Saturday, May 31, 2014
Learning to Manipulate and Categorize in Human and Artiﬁcial Agents
Giuseppe Morlino, Claudia Gianelli, Anna M. Borghi, Stefano Nolﬁa
Cognitive Science (2014) 1–26
This study investigates the acquisition of integrated object manipulation and categorization abilities through a series of experiments in which human adults and artiﬁcial agents were asked to learn to manipulate two-dimensional objects that varied in shape, color, weight, and color intensity. The analysis of the obtained results and the comparison of the behavior displayed by human and artiﬁcial agents allowed us to identify the key role played by features affecting the agent/environment interaction, the relation between category and action development, and the role of cognitive biases originating from previous knowledge.
The paper looks at issues in the effect of action on categorisation. They present that categorisation is grounded in in the sensorimotor system, according to present experiments and theory. And again suggest the central role of action in cognition.
They also look at the issues around how categories enable the flexible usage of objects, and how the grasping of objects changes according to the tasks needed, as per the classic idea of affordances by Gibson (1979).
Important quote: "Affordances are proposed to be the product of the conjunction, in the brain, of repeated visuomotor experiences." Probably a no-brainer to the design community, but important to me, as I need to see this generalise to virtual worlds. