Skip to content

Latest commit

 

History

History
30 lines (23 loc) · 1.54 KB

README.md

File metadata and controls

30 lines (23 loc) · 1.54 KB

Text Embedding for Question and Answering

Text Embedding:

The embedding process converts text to an N-dimensional vector

Use Cases:
  • Search : where results are ranked by relevance to a query string
  • Clustering : where text strings are grouped by similarity
  • Recommendations : Where items with related text strings are recommended
  • Anomaly Detection : where outliers with little relatedness are identified
  • Diversity Measurement: where similarity distributions are analysed
  • Classification : where text strings are classified by their most similar label

Project Description:

Developed a sophisticated approach to enhance question answering by utilizing text embedding techniques. The project focuses on converting textual information related to start-ups into vectors, subsequently integrating these vectors to add contextual understanding to queries. The central objective was to improve the performance of a query completion model by providing relevant context.

Project Highlights:

Employed advanced text embedding methodologies to transform start-up information into numerical vectors. Integrated these vectors to enrich queries with context, enhancing the query completion model's response accuracy. Implemented document similarity using cosine similarity to identify the most relevant context for a given query. Achieved enhanced performance in question answering through the injected contextual understanding.

Skills Utilized:

  • Text Embedding
  • Natural Language Processing (NLP)
  • Document Similarity
  • OpenAi API
  • Prompt Engineering