Business Of Linked Data: Opportunities re., Smart Agents (Bots)

33 views
Skip to first unread message

Kingsley Idehen

unread,
Jul 8, 2016, 10:40:38 AM7/8/16
to business-of-li...@googlegroups.com
All,

Smart Agents and Bots are now hot topics across the industry at large.

Bearing in mind years of knowledge and experience this community has in
regards to Bots and Smart Agents [1], are there any Linked Data driven
Smart Agents out there? Basically, conversational bots that have the ability
to leverage semantically-rich Linked Open Data clouds.

Personally, I believe this new wave of interest in Bots provides a great
opportunity to showcase the value proposition of a Semantic Web build
using Linked Open Data.

Historically, there has been a disconnect between the technology and business opportunities related to Linked Data and a Semantic Web. Thus, I am opening a discussion thread to explore the business related aspects of this matter (i.e., this has nothing to do with research, papers, and conferences).
Links:

[1]
http://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Ftwitter.com%2Fhashtag%2FSmartAgent%23this&distinct=1
-- Smart Agent Notes collated over time.

[2] http://techemergence.com/valuing-the-artificial-intelligence-market-2016-and-beyond/ -- Market Size & Projections Example


--
Regards,

Kingsley Idehen
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this


Seth Russell

unread,
Jul 8, 2016, 4:04:53 PM7/8/16
to business-of-li...@googlegroups.com
All,

i have a extensive corpus of English thought compiled since 2005 that is now coded in a relational database.  It runs http://fastblogit.com which is the first http://thinking.domains .   It is extensively tagged in English folksonomy and the items themselves are densly linked.   All of that could be represented in linkeddata.   My intention is to find a bot with which a person could interact with this corpus in natural language … say with tag alexa or tag siri. 

If anyone wants to play with this … i would be more than willing to make the data and the server available.

Seth Russell
Developer at Thinking Domains




--
You received this message because you are subscribed to the Google Groups "Business Of Linked Data (BOLD)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to business-of-linked-d...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Kingsley Idehen

unread,
Jul 8, 2016, 7:00:24 PM7/8/16
to business-of-li...@googlegroups.com
On 7/8/16 4:04 PM, Seth Russell wrote:
All,

i have a extensive corpus of English thought compiled since 2005 that is now coded in a relational database.  It runs http://fastblogit.com which is the first http://thinking.domains .   It is extensively tagged in English folksonomy and the items themselves are densly linked.   All of that could be represented in linkeddata.   My intention is to find a bot with which a person could interact with this corpus in natural language … say with tag alexa or tag siri. 

If anyone wants to play with this … i would be more than willing to make the data and the server available.

Seth Russell
Developer at Thinking Domains


Hi Seth,

Could you make CSV dumps of this data? Basically, a CSV file per Table or View.

With the above in place, an ontology could be generated which is then mapped to existing shared ontologies.

That's the first critical step.

On the Smart Agent front, it simply needs to understand relationship type semantics exposed by ontologies related to the activity I outlined above. 

Kingsley

Seth Russell

unread,
Jul 8, 2016, 7:43:44 PM7/8/16
to business-of-li...@googlegroups.com
Hi Kingsley,

Yes i could make the tables available in a CSV dumps.  Bear in mind that there is a lot of data at least by my standards.  Is there some place in the cloud to which that i could upload?  

We also have privacy to respect here, not all of the data is public, some is marked private.   So a CSV dump would be only of the public data.  There is plenty of that.

By the way we have an API ... that might be a better interface to the data than a CSV dump ... i don't know ... what do you think?



Kingsley Idehen

unread,
Jul 9, 2016, 12:15:44 PM7/9/16
to business-of-li...@googlegroups.com
On 7/8/16 7:43 PM, Seth Russell wrote:
Hi Kingsley,

Yes i could make the tables available in a CSV dumps.  Bear in mind that there is a lot of data at least by my standards.  Is there some place in the cloud to which that i could upload?  

We also have privacy to respect here, not all of the data is public, some is marked private.   So a CSV dump would be only of the public data.  There is plenty of that.

