quick inquiry

Semantic Search

AI-driven scalable semantic search solution for university

  • Development time

    14 Months

  • Resources involved

    2 Professionals

Overview

One of the top universities in the US wanted to have a solution for finding the documents which are contextually similar. It can be a useful technique to identify potentially relevant documents based on contextual similarity. The goal of the system is to find related documents which represents similar questions/answers with different content representation.

Business Goal

Extract text segments from text file.

Build semantic search models.

Deploy models to serve users ad hoc queries.

Document Viewer.

Highlight matched text segments.

Reporting on results and export/save documents.

Challenges

There were multiple challenges while developing the solution like,

Handling large amount of documents,

Providing distributed search with second level latency.

Developing custom document viewer for displaying quick result of contextual similar paragraphs.

Solution

System architecture

  • Our engineers have divided this complex problem into sub modules. Major modules include extract text segments from text documents, build and tune semantic search models, searching facility on deployed model and semantic concept highlighting.
  • Distributed processing framework is used to extract paragraph from text documents and output is stored on central storage servers.
  • Semantic search models built from the extracted text segments, tune models, deploy on web server.
  • User select any text segments from web application built in text viewer and search for conceptually similar documents. System has also functionality to highlight resulted text segments, filter criteria, save and export results for further analytics.

Technologies

Implementation

We followed an iterative model approach to implement the solution that included the following phases:

Business requirements analysis.

Architecture design.

Application development.

Setup environment and application deployment.

Road map

Requirement Analysis
  • Brainstorm
  • Innovation
  • Research
  • Resource planning
Architecture Planning
  • Architecture Design
  • System Prototyping
  • Documentation
Development
  • Back end development.
  • Front end development.
Quality Testing
  • Quality Assurance
  • User Acceptance Testing
Deployment
  • Production
  • Evaluation

Business Benefits

Our Big Data, Machine learning development team have delivered elegant solution to the client organization as a result the product users were able to search, categorize, highlight, documents easily through the semantic search process which helped them in analyzing documents similarity, plagiarism and relevancy.

Got idea in mind?

Let's Talk

Contact us for your business needs anytime and our support team will be available to you 24 * 7 to answer your queries.