Social Reasoner

PROVA

Probabilistic Reasoning System

PROVA is an Online Probabilistic Reasoning System, where the community can evolve the reasoning for any hypotheses supported by verifiable sources.

Motivation

  • Official media loses its credibility.
  • People are confused.
  • People are looking for truth.
  • “Conspiracy Theories” and “Fake News” become more influential.
  • World is getting divided by polar beliefs and opinions.
  • Everyone thinks they know the Truth.
  • We need a solution.

Solution

The path to the Truth is the Proof

Because a Proof has the greatest power to change people’s beliefs.

The Truth exists but it consists of many pieces that are spread out and not easy to find.

If we collect the pieces and put them in order we will see the path to the Truth: the Proof.

During a time of falling monuments, we need a clear understanding of true values. It’s time for reasoning.

How it works

The user creates a hypothesis supported by at least one citation from a verifiable document. The hypothesis may be a stand-alone or be used as a premise for other hypotheses.

Other users bring new premises (hypotheses, documents) and suggest logical connectives that they believe may support or reject the hypothesis effectively building the reasoning tree. 

Users vote for Validity, Premises and Logical Connectives. The more votes, the more significant the reasoning becomes. 

Based on votes the system calculates the total probability and significance of the hypothesis. 

People can see the Most Probable Reasoning and all alternative views.

Here is more detailed description of PROVA

Applications

Free Open Probabilistic Reasoning Analytical System.

News and Media: The system filters-out fake news and promotes valid information. The Internet articles will include references to the reasoning.

Internet Discussions and Chats: The users can ban other abusive or spamming users by providing valid reasoning.

Community Decision Validator: The reference point for finding motivations for social, legal, political, economic decisions. The decision of government or corporation may be compared with a decision generated by a community using the reasoning system.

Product and services reviews by applying reasoning.

Analogy with Archeology

Someone has found the piece of a pot.>New hypotheses with supporting evidence is created.
Archeologists excavate. Someone finds another piece of a pot. Someone else finds the fragment of a bone.>New premises are added to the hypothesis.
The piece of a pot is a fragment of an ancient vase. The bone belongs to a chicken from the supermarket.>Causal connection by voting and validation.
Dating tells that the vase is from neolithic age. Here used to live ancient people.>Applying logical inference rules by voting.
The reconstructed vase is presented in a museum.>Presentation of reasoning as Article, Graph, etc.
Maybe there is a hidden neolithic village?>Analysis of the proof.

Voting

Users vote for:

  • Validity of Supporting Documents
  • Validity of Hypothesis
  • Conditional Dependency between Hypotheses 
  • Logical Connectives and Inference Rules for Premise

Reasoning

All Hypotheses are connected by conditional dependencies and logical rules into a single graph.

The system calculates the Probability and Significance of all Hypotheses of the graph.

The base math is Bayesian Networks.

Two networks for each reasoning: Proof and Disproof.

Significance, calculated as the Likelihood Ratio.

Reasoning Example

The green arrow shows that the premise hypothesis contributes to the proof and the red to the disproof.
The green hypotheses are True and the red are False. The grey color means low hypothesis significance.
All colors are obtained from the calculation of a complete reasoning tree.

Reasoning Representation

Home

Just after logging-in the user can search the hypotheses.
If the user wants to elaborate on the hypothesis he/she can do it in Workspace.
Ex: Search for “covid” and click on “The decision to close the economy during COVID was right”.

Graph

Graph shows logical connectives and conditional dependencies. Here nodes represent Hypotheses and directed links – the Premises. The branch all the way to the left represents the Most Probable Reasoning Path.

Article

It’s organized according to the Most Probable Reasoning Path in reverse order. Each section is a Hypothesis. By clicking on the title you can see the full text with the best voting comment.

List

Each row corresponds to the Hypothesis with sortable columns of Name, Vote Counters, Validity, Probabilities, and Significance. The advanced users can experiment with reasoning by switching to HypoLab mode.

Pies

Reasoning Script

To obtain a result the System generates a Program Script. The script can be downloaded and independently executed on any computer. It may be used for experimentation and as a certification of validity of the reasoning.

Voting Comments

During the voting the User may enter a comment or select the existing one on either the green or the red side.

Hypo Lab 

Here the user can experiment with reasoning and study how parameters may affect the outcome of the reasoning.

The user may set his own threshold  to accept a belief – the Scepticism of Significance.

The user can set the decay level that discounts older votes in favor of newer ones. The decay with different rates allows the user to see how the probability of the proof changes with time.

The user can play with votes by changing its counters.

The user can set a True/False evidence for any node of the hypothesis.

User Voting Power

User’s Vote increments the specific counter by the User Voting Power value.

The user gains Voting Power based on his total contributions for all proofs.

By creating a Hypothesis or adding new Premise the user gains his Voting Power. The User’s Votes do not impact his Voting Power. 

The more the user’s hypothesis becomes significant – the more the User Voting Power grows.

The User Voting Power makes the system not just vote-driven but Expert-Driven. Based on User Voting Power the crowd selects experts.

PROVA is not just a fantasy

It exists and ready for demonstration. Contact us for a live demo.

Most of its functionality is already implemented and tested.

It runs on Linux in Tomcat. 

Its source code is currently in a private Git repository.

It waits for exposure.

Following Wikipedia

We try to follow the proven success of Wikipedia which is based on truth-leveraging policies and editorial that respects the policies. 

We should adopt the most important Wikipedia’s policies and practices:

  • Verifiable Sources
  • No Original Research
  • Neutral Point Of View

Future

We believe that society should be driven not just by vote-based democracy but also by intelligence.

Democracy may be implemented as an equal opportunity to expose your own reasoning or to support the existing one.

The individuals who contribute more to the reasoning should have a larger impact on society.

Goals

The reasoning is like a path that leads you to the conclusion that then becomes your belief. The main goal of the system is to show the path.

The goal of the project is to create a team that will:

  • bootstrap the system
  • continuously elaborate on it
  • make it popular
  • maintain hosting

Help Us

We are currently the team of only two people. With your help we can grow.

To bring the project to success we need financial help.

YouTube Fundraising Presentation

Donate

Contact Us

Eugene Yakushev and Alexander Yakushev
contact@provareasoner.org
June 27, 2020

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s