Skip to content

A mathematically grounded, reproducible misinformation verification engine based on evidence aggregation: V(C)=f(R,S,N,W)

License

Notifications You must be signed in to change notification settings

gurarpitzz/ARES_POC

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GitHub stars GitHub license

ARES_POC (Autonomous Real-time Evaluation System Proof) — Research Verification Engine

ARES_POC is a mathematically grounded misinformation verification system.
Unlike LLM-based fact checkers, ARES computes truth using a deterministic evidence functional:

$$ Truth'(C) = \frac{\sum_{i=1}^{M} S(C,E_i)\cdot N(C,E_i)\cdot W(E_i)}{\sum_{i=1}^{M} S(C,E_i)} $$

This repository is the Proof-of-Concept implementation accompanying the research manuscript.


What ARES Does

Given a claim C, ARES:

  1. Retrieves evidence passages (R)
  2. Selects Top-M semantically relevant passages (S)
  3. Evaluates entailment / contradiction (N)
  4. Applies source credibility priors (W)
  5. Computes a reproducible truth score

Quick Start

Install dependencies

pip install -r requirements.txt

Web Interface Preview (ARES_UI)

ARES also includes a research-grade visualization interface for inspecting the mathematical verification process.

Overall Interface

ARES UI Overview


Verdict Dashboard

The verdict panel shows the final decision along with the computed Truth Score and Confidence derived from the functional.

Verdict Panel


Mathematical Evidence Trace

Every evidence passage contributing to the verdict is shown with:

  • Semantic Similarity (S(C,E))
  • Entailment Score (N(C,E))
  • Source Weight (W(E))
  • Individual Contribution to Truth'

This ensures full transparency of the decision process.

Evidence Trace


Run single claim verification

py main.py "The Earth orbits the Sun" --k 10 --m 5

Run benchmark evaluation

py evaluate.py --dataset liar --samples 20

File → Mathematical Operator Mapping

File Operator Role
retriever.py R(C) Evidence retrieval
similarity.py S(C,E) Top-M ranking
entailment.py N(C,E) Logical validation
credibility.py W(E) Source prior
verifier.py V(C) Truth functional

Determinism

ARES uses a local cache/ to ensure that the same claim with same parameters always produces the same truth score.


Research Integrity

ARES does not generate answers.
It derives verdicts strictly from discovered evidence.


License

MIT License

About

A mathematically grounded, reproducible misinformation verification engine based on evidence aggregation: V(C)=f(R,S,N,W)

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published