Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Since Aspire 5.2 it integrates with Large Language Models (GenAI) services like Azure Open AI, among others on premise models to enrich content and allow businesses to leverage Aspire vast set of connectors to power GenAI applications.

Aspire enables understanding and generation of content based on existing business data, for instance



For content generation

Content summarization

  • Summarizing ESG goals from web pages, annual reports, earning calls, news, etc.
  • Summarizing project status, data tables, e-mails, complex documents

Automatic description generation

  • Describe a table/function/database/view based on its schema, data samples and context.

For content understanding

Deep meaning search (vector embeddings)

  • Find the best sentence or paragraph in policies & procedures, FAQs, documentation, help files, web content, etc.
  • Find duplicative content – across web sites, documentation, etc.

Content classification

  • Identifying risky clauses in contracts, find improper language, locate terms & conditions, identify root cause statements, connecting regulatory statements to corporate obligations, etc.
  • Export controls, secret classification, intellectual property, privacy, Material Non-Public Information, etc.

Meaning Enriched Business Identifiers

  • Users can find identifiers with simple names and descriptions
  • Business entities enriched with deep meaning vectors

To learn more about these applications and how to configure them go to Generative AI Components


Models supported

  • Azure Open AI
    • GPT 3.5/4
    • Text Embedding ADA v1 & v2 (text-embedding-ada-002)
    • Customized chat & embeddings models
  • On premise Python models (with Aspire + Python Bridge model)
    • BERT
    • T5
    • GTR-T5
    • MiniLM
    • any other embeddings models

To learn more about these applications and how to configure them go to Generative AI Components