By the way we have an API ... that might be a better interface to the data than a CSV dump ... i don't know ... what do you think?



Seth,

If you a comfortable with RDF-Turtle (Turtle) notation, for structured data representation, how about simply describing the actions performed by your API using Turtle? That way, in line with the theme of this thread, a Smart Agent could dynamically understand the URLs call patterns required for performing specific actions?

Examples of APIs described using Nanotation (RDF-Turtle embedded in places where plain text is allowed):

[1] https://medium.com/virtuoso-blog/describing-web-services-using-a-linked-data-doc-using-schema-org-terms-6caa399fc97a#.lv5a4o45o -- Describing Search & Find API

[2] https://medium.com/virtuoso-blog/describing-a-new-york-times-web-service-7a07a11019ae#.5eya6sszi -- Describing NYT API

[3] http://linkeddata.uriburner.com/about/html/http://kingsley.idehen.net/DAV/home/kidehen/Public/Swagger/uber_api_swagger.json -- shows conversion of JSON variant of Swagger Description of Uber's API converted to RDF using Schema.org terms

[4] http://linkeddata.uriburner.com/describe/?uri=http%3A%2F%2Flinkeddata.uriburner.com%2Fabout%2Fid%2Fentity%2Fhttp%2Fkingsley.idehen.net%2FDAV%2Fhome%2Fkidehen%2FPublic%2FSwagger%2Fuber_api_swagger.json -- alternative view

[5] http://kidehen.blogspot.com/2014/07/nanotation.html -- About Nanotation

Kingsley Idehen

unread,
Jul 14, 2016, 11:16:47 AM7/14/16
to business-of-li...@googlegroups.com, publi...@w3.org
On 7/8/16 10:40 AM, Kingsley Idehen wrote:
> All,
>
> Smart Agents and Bots are now hot topics across the industry at large.
>
> Bearing in mind years of knowledge and experience this community has in
> regards to Bots and Smart Agents [1], are there any Linked Data driven
> Smart Agents out there? Basically, conversational bots that have the ability
> to leverage semantically-rich Linked Open Data clouds.
>
> Personally, I believe this new wave of interest in Bots provides a great
> opportunity to showcase the value proposition of a Semantic Web build
> using Linked Open Data.
>
> Historically, there has been a disconnect between the technology and business opportunities related to Linked Data and a Semantic Web. Thus, I am opening a discussion thread to explore the business related aspects of this matter (i.e., this has nothing to do with research, papers, and conferences).
> Links:
>
> [1]
> http://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Ftwitter.com%2Fhashtag%2FSmartAgent%23this&distinct=1
> -- Smart Agent Notes collated over time.
>
> [2] http://techemergence.com/valuing-the-artificial-intelligence-market-2016-and-beyond/ -- Market Size & Projections Example

All,

Here's an example of a Bot [1] that falls in line with what I was trying
to instigate discussion about. I've also joined a thread on Product Hunt
where some work is taking shape re., Facebook Messenger [2].

[1] http://growthbot.org/

[2] https://www.producthunt.com/tech/growthbot#comment-326494

Kingsley Idehen

unread,
Jul 15, 2016, 10:41:49 AM7/15/16
to Wetz Peter, business-of-li...@googlegroups.com, publi...@w3.org
On 7/14/16 11:54 PM, Wetz Peter wrote:
> Sorry for stepping in here, but for people involved in this discussion, the (first) Linked Startup Workshop at the upcoming ISWC 2016 would be a great opportunity to talk about this and everything related to Linked Data businesses/startups.
>
> There is still some time left to submit (informal) 2-4page proposals: http://linkedstartup.ifs.tuwien.ac.at/2016/call-for-papers/
>
> Best,
> Peter

Peter,

I am not one to generally complain about forum postings, but please do
not use posts like this to negate the CFP "best practices" already
outlined for the public-lod mailing. In addition, assume the same "best
practices" with regards to the BoLD list (which is about the Business of
Linked Data).

Paper submission postings have pretty much destroyed the utility of a
variety of forums. IMHO.

