from natural language to Cypher for dynamic answers
• -Open-domain Question Answering over knowledge graph
• Q&A over text docs
• Dashboards (custom delivery)
  • -Combine filtering and search
  • -Context aware graph viz & exploration, geo dashboards

Key Features

• Collaborative Knowledge Graph (CKG)
• Robust NLP/NLU engine (up to 99% precision, 90% recall)
• Part of Speech tagging
• Dependency Parsing
• Named Entity Recognition
• Co-Reference
• First production usable TextRank algorithm
• Keyword & Concept Extraction (using TextRank)
• Sentiment Analysis
• Enrichment
• Multi-language doc & concept comparison/analysis
• Ontology support
• Using existing customer ontologies (or from scratch using Hume.Labs)
• Entity Linking & Disambiguation to leverage other KBs more effectively
• Routing/Alerting/Notifications/Doc Classification
• Management API and UI (including schemas)
• Ingestion / dsource transform & mapping to schema
• REST API
• Indexing (for contextual info retrieval) combo of Graph & Elasticsearch
• Dynamic search/retrieve from graph
• Generating more connectedness using other KBs (WikiData, ConceptNet, internal KBs, taxonomies etc)
• Image recognition / processing / enrichment

• Word Embeddings (Word2Vec)
• Doc Embeddings (Doc2Vec)
• Entity Relationships extraction (RNN, CNN, SVM)
• Classification based Doc2Vec & K-means
• Intent based on Bayesian and others
• Text Summarization
• Text Clustering (LDA & LDA over Apache Spark)
• Entity Linking / Disambiguation

Scalable & highly-available

• Infinitely scalable NLP engine & HA Knowledge Graph

Hume.Labs

• Map domain so Hume understands the language of your enterprise