AXI

STEALTH PROJECT: Actual_Extra_Intelligence

How smart will humanity become, with the help of AI?

This paper is AXI by aWORDZa.

GOALS

Start AI Company, find people, research/explore exciting linguistic breakthroughs in NLP.

EXAMPLES:

a. Solution to "Junk Food" training data.

b. Jet Fuel for Artificial Intelligence.

c. Open Source AI MIT, with AI SERVICES for:

  1. Illusion Detection Systems

  2. Hallucination Remediation Systems

  3. Actual_Extra_Intelligence (in humans)

STEALTH PROJECT

MIT Open-Source Artificial Intelligence platform.

NetCinematics 2023 (tm) aWORDZa, AXI (tm) tech.

What_exactly?

  1. Illusion_Detection and Confusion_Classification Systems

Business intelligence needs hallucination_detection, remediation

and continued assurance of AI efficacy and correctness.

A SERVICE to review and confirm AI performance.

For any foundation model, of AI "junk food", will hallucinate around those areas.

3rd Party AI TESTS specific to company, across application layer to find hallucinations in fine-tuned data models. Like INFOSEC for AI.

  1. KEY TECH INNOVATIONS:

how to precisely discern:

  • actual reality from actual false premise.

  • called LINGUISTIC_MECHANISMS (as a new science)

  • Massive Unexplored space. Myriad of surprises.

  • New ways to MEASURE CONCEPTS - by WORD COUNT.

    …with linguistic_mechanisms.

    1. EXAMPLES:

      How do we discern "junk food" exactly?

    a. aFalse_or_aTrue ANALYTICS.

    • how can we say ai text is 99% confirmed actual?

    b. Constructive actor or Distruptive actor METRICS.

    • how can we say ai text is 99% trained by bad actor?

    c. Deception_detection_mechanisms

  • how can we label ai text as 100% misinformation?

    > Measurements, never before possible, are now possible.

    And they can be provided AS A SERVICE.


    Measuring AI input and output in creative ways.

    WORDCOUNT and WORDCRAFT, with visual charts.

    Like the stock market, but for ai.

    For any business using ai - in the future.

    Like a new science for the ai era.

    2._Hallucination Remediation Systems

    • Once "Junk Food" is found in ai...

      ... what should a company do?

    a. Track_back "Reflections" to Neural Network.

    b. configure vast FILTER_ARRAYS. But also...

    c. Use each instance, to inform business_intelligence.

    d. Inform CONFUSION_CLASSIFICATION taxonomy. AND

    e. eventual, Ontology of all_human_awareness.

    Measurements, never before possible, now possible.

    Provided AS A SERVICE. Via ai reflections.

    USE Language Mechanisms to exactly define confusions.

    If we can know where confusions exists in humans,

    then we know exactly where hallucinations exist in

    training data, and downstream neural networks.

    If we do not know how humans are actually confused,

    how can we know if training data is "junk" or "treasure"?

    NOT an IMPOSSIBLE_PROBLEM, with ai.

    ai test_and_filter_systems on INPUTS and OUTPUTS.

    • exactly by COUNTING_WORDS, and by WORDCRAFT_CLASSIFICATION.

    • All_Human_CONCEPT classification systems.


      These are the tools use to arrive at actual_extra_intelligence.

    • WORD_MATHEMATICS (index)

    • SOCIAL_PHRASES

    • CONFUSION_TAXONOMY

    • Illusion_Remediation_Ontology

    • 3D_CONCEPTUAL_MAPPING

    • CONCEPTS_BEYOND_US, or CONCEPTS_NOT_YET_ARTICULATED,

      as CONCEPTS_BENEATH_WORDS.

      Automated by Naming Convention and Syntax.

      Looking for good people. Please let them know.

      Thank you,

      NetCinematics. 11/11/2023