Kingsley
> ________________________________________
> Von: Kingsley Idehen <kid...@openlinksw.com>
> Gesendet: Donnerstag, 14. Juli 2016 17:16
> An: business-of-li...@googlegroups.com
> Cc: publi...@w3.org
> Betreff: Re: Business Of Linked Data: Opportunities re., Smart Agents (Bots)

Melvin Carvalho

unread,
Jul 16, 2016, 6:20:08 AM7/16/16
to business-of-li...@googlegroups.com, publi...@w3.org
On 14 July 2016 at 17:16, Kingsley Idehen <kid...@openlinksw.com> wrote:
On 7/8/16 10:40 AM, Kingsley Idehen wrote:
> All,
>
> Smart Agents and Bots are now hot topics across the industry at large.
>
> Bearing in mind years of knowledge and experience this community has in
> regards to Bots and Smart Agents [1], are there any Linked Data driven
> Smart Agents out there? Basically, conversational bots that have the ability
> to leverage semantically-rich Linked Open Data clouds.
>
> Personally, I believe this new wave of interest in Bots provides a great
> opportunity to showcase the value proposition of a Semantic Web build
> using Linked Open Data.
>
> Historically, there has been a disconnect between the technology and business opportunities related to Linked Data and a Semantic Web. Thus, I am opening a discussion thread to explore the business related aspects of this matter (i.e., this has nothing to do with research, papers, and conferences).
>  Links:
>
> [1]
> http://linkeddata.uriburner.com/describe/?url=https%3A%2F%2Ftwitter.com%2Fhashtag%2FSmartAgent%23this&distinct=1
> -- Smart Agent Notes collated over time.
>
> [2] http://techemergence.com/valuing-the-artificial-intelligence-market-2016-and-beyond/ -- Market Size & Projections Example

All,

Here's an example of a Bot [1] that falls in line with what I was trying
to instigate discussion about. I've also joined a thread on Product Hunt
where some work is taking shape re., Facebook Messenger [2].

Why cant we make smart agents run anywhere?

e.g. as a desktop app, from the command line, one your phone, on a server, and in the browser?

Using javascript right now seems to a great way to facilitate this.

I have programmed chat bots in js using hubot and they are useful.

Right now I'm using kue [1] and am thinking it would be nice to have something like a distributed URIBurner

We also need to think about what would be a compelling demo.  I keep coming back to semantic web + multi media as being something anyone can instantly understand.
 

--
Regards,

Kingsley Idehen
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this


Melvin Carvalho

unread,
Jul 16, 2016, 6:28:45 AM7/16/16
to business-of-li...@googlegroups.com
On 8 July 2016 at 16:40, Kingsley Idehen <kid...@openlinksw.com> wrote:
All,

Smart Agents and Bots are now hot topics across the industry at large.

Bearing in mind years of knowledge and experience this community has in
regards to Bots and Smart Agents [1], are there any Linked Data driven
Smart Agents out there? Basically, conversational bots that have the ability
to leverage semantically-rich Linked Open Data clouds.

Personally, I believe this new wave of interest in Bots provides a great
opportunity to showcase the value proposition of a Semantic Web build
using Linked Open Data.

Historically, there has been a disconnect between the technology and business opportunities related to Linked Data and a Semantic Web. Thus, I am opening a discussion thread to explore the business related aspects of this matter (i.e., this has nothing to do with research, papers, and conferences).


Let's think about an architecture for this.

At the heart of the smart agent I think you need a priority based queue of jobs.

There will be a number of interfaces to read and write to this queue, ideally one of which is HTTP.

It should be able to discover other nodes and talk to them in a common language (Linked Data)

It should be able to add jobs and request jobs from other nodes.

A incentives based system should help prioritze the jobs (this is exactly the function of bitcoin btw) -- this can also be linked data based

The bots should be able to interact with knowledge bases (their own and remote) -- querying, syncing, posting

The bots should do useful things like distributed multi media search, entertainment, unexpected discovery, advocacy, games etc.  Simple demos should mock this up.

The bots should run anywhere, easily down loadable, as an app etc.

