Predicate Proposal - CEO/CEO of

Predicates
#1

CEO

  • Value statement: Knowing the CEO of an organization provides valuable information about the identity of the organization and its management.
  • Definition: “The CEO of an organization is the individual charged with the management of the organization – especially an independent legal entity such as a company or nonprofit institution. The CEO of an organization typically reports to the board of directors.“
  • Tooltip definition of the predicate: “The CEO of an organization”
  • Type of value: Entity
  • For enums - all possible values: N/A
  • # of accepted values: One
  • Inverse Properties and Name: ‘CEO of’
  • Examples of proper use:
  • Examples of improper use:
    • ‘Amazon’ → ‘CEO’ → ‘Jeff Bezos’ citation: Interview: Amazon CEO Jeff Bezos - YouTube. While this is an official video that states Jeff Bezos is the CEO of Amazon, it is now outdated.
    • ‘Reddit’ → ‘CEO of’ → ‘Ellen Pao’ citation: Meet Reddit’s Interim CEO: Ellen Pao. This source states that Ellen Pao has been appointed as ‘interim CEO’, but not as a formal ‘CEO’ whose position will continue indefinitely. An interim CEO should not be considered ‘CEO’ until either 1. their interim status is removed and they are made CEO indefinitely 2. qualifiers enable the nature of the statement to be more specifically described.
  • Usage in other schemas:
  • Constraints: Entity should have value ‘is a’ → ‘Person’
  • Masculine/Feminine Form: N/A
  • Restrictions (ie PII concerns): N/A

CEO of

  • Value statement: Knowing what company this person is CEO of is a valuable contextual information for this person’s identity.
  • Definition: “The organization this individual is charged with the management of – especially an independent legal entity such as a company or nonprofit institution. The CEO of an organization typically reports to the board of directors.“
  • Tooltip definition of the predicate: “Organization a person is CEO of”
  • Type of value: Entity
  • For enums - all possible values: N/A
  • # of accepted values: One
  • Inverse Properties and Name: ‘CEO’
  • Examples of proper use:
  • Examples of improper use:
  • Usage in other schemas:
  • Constraints: Entity should have value ‘is a’ → ‘Organization’
  • Masculine/Feminine Form: N/A
  • Restrictions (ie PII concerns): N/A
#2

Applying to both of these predicates:

  • Estimated cardinality of this predicate: ~2 billion, twice the rough expectation on the number of companies/organizations that will be in Golden’s graph (assuming on average two CEOs per organization are captured in this field, where only one CEO is active but another might be captured with a valid time qualifier).
  • Estimated frequency of new values or changes: Once every 5-10 years - the tenure of a CEO in a company tends to be quite long (source)
#3

@leeds This looks good to me!

One very minor point that may be worth noting, assuming no disagreement: I think an example of Improper use for might be worth adding for the position of Interim CEO.

1 Like
#4

Good point - ive updated the proposal with an example for Ellen Pao’s interim CEO status at reddit. Longer term, I suppose we should look to capture interim CEO’s with some sort of qualifier to note that the “nature of statement” is only considered temporary/interim?

#5

In this case, you can implement a timer for verification, for example: interim CEO from 6 months to 1 year, founder 3-5 years, investor 3-10 years. The introduction of a check timer will keep the knowledge graph up to date.

1 Like
#6

This is a great point. 6 months to 1 year sounds about right to me as well. We will need to fully solve deletion/mutation of triples (this is a non-trivial problem with their being stored on chain) and then we can integrate this into the predicate.

1 Like
#8

When implementing check timing, you need to think about the economics of rewarding such checks. The economy will require serious study, since according to the timing of the check, it will be necessary to check a significant amount of data in 2-3 years, according to my estimate, this is 20-30% of all data