RE: Closed Loop Feedback

 Sharon ,

 

Thanks for clarification . I understand that merely clicking on any STAR * , would not be sufficient feedback to my AI , to get “ trained “

 

I am considering providing following text ( Feedback Statement )  on U /  I  :

 

 

Totally Irrelevant

 

Somewhat Relevant  

 

Quite Relevant

 

Excellent and Precise

 

Insightful

 

 

 

As soon as a visitor ticks inside square block against any FEEDBACK  STATEMENT , that will get forwarded to My AI , along with the concerned QUESTION ( submitted ) – and the ANSWER ( that got rated )

 

Pl advise if this will help

 

hemen

 

From: Sharon Zhang [mailto:sharon@personal.ai]
Sent: 06 November 2023 19:56
To: Suman Kanuganti
Cc: Hemen Parekh; Kishan Kokal
Subject: Re: Closed Loop Feedback

 

Hemen - 

 

Currently reinforcement learning happens when your team corrects the response that are scored low / unsatisfactory and feed it back to the API with the correct response. The ratings themselves does not do anything to help with further learning. I would recommend building this layer yourself and leveraging memory api to feed in the corrected data.

 

Cheers,

Sharon

 

 

 

Sharon Zhang CTO & Co-founder, Personal AI 617-909-1053  |  sharon@personal.ai Personal.ai - I Remember   Linkedin • Twitter • Instagram • Medium

 

 

 

 

On Mon, Nov 06, 2023 at 6:19 AM, Suman Kanuganti <s@personal.ai> wrote:

Hemen, let me touch base with team on this. 

 

On Mon, Nov 6, 2023 at 3:47 AM Hemen Parekh <hcp@recruitguru.com> wrote:

Suman,

 

Sometime back , I urged you to integrate “ Rate My Answer /  * * * * *  “  on  www.hemenparekh.ai  , into personal ai API , so that feedback of visitors could TRAIN my ai

 

You wrote back , saying ( something ) .. “ I will put it on sprint … “

 

I wonder if this could be done before my Video Meet with Marrisa on Friday ,  10^th

 

 

hemen