Ultimately the smart agents should generate revenue for those that have opted in.  Providing a basic income for anyone that wants it.
 

Kingsley Idehen

unread,
Jul 18, 2016, 8:48:38 AM7/18/16
to business-of-li...@googlegroups.com
On 7/16/16 6:28 AM, Melvin Carvalho wrote:


On 8 July 2016 at 16:40, Kingsley Idehen <kid...@openlinksw.com> wrote:
All,

Smart Agents and Bots are now hot topics across the industry at large.

Bearing in mind years of knowledge and experience this community has in
regards to Bots and Smart Agents [1], are there any Linked Data driven
Smart Agents out there? Basically, conversational bots that have the ability
to leverage semantically-rich Linked Open Data clouds.

Personally, I believe this new wave of interest in Bots provides a great
opportunity to showcase the value proposition of a Semantic Web build
using Linked Open Data.

Historically, there has been a disconnect between the technology and business opportunities related to Linked Data and a Semantic Web. Thus, I am opening a discussion thread to explore the business related aspects of this matter (i.e., this has nothing to do with research, papers, and conferences).


Let's think about an architecture for this.

At the heart of the smart agent I think you need a priority based queue of jobs.

Naturally.

There will be a number of interfaces to read and write to this queue, ideally one of which is HTTP.

Yes.



It should be able to discover other nodes and talk to them in a common language (Linked Data)

RDF would be the Language with packets in Linked Data form.

It should be able to add jobs and request jobs from other nodes.

A incentives based system should help prioritze the jobs (this is exactly the function of bitcoin btw) -- this can also be linked data based


The bots should be able to interact with knowledge bases (their own and remote) -- querying, syncing, posting

Yes.



The bots should do useful things like distributed multi media search, entertainment, unexpected discovery, advocacy, games etc.  Simple demos should mock this up.

In a nutshell, they are your eyes and ears in cyberspace while also providing local productivity in regards to meeting scheduling, task management, calendar management etc..



The bots should run anywhere, easily down loadable, as an app etc.

Bots should be accessible via a variety of messaging interfaces ranging from those provided by Facebook to Slack, for instance.

Ultimately the smart agents should generate revenue for those that have opted in.  Providing a basic income for anyone that wants it.

Smart Agents can generate revenue or simply provide productivity value to their users.

Kingsley

Melvin Carvalho

unread,
Jul 18, 2016, 9:48:16 AM7/18/16
to business-of-li...@googlegroups.com
+1 on everything

I think we are aligned on the high level architecture.

What's next steps?

Should we start some basic design discussions?  Dogfooding?  Prototyping?  Use cases?

Kingsley Idehen

unread,
Jul 18, 2016, 10:12:27 AM7/18/16
to business-of-li...@googlegroups.com
Melvin,

Ideally, we should at least hear from others who may be working on similar issues. I don't want to believe that porting Phenny, and other #swig bots isn't an obvious endeavor here.

I would give it 48 hours, and then proceed to one of the steps you've suggested.

Kingsley

Sherman Monroe

unread,
Sep 24, 2016, 2:20:58 AM9/24/16
to business-of-li...@googlegroups.com, Sarven Capadisli
Greetings All,

I've like to revive this thread with a description of a protocol, architecture, and implementation details for a smart agent framework based on Linked Data, which addresses many of Melvin's earlier points. The work draws heavily on ideas from DARQ, but applies them to web services in general, not just SPARQL endpoints. I would be grateful if you could provide feedback on my design choices, and/or help in developing it into a general recommendation for smart agent deployment.

RDF Universally Distributed Index (RUDI) - is a protocol for web services message routing and response federation. It is essentially an intelligent API Gateway that plugs RESTful services into a hyperdata access layer (HAL).

