Thad, All,

Thank you for taking a look at the proposal.

Quality assurance is an important topic for the proposed project, including and beyond vandalism detection.

As envisioned, each question can have multiple answers and each answer can be supported by multiple justifications. Perhaps users (and eventually AI systems) will be able to indicate that particular answers were, in their opinions, either of quality or in need of attention. Perhaps there can be solutions devised – and perhaps beyond upvoting and downvoting – for sorting candidate answers for presentation and display purposes (this is on the way to automated decision-making systems, automated debate judges, which can be developed and trained using wiki-based resources such as Wikianswers).

Then to the matter of whether users will choose to create, edit, and work with justifications to answers. I go into some detail in the proposal about emergent benefits and features were they to.

If we get the collaboration design and related features right – and there are some details remaining to be considered – the integration of multiple AI question-answering systems (these can be domain-specific) with a wiki question-and-answers platform should become increasingly performant and useful. Crowdsourced, or social, question-and-answer platforms have been previously proven. The cool new features come from the integration of AI technologies such as question-answering systems.

Also, what might you think about the computer-aided document-authoring ideas (https://meta.wikimedia.org/wiki/Wikianswers#Technical_discussion)? These MediaWiki software feature ideas are applicable to Wikifunctions in that extensible server-side document-processing tools can include parsers and compilers for each supported programming language; the outputs of these document-processing tools can be represented as annotations; and there are also user-interface topics to consider pertaining to displaying these annotations while users edit, preview, view, and view more information about wiki documents.

These same software features would be increasingly useful for other wiki projects – including the proposed Wikianswers project – though the server-side document-processing tools would be, instead of parsers and compilers, natural-language processing tools and systems (e.g., spellchecking, grammar checking, readability analysis, sentiment analysis, analysis of subjectivity and objectivity, fact-checking, reasoning checking, dependency graph analysis, argument verification and validation, and automated decision-making systems).

Best regards, Adam

From: Thad Guidrymailto:thadguidry@gmail.com Sent: Monday, December 27, 2021 10:16 AM To: General public mailing list for the discussion of Abstract Wikipedia and Wikifunctionsmailto:abstract-wikipedia@lists.wikimedia.org Subject: [Abstract-wikipedia] Re: Wikianswers

Cool, but let's have fun with a quality perspective, and the reality of needs sometimes...

Girl: <>I need an answer for why birds have hard beaks. Wikianswers: <blank> - We know all birds have hard beaks, but do not know the reason why. Perhaps you know the reason and can type this in? Boy: <typing> Because... Mom said... God made them that way. Wikianswers: Thanks! Boy: <puzzled and still no answer> OK, let me try Google. I personally think folks appreciate much more when they don't have to participate so much when they want answers, but often are quite content to clarify their questions to get a quality answer. The reverse is not often true.

Thad https://www.linkedin.com/in/thadguidry/ https://calendly.com/thadguidry/

On Mon, Dec 27, 2021 at 2:20 AM Adam Sobieski <adamsobieski@hotmail.commailto:adamsobieski@hotmail.com> wrote: Abstract Wikipedia,

Hello. I am pleased to share a new project proposal: Wikianswers (https://meta.wikimedia.org/wiki/Wikianswers).

Like similar projects, users will be able to receive one or more answers to their questions from artificial intelligence question-answering systems, but, unlike other projects, they will subsequently be able to edit these answers and any provided justifications using wiki technology. User-corrected content could later be utilized to retrain consulted question-answering systems, resulting in continual improvement.

The proposed project would greatly benefit end-users and would create a valuable resource for the training of artificial intelligence systems.

As broached in the proposal’s technical discussion section (https://meta.wikimedia.org/wiki/Wikianswers#Technical_discussion), components and features which could add value to Wikianswers could also add value to Wikifunctions: server-side document-processing, annotations, and client-side user interface topics. With these components and features, in a manner resembling an IDE, users could receive informational, warning, and error messages when working with source code, e.g., Python, in a wiki document context.

Thank you for any ideas, comments, questions, or suggestions with which to improve the project proposal.

Thank you for taking the time to express your support for or opposition to the newly proposed project.

Best regards, Adam Sobieski http://www.phoster.com

_______________________________________________ Abstract-Wikipedia mailing list -- abstract-wikipedia@lists.wikimedia.orgmailto:abstract-wikipedia@lists.wikimedia.org List information: https://lists.wikimedia.org/postorius/lists/abstract-wikipedia.lists.wikimed...

There is an existing proposal to support answering and context - wikipragmatica.

On Mon, Dec 27, 2021 at 2:02 PM Thad Guidry thadguidry@gmail.com wrote:

My thoughts,

All of it is incredibly ambitious. However, knowledge collection with supporting AI technologies to make collection and curation easier is highly needed. There are no doubts about that.

...

If we get the collaboration design and related features right...

I would first lay some foundation stepping stones of knowledge collection of questions with simple AI augmentation as you detail (context feedback loops will be especially important. Example: Car(train carriage), Car(automobile) ). And then iterate and expand with community involvement on providing answers through whatever mechanisms, AI, programming, might make a young student smile.

Context will be the thorn to constantly pluck out and disambiguate. But don't be disheartened since even the largest Q/A and Search companies have a constant struggle and have slow improvement processes. Good things sometimes take time.

I wish you and the team the best of luck, Adam.

Thad https://www.linkedin.com/in/thadguidry/ https://calendly.com/thadguidry/

---

Abstract-Wikipedia mailing list -- abstract-wikipedia@lists.wikimedia.org List information: https://lists.wikimedia.org/postorius/lists/abstract-wikipedia.lists.wikimed...