RE: Synthetic Media Community Group

Owen,

A longer-term initiative might involve appealing to AAAI for them to recommend including more discussion of justification, provenance and explanation in AI curricula. Knowledge representation and reasoning courses can equip students with the skills needed to model justification, provenance and explanation and to add these features, as desired, to database and knowledgebase projects.

On these topics, an interesting futuristic technology topic is real-time fact-checking.

Journalists, debate moderators, their support staffs, and audience members can, today, make use of Web search engines to obtain information pertinent to what politicians and orators are saying. In the near future, people might be able input spoken language audio into their computers to obtain, on their screens, real-time fact-checking and other information relevant to spoken utterances.

This fact-checking or for-more-information content could appear on screen seconds after orators speak utterances. With a brief delay in video feeds, the fact-checking and informational content could appear on-screen as the utterances occur. There would likely be multiple such services available to journalists and debate moderators could make use of them simultaneously and select and assemble multi-source content.

Fact-checking technologies have more uses than journalism and debate moderation during election seasons. People could make broad use of real-time fact-checking and for-more-information services when working on documents, presentations and slideshows in various business and organizational contexts.

Argumentation technologists can envision services which provide, in real-time, materials which support and materials which argue against provided spoken language claims.

With real-time fact-checking technologies and related services, developers could easily add useful features to their dialogue systems, Q&A systems and intelligent tutoring systems. Developers could access real-time fact-checking, for-more-information, and argumentation-based services to overlay provided content atop their hypervideo streams. Multiple services could be accessed when dialogue systems form their utterances and multiple services could be used to composite content into what end-users interact with as a single WebRTC hypervideo stream.


Best regards,
Adam

From: Owen Ambur<mailto:Owen.Ambur@verizon.net>
Sent: Tuesday, February 25, 2020 11:13 PM
To: Adam Sobieski<mailto:adamsobieski@hotmail.com>; public-aikr@w3.org<mailto:public-aikr@w3.org>
Subject: Re: Synthetic Media Community Group


Adam, conceptually speaking, I believe we are violent agreement.

Provenance is also a key concept at the heart of records management.  https://www.archives.gov/research/alic/reference/archives-resources/principles-of-arrangement.html

What I look forward to learning is what, practically speaking, we might be able do about it by working collaboratively together -- particularly since basic document/records management principles seem to be too boring for the W3C ... and it's hard to learn to run if we do not first learn to crawl and walk.

To coin a Yogi Berraism<https://ftw.usatoday.com/2019/03/the-50-greatest-yogi-berra-quotes>:  You won't get there if you don't go there.

Owen
On 2/25/2020 7:49 PM, Adam Sobieski wrote:
Owen,

Also on the topics of the veracity and reliability of content is, perhaps more relevant to AIKR, epistemology and the means by which to ensure the accuracy of information relayed by digital characters, e.g. intelligent tutoring systems and question answering systems.

Brainstorming: referenced materials could appear atop synthesized video content, in a lower third manner, as digital characters answer users’ questions. This approach could be of use for both video and hypervideo scenarios.


Best regards,
Adam

https://en.wikipedia.org/wiki/Lower_third
https://en.wikipedia.org/wiki/Hypervideo

From: Adam Sobieski<mailto:adamsobieski@hotmail.com>
Sent: Tuesday, February 25, 2020 4:47 PM
To: Owen Ambur<mailto:Owen.Ambur@verizon.net>; public-aikr@w3.org<mailto:public-aikr@w3.org>
Subject: RE: Synthetic Media Community Group

Owen,

For those desiring more information about synthetic media, here are some hyperlinks:


  1.  http://smartbody.ict.usc.edu/
  2.  https://www.synthesia.io/
  3.  https://medium.com/@vriparbelli/our-vision-for-the-future-of-synthetic-media-8791059e8f3a

I enjoyed the concept of democratized creativity mentioned in the article indicated by the third hyperlink. I am hoping that new standards can facilitate the artificial-intelligence-enhanced creation, modification, interchange, use and reuse of media including, but not limited to: imagery, video, audio, speech, text, documents, scripts and screenplays, dialogue, and 3D graphics. I have some ideas with respect to advancing computer animation (like BML) and articulatory synthesis (like SSML).

