CFP: Cognitive Robot Architectures

Recently I was appointed to the Editorial Board of the journal Cognitive Systems Research. We have just announced a call for submissions to a special issue that I am co-editing along with the other s13890417organisers of EUCognition2016.  Although we expect some authors of papers for that meeting to submit their papers for inclusion in this special issue, this is an open call: one need not attend EUCognition2016 to submit something for inclusion in the special issue.  The call, reproduced below, can also be found at:

http://www.journals.elsevier.com/cognitive-systems-research/call-for-papers/special-issue-on-cognitive-robot-architecture

Special Issue on Cognitive Robot Architectures


Research into cognitive systems is distinct from artificial intelligence in general in that it seeks to design complete artificial systems in ways that are informed by, or that attempt to explain, biological cognition. The emphasis is on systems that are autonomous, robust, flexible and self-improving in pursuing their goals in real environments.  This special issue of Cognitive Systems Research will feature recent work in this area that is pitched at the level of the cognitive architecture of such designs and systems.  Cognitive architectures are the underlying, relatively invariant structural and functional constraints that make possible cognitive processes such as perception, action, reasoning, learning and planning.  In particular, this issue will focus on cognitive architectures for robots that are designed either using insights from natural cognition, or to help explain natural cognition, or both.

Papers included in this issue will address such questions/debates as:

  • What makes an architecture a cognitive architecture? A good cognitive architecture?
  • Implementation-independent architectures vs strongly embodied architectures
  • Should/could there be a common framework for comparing cognitive architectures?
  • What are the tradeoffs of various architectural features?
  • What are the best examples of architecture-based designs of cognitive systems?
  • What is the role of architecture in the design-implement-evaluate cycle?

Possible topics to be explored by the contributions to this special issue include:

  • The role of cognitive architectures in robotic systems
  • The role of neural networks in the design of cognitive architectures
  • The role of cognitive science in the design of robotic architectures
  • Embodiment and robotic architectures
  • Architectures for robotic perception (esp. vision, audition, haptics, proprioception, kinaesthesia, interoception)
  • Architectures for motor control and behavioural organisation
  • Architectures for planning and mapping
  • Architectures for HRI
  • Machine learning and robotic architectures
  • Architectures for developmental robotics
  • Embodied deliberative architectures
  • Embodied reflective architectures

SUBMISSION GUIDELINES:

Submissions to the special issue must include original research. Papers must be new and have not been published or submitted to other journals. Authors should prepare their manuscript according to the “Guide for Authors” available at the journal homepage: http://www.journals.elsevier.com/cognitive-systems-research/.

Submission should be made via the EVISE system:
https://www.evise.com/evise/faces/pages/homepage/homepage.jspx?_adf.ctrl-state=10f3bazys2_132

Authors must select “VSI: Cognitive Robots” when they reach the “Article Type” step in the submission process. All papers will be peer-reviewed following the reviewing procedures of the Cognitive Systems Research (CSR) journal.

All papers will undergo a preliminary screening to ensure relevance to the special issue prior to be the peer-review phase; research papers that do not sufficiently address the special issue call may not be selected for a full peer review (such a decision will be communicated rapidly).

IMPORTANT DATES:

Submission open: November 30th, 2016
Deadline for paper submission: January 31st, 2016
Notification of acceptance: July 31st, 2017
Publication date: October 31st, 2017

GUEST EDITORS:

Ron Chrisley
University of Sussex, United Kingdom
ronc@sussex.ac.uk
http://www.sussex.ac.uk/profiles/476

Vincent Müller
Anatolia College/ACT & University of Leeds
vmueller@act.edu
http://www.sophia.de

Yulia Sandamirskaya
University of Zürich & ETH Zürich, Switzerland
yulia.sandamirskaya@ini.uzh.ch
http://sandamirskaya.eu

Markus Vincze
TU Wien, Austria
vincze@acin.tuwien.ac.at
http://www.acin.tuwien.ac.at/institut/mvincze/

Advertisements

3 thoughts on “CFP: Cognitive Robot Architectures

  1. [Oh, no: it’s HIM again …]

    Ron, has the argument that there are not [need not be] any “relatively invariant structures” in cognition (sic) been taken seriously in robotics or AI more widely?

    It was made by a highly respected but little published Oxford college Fellow (tutor): cognitive structure is entirely determined by content, he wrote. An even better known neurophysiologist made the same point about visual cortex. Current work on sparse adaptive synapses / the memory engram can be read that way too. The evidence can be shoe-horned into the notion of ‘modules’ which sound pretty invariant but are increasingly found to be highly ‘dynamic’ at integrating complex scenes or movements.

    Aaron’s architecture (of a decade ago) looked to me like fairly arbitrary lines drawn between (unspecified) sets of identical causal processing elements.

    Of course, all the cognitive structure in my causal maths comes entirely from the most effective way of using past environments to act in a present environment.