Actors:

- Client
- Router
- Service
- Sponsor

Architectural Constraints:

1) Stateless - the router is unaware of transaction states
2) Infinite Scrolling - the Client browses Service responses in "scrollable" groups so as not to overwhelm Services, and deep fetching should be discouraged
3) Messaging Queue - the Client and Service interact through a message queue, never directly
4) Planning Support - the Service should provide metadata to support Service call planning and optimization, e.g. number of triples and supported classes, properties and namespaces for SPARQL endpoints
5) Built-in Economy - native support for incentivizing Services through (optional) royalties payments paid for by non-intrusive Sponsor messages, or by clients, or both
6) Distributed Indexing - channels in the message queue should be syndicated to subscribing Services, the Service owner's involvement should not be required in the subscription process
7) Trust - all Actors are identified by Public Key, and the queue should expose a web of trust rating for each
8) Federated Results - Clients should be able to subscribe to a virtual channel corresponding to a request and along which responses from various Services are asynchronously transmitted by the Router

I achieve these goals with Adaptors (6), which are RDF documents that describe how to:

1b) Extract patterns from Client input
2b) Use extracted data to fill placeholders in Service call templates
3b) Transform Service output into Linked Data
4b) How, and who, to charge for adapted responses

A pattern (1b) can be regex with captured groups, a type of RDF class, a type of RDF property, or one of many parsed data types supplied by the system (e.g. dateTime, address, geo, etc). JSON, and other non-XML Service output, is converted to XML automatically by the system. Translations (3b) are provided by XSLT documents whose URLs are contained in the Service call templates. Credentials for the Services are obtained and managed by the Routers, and are accessible to Adaptor authors via reserved placeholders. Adaptors are signed by their authors and submitted to the queue via the Router. Untrusted Adaptors are not accepted. Service trust is based on its SSL Certificate. Adaptors involving Services new to the Router and which require credentials must be manually credentialed by the Router's maintainers. The spec gives no guidance on how such cases are to be caught.

To suppress spam, Client UIs can filter the content in response channels based on trust. Special Services called Sponsors can subscribe to request channels and attach a message to the response in exchange for a micropayment to the Service and Adaptor author (adaptors sign their adaptors and Services prove they own the URL). These payments are based on Public Keys and are managed by the blockchain (e.g. Ethereum). Sponsor channels are SPARQL queries that return trusted Sponsor messages maintained in the queue. Like all channels, the Sponsor query is continually polled and results are placed a non-intrusive inbox (the Sponsor channel).


Inline image 2

There are several potential business opportunities in this model. A company could provide secure Routers in exchange for a small portion of the Sponsor revenue. Sponsors in turn would subscribe to Routers based on various preferences (Router fees, API coverage, etc). Adaptor authors could earn revenue by adapting existing APIs on the web. Services owners could write Adaptors to plug their own web services into this economy. Since the blockchain manages the economy, users could even attach a promised payment to their requests. Many types of Clients could be built that allow human interaction with Routers. These actors should partner together to build out the architecture and provide open source implementations.

I've been working on a reference Router and some simple Adaptors, which I plan to share sometime next week.

Cheers,
-sherman

 

Kingsley
 
To unsubscribe from this group and stop receiving emails from it, send an email to business-of-linked-data-bold+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "Business Of Linked Data (BOLD)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to business-of-linked-data-bold+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.


-- 
Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software     
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this
--
You received this message because you are subscribed to the Google Groups "Business Of Linked Data (BOLD)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to business-of-linked-data-bold+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "Business Of Linked Data (BOLD)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to business-of-linked-data-bold+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.


-- 
Regards,

Kingsley Idehen	      
Founder & CEO 
OpenLink Software     
Company Web: http://www.openlinksw.com
Personal Weblog 1: http://kidehen.blogspot.com
Personal Weblog 2: http://www.openlinksw.com/blog/~kidehen
Twitter Profile: https://twitter.com/kidehen
Google+ Profile: https://plus.google.com/+KingsleyIdehen/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen
Personal WebID: http://kingsley.idehen.net/dataspace/person/kidehen#this

--
You received this message because you are subscribed to the Google Groups "Business Of Linked Data (BOLD)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to business-of-linked-data-bold+unsub...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.



--

Thanks,
-sherman

Every good gift and every perfect gift is from above, and cometh down from the Father of lights, with whom is no variableness, neither shadow of turning.
(James 1:17)
Reply all
Reply to author
Forward
0 new messages