Digital characters are a component of intelligent tutoring systems which have interested me for some time.

The output formats points you broached seem applicable to MPEG, WebM and WebRTC.


Best regards,
Adam

From: Owen Ambur<mailto:Owen.Ambur@verizon.net>
Sent: Tuesday, February 25, 2020 1:34 PM
To: Adam Sobieski<mailto:adamsobieski@hotmail.com>; public-aikr@w3.org<mailto:public-aikr@w3.org>
Subject: Re: Synthetic Media Community Group


Adam, I'll be interested to see how your interest cloud-based automated planning relates to the StratML standard.  As your plan comes together, I look forward to rendering it in StratML format for inclusion in our collection at https://stratml.us/drybridge/index.htm#W3C

In the meantime, I encourage you to include StratML (ISO 17469-1) in your list of related standards.

Based upon the televideo conference that Carl Mattocks arranged today, several of us will be using Chris Fox's StratNavApp to collaboratively draft a plan addressing knowledge representation for AI.  https://www.stratnavapp.com/

Chris' app outputs plans and reports in StratML format.

Rendering content in open, schema-compliant, machine-readable format will greatly facilitate synthesis.

In the CredWeb CG, I am also arguing that it will support evaluation of veracity and reliability.  Why should we pay any particular attention to content which ignores applicable standards?

Owen
On 2/25/2020 12:06 AM, Adam Sobieski wrote:
Artificial Intelligence Knowledge Representation Community Group,

Introduction

A new W3C Community Group is launching: the Synthetic Media Community Group. I would like to invite you to support the creation of this group and to join this new group: https://www.w3.org/community/groups/proposed/#synthetic-media .

Our group is interested in every component of the modern architectures of artificial intelligence with which to generate or synthesize media and interactive media content.

We intend to advance the state of the art with a number of new standards. With new and improved standards, teams and organizations will be better able to explore and develop the interoperable components which comprise these modern architectures of artificial intelligence.

We hope that you can join us in the advancement of artificial intelligence standards and technologies.

Synthetic Media

Synthetic media is any content created or modified algorithmically.

Applications of synthetic media include: education, journalism, and entertainment.

Digital Characters

Our group is interested in digital characters, behavior modeling, and in the numerous other topics of computer animation such as: lighting, cameras, timelines, tracks, and scene graphs.

Dialogue Systems

Our group is interested in the topics of multimodal mixed-initiative dialogue systems including: speech recognition, natural language understanding, natural language generation and speech synthesis.

Modeling and Simulation

Our group is interested in modeling and computer simulation topics.

Computational Narratology, Digital Screenplays and Interactive Narrative

Computer-generated stories, scripts and screenplays can facilitate the complex and interactive behaviors of one or more digital characters including interactions between digital characters and on-screen imagery, presentation slides, videos, 3D graphics, virtual sets and video walls.

Our group is interested in the generation and real-time generation of stories, scripts, screenplays, storyboards and cinematography.

Automated Planning and Scheduling

Automated planning is a traditional approach to generating media content. Our group is interested in cloud-based automated planning and scheduling.

Deep Learning

Deep learning is a recent approach to generating media content. Our group is interested in cloud-based deep learning.

Content Evaluation and Computational Aesthetics

Our group is interested in the automatic evaluation of plans, stories, scripts, screenplays, storyboards, cinematography, visual imagery, audio, dialogues, natural language, explanations, presentations and other media content.

Computational aesthetics and evaluation technologies can be of use in guiding search, when comparing alternatives, when making design decisions, and when training neural systems.

Existing Standards and Technologies

Existing standards and technologies relevant to our group include, but are not limited to: FML, BML, MURML, SSML, SMIL, WebRTC, HTML, RDF, and JavaScript.

Conclusion

Thank you. We hope that you can join us in the advancement of artificial intelligence standards and technologies.



Best regards,
Adam Sobieski