    IF this argument MIGHT be of any interest, I may be able to prepare a short and rough draft (not just a bunch of notes like the above) for SCREENING before the end of December. I’d consider from any approving feedback whether I could give a reviewable write-up enough priority to meet the 31 Jan deadline.

    Best regards.

    – David

    David Booth
    School of Psychology, University of Sussex
    Web: http://www.sussex.ac.uk/profiles/335100
    Email: D.A.Booth@sussex.ac.uk

    ________________________________

  2. David,

    Thanks for your comments.

    “[Oh, no: it’s HIM again …]”

    Not at all.

    “Ron, has the argument that there are not [need not be] any “relatively invariant structures” in cognition (sic) been taken seriously in robotics or AI more widely?”

    I doubt many people know of the particular argument you have in mind (I don’t), so probably not. The claim itself, though, has affinities with several *critiques* of traditional AI and Cognitive Science. But even many emergentists concede there are invariances, it is just that such invariances are the result of self-organisation in an environment with (some) invariances.

    “It was made by a highly respected but little published Oxford college Fellow (tutor): cognitive structure is entirely determined by content, he wrote.”

    Well, not entirely, surely? Expose a stone to a classical conditioning contingency schedule, and no association occurs. The aspects of a rat that make it, and not the stone, the kind of thing whose cognitive structure will depend on content can be referred to as its architecture. That architecture is not itself entirely determined by content.

    By the way, what is the name of the argument’s author?

    “An even better known neurophysiologist made the same point about visual cortex. Current work on sparse adaptive synapses / the memory engram can be read that way too.”

    See above. There must be something special about neurons and how they are coupled to the world via perception and action that allow them to self-organise in this way.

    “The evidence can be shoe-horned into the notion of ‘modules’ which sound pretty invariant but are increasingly found to be highly ‘dynamic’ at integrating complex scenes or movements.”

    If your point is that architecture is more fluid than particular theorists have maintained, then that seems like a very reasonable hypothesis.

    “Aaron’s architecture (of a decade ago) looked to me like fairly arbitrary lines drawn between (unspecified) sets of identical causal processing elements.”

    No, I don’t think that is a fair analysis of his proposed architectures. Also note than much of the time he was/we were talking about the COGAff architecture schema, which was not itself an architecture, but a framework for comparing them.

    “Of course, all the cognitive structure in my causal maths comes entirely from the most effective way of using past environments to act in a present environment.”

    What is it about an organism that ensures that it will always do the optimal thing?

    “IF this argument MIGHT be of any interest, I may be able to prepare a short and rough draft (not just a bunch of notes like the above) for SCREENING before the end of December. I’d consider from any approving feedback whether I could give a reviewable write-up enough priority to meet the 31 Jan deadline.”

    The other submissions to the special issue will likely not question the notion of cognitive architecture, so that could work either for or against you with the reviewers. As for screening, you’ll have to check with Jonny about whether there are any EI slots left before the end of December.

    Ron

    • Sorry, Ron, I’ve only just seen that my (?emailed) questions about a ‘No Cognitive Architecture’ submission to the SI on Cognitive Robot Architectures went up under your posting about that journal Issue submission. I receiving no email notification of your reply (or the posting of my questions), as I have done for other PAICS postings.

      Thank you for the information that ‘screening’ would be in an E-Int session (please correct me if I’ve misunderstood). That may not be possible by the end of December and so I’m taking your comments as a ‘pre-screening’ …

      I should not have written “the most effective way of” in my summary “all the cognitive structure in my causal maths comes entirely from the most effective way of using past environments to act in a present environment.” Delete that phrase, to read “… cognitive structure … comes entirely from using past environments to act in a present environment.” The effect of that act on the next occurrence of a similar environment determines the action then. The (cognitive) “uses” of past environments are instances of forms/formulae generated by subtraction of present from past. There is no optimising process within the system.

      A stone has no cognition to have architecture or not. The ‘special’ aspects of a rat, a cognitive robot or a human being – of the systems’ neurons, chips or whatever – are the adapting synapses distributed among many non-adapting synapses, as claimed by Hebb 1949, (at Sussex) Uttley 1969, and the consensus by the 1980s at the latest. The non-adapting synapses transmit present information content to and from the environment and through the sparse adaptive synapses which hold the past information content (memory). It’s not clear to me what invariants emerge which are not in the history of the system’s transactions with the material and social environments.

      In the diagram Aaron showed of CogAff’s schema, I could not see what the boundaries between unstructured sets of interconnecting arrows were adding to a schema without those few boundaries. ‘Hierarchies’ can be redrawn ‘flat’.

      I have half a dozen formulae for interactions between past-present disparities but differently shaped bricks do not make a building have architecture.

      I suppose I could focus the article on examples of processing of environmental content and structure without supervening (beg pardon!) cognitive architecture.

      – David

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s