Computer Knowledge Management Fault Analysis (CHMF)

  ProjectAMPLE
Monday, Apr. 04 2016, 07:10:50 PM
Edited: Saturday, Mar. 23 2024, 09:10:17 AM
Laddice InsightsAnalysis0. Guidelines

Purpose

  • The purpose of this document is to aid in developing a robust note taking system which emphasizes fail safe systems while getting out of the way of humans doing what they are made to do: being creative and connecting ideas to find new and useful insight.

Determinant KM Faults

  • Glitches

    • Interface

      • Frozen

      • Function stops working

    • Server

      • False save

      • Data loss

    • Browser

      • Chrome error

  • Usability issues

    • Function is particularly hard to use/understand

  • Lag

    • Slow savings

    • Slow opening

    • Slow filtering

    • Slow searching

Generative KM Faults

With the advent of AI (Machine Leaning, Language Models, etc.) —> Some computer faults have similarities to human faults:

  • Human Knowledge Management Faults (HKMF)

    • Memory

      • Loss of contextual awareness

    • Unawareness

      • Lacking sufficient training data

    • Errors

      • Lacking sufficient training

      • False interpretation of intent leading to wrong operations

        • Data loss

        • Data pollution

    • Bias

      • Training on biased data

      • Inability to distinguish bias in training data

    • False knowledge

      • Trained with bad data

      • Inability to distinguish truth


 ProjectAMPLE - 6 years, 10 months ago Open
Improvement Options for CHMF Lag

Right now (6/4/2017 - Monish-Mamba) server seems to be the main bottleneck for speed. We have options for increasing server speed since right now we are operating on a free (and slow) server.

  • Use a different form of SQL (CTE) - biggest gains here for graph querying and list view fetching

  • Use a faster server

  • Sharding database

  • More caching

reply