search

HyperAgents / hmas

An ontology to describe Hypermedia Multi-Agent Systems, interactions, and organizations.
https://purl.org/hmas/
1 stars 0 forks source link

Are `:Platform` and `:hosts` inspired by SSN ? #20

Closed maximelefrancois86 closed 2 years ago

maximelefrancois86 commented 2 years ago

:Platform and :hosts are also defined in SSN https://www.w3.org/TR/vocab-ssn/ .

Maybe we could adapt their definition to make explicit that we are borrowing these terms ?

maximelefrancois86 commented 2 years ago

Platform - A Platform is an entity that hosts other entities, particularly Sensors, Actuators, Samplers, and other Platforms.

And

hosts - Relation between a Platform and a Sensor, Actuator, Sampler, or Platform, hosted or mounted on it.

maximelefrancois86 commented 2 years ago

one could also link to these terms in SOSA with property rdfs:seeAlso ?

andreiciortea commented 2 years ago

@maximelefrancois86 hmas:Platform and hmas:hosts are inspired by (and closer to) their homonyms in the socio-technical networks (STN) ontology (btw, just noticed the STN website is down: https://w3id.org/stn).

+1 for a formal alignment with SSN. The definition for hmas:Platform is currently incomplete, I'll propose a list of features over the next couple of days (e.g., runtimes for agents, communication and interaction among agents): https://github.com/HyperAgents/ns.hyperagents.org/blob/51de578fdc490d734d09eb7750b7f543980b35bb/src/core.ttl#L77

Examples of hmas:Platforms include Twitter, Facebook, a JADE node/delpoyment, or a JaCaMo node/delpoyment — all of which provide features that could support Hypermedia MAS & hybrid communities. The definition of ssn:Platform is focused on entities for sensing and actuation, which could also support Hypermedia MAS & hybrid communities.

Should then hmas:Platform be a superclass of ssn:Platform?

smnmyr commented 2 years ago

The definition of ssn:Platform is focused on entities for sensing and actuation, which could also support Hypermedia MAS & hybrid communities.

Can this really be seen in this way? sosa:Platform is specifically meant for physical entities (including humans). But I'm unclear about what you mean with "support" in this context.

Should then hmas:Platform be a superclass of ssn:Platform?

I (unfortunately) have no quick good arguments against this, but it feels wrong to me - the intermixing of virtual and physical platforms, that is.

Message ID: @.*** com>

DrLeturc commented 2 years ago

I do not see any problem between intermixing of virtual and physical platforms ? Could you develop your idea ? : )

In MAS literature an agent may be physical, or virtual. Artificial, or biological. MAS are all kinds of systems in which you have several agents interacting in a same environment. hMAS is a particular case of MAS, where agents interact in their own environment (workspace) through a platform. We aim to find the most high-level ontology to catch every existing MAS-types within the core-ontology. Of course it would not be exactly as the class of (general) MAS (by definition of hMAS), but we aim to be as close as possible to MAS, in order to let the possibility of defining all of these MAS as hMAS. Do you see what I mean ? Thus, according to me, there is no problem. But maybe i didn't understand your claim. : )

andreiciortea commented 2 years ago

Actually, on a 2nd thought, I also tend to agree with @smnmyr. When I wrote my previous comment I was focused on the definition of sosa:Platform as:

Platform - A Platform is an entity that hosts other entities, particularly Sensors, Actuators, Samplers, and other Platforms.

But here is the reference example of how sosa:Platform is meant to be used:

A post, buoy, vehicle, ship, aircraft, satellite, cell-phone, human or animal may act as Platforms for (technical or biological) Sensors or Actuators.

So a sosa:Platform is a (physical) entity on which we can mount sensors, actuators, etc. This definition is closer to the definition of platform as "a raised level surface, for example one that equipment stands on or is operated from" (e.g., an oil/gas platform).

Therefore an hmas:Agent (e.g., a person) would be more suited to be considered a sosa:Platform. Now it's hard for me as well to see an hmas:Platform as a sosa:Platform (or the inverse) because here an hmas:Platform is a (Web-based?) system, not a physical entity on which we could mount sensors or actuators: https://github.com/HyperAgents/ns.hyperagents.org/blob/fb403478a482639dd8a44d7a6374590c7a4b7acc/src/core.ttl#L76-L78

@maximelefrancois86 do you agree with this view or is there something we might be missing?

FabienGandon commented 2 years ago

One possibility also is to qualify our notion of platform to make it unambiguous e.g. instead of hmas:Platform we could have hmas:SoftwarePlatform or hmas:WebPlatform or hmas:VirtualPlatform or ... if the word "Platform" is too ambiguous or overloaded we can deactivate it and prefer another one

FabienGandon commented 2 years ago

Following our meeting with @DrLeturc this morning I would suggest we clarify the concept of Platform by focusing on hmas:MASPlatform leaving other types of platforms (including physical ones) to other considerations. If we agree, we can close this issue and implement this change.

smnmyr commented 2 years ago

If we agree, we can close this issue and implement this change.

Did we agree?

andreiciortea commented 2 years ago

@FabienGandon I agree hmas:Platform is ambiguous, but I think hmas:MASPlatform would also be misleading. "MAS platform" has a specific meaning in the AAMAS community (e.g., JADE, JaCaMo) and, for instance, Facebook would not intuitively fit as a MAS platform.

To further clarify the concepts:

In [1], we defined Hypermedia MAS as a socio-technical system "composed of people and autonomous agents – henceforth agents – situated in a shared hypermedia environment that is distributed across the open, world-wide Web". This is aligned with the initial definitions we gave in [2].

In [3, 4], we did a deployment with socio-technical networks (STN) in which we had a hypermedia environment distributed across Facebook, Twitter, and multiple instances of an STN platform. Autonomous agents were then able to use Facebook and Twitter to navigate a distributed social graph (and they could also use Twitter to interact with people or other autonomous agents). In that deployment:

To avoid this confusion, we could instead use hmas:HypermediaMASPlatform. This would fit with all the previous definitions and publications on the topic (Hypermedia MAS and socio-technical networks).

[1] Andrei Ciortea, Simon Mayer, Fabien Gandon, Olivier Boissier, Alessandro Ricci, and Antoine Zimmermann. 2019. A Decade in Hindsight: The Missing Bridge Between Multi-Agent Systems and the World Wide Web. AAMAS 2019.

[2] Ciortea A., Boissier O., Ricci A. (2019) Engineering World-Wide Multi-Agent Systems with Hypermedia. In: Weyns D., Mascardi V., Ricci A. (eds) Engineering Multi-Agent Systems. EMAS 2018.

[3] Andrei Ciortea, Antoine Zimmermann, Olivier Boissier, and Adina Magda Florea. 2016. Hypermedia-driven Socio-technical Networks for Goal-driven Discovery in the Web of Things. WoT 2016.

[4] Andrei Ciortea, Olivier Boissier, Antoine Zimmermann, and Adina Magda Florea. 2017. Give Agents Some REST: A Resource-oriented Abstraction Layer for Internet-scale Agent Environments. AAMAS 2017.