Chief scientist of major corporation can’t handle criticism of the work he hypes

Re:Chief scientist of major corporation can’t handle critici... by dekhn on Wednesday November 23, 2022 @11:07AM (#33720949)

The points you are making are entirely political and apply generally, not just to insurance. Insurance is a business and it's not the job of the insurer to correct long-standing social ills.

[ Parent | on HN ]

Re:Chief scientist of major corporation can’t handle critici... by PuppyTailWags on Wednesday November 23, 2022 @11:21AM (#33721135)

Whether or not my points apply generally does not mean it does not apply to insurance. In fact, it very likely means insurance is a subset of generally, and therefore insurance also has to deal with racism. You cannot extricate models of society from their ills by pretending the ills aren't relevant.

[ Parent | on HN ]

Re:Chief scientist of major corporation can’t handle critici... by mrguyorama on Wednesday November 23, 2022 @01:06PM (#33722511)

It's not political to point out bad data will make bad stats.

[ Parent | on HN ]

Re:Chief scientist of major corporation can’t handle critici... by dekhn on Wednesday November 23, 2022 @01:43PM (#33722996)

Is it bad data if it reflects the underlying observed frequencies regardless of their social origins? I don't think so.

[ Parent | on HN ]

Re:Chief scientist of major corporation can’t handle critici... by mcguire on Wednesday November 23, 2022 @03:10PM (#33723932)

"But our business processes avoid all bias; all decisions are based on a ML model, not on human decisions."

[ Parent | on HN ]