FeasibilityLens

FeasibilityLens : AI-driven agent assessing research feasibility by analyzing theory, practicality, and computational constraints with RAG

Overview

FeasibilityLens is an AI-powered agent designed to assess the feasibility of implementing novel approaches described in research papers. It leverages Retrieval-Augmented Generation (RAG) to extract relevant methodologies from related papers and uses a Large Language Model (LLM) to evaluate the feasibility and novelty of a given research methodology.

Features

Automated Feasibility Assessment: Analyzes research methodologies for feasibility, reproducibility, and improvement over existing approaches.

RAG-based Information Retrieval: Retrieves relevant research papers and extracts key methodological details.

LLM-powered Evaluation: Uses a Large Language Model (e.g., GPT-4) to assess feasibility and provide structured feedback.

Vector Database & Similarity Search: Stores research embeddings and performs efficient similarity searches using Facebook AI Similarity Search (FAISS).

Iterative Refinement: Allows researchers to refine methodologies based on AI-generated insights.

Workflow

• Input Research Papers: User provides the current research paper and related papers for comparison.
• RAG-based Retrieval & Extraction:
  • Embeds and retrieves related papers using T5.
  • Stores embeddings in a vector database (FAISS) for efficient similarity searches.

Feasibility Analysis via LLM:

Evaluates the methodology for feasibility, practicality, and improvements.

Compares with retrieved methodologies to assess novelty.

Results:

✅ Feasible & Novel

⚠️ Feasible but Not Novel

❌ Not Feasible

Technical Details

Technologies Used:

RAG for information retrieval (using T5 for embedding)

Vector Database (FAISS) for similarity searches

LLM for feasibility evaluation (GPT-4 or similar models)

Implementation Stack:

Python

FAISS for efficient vector search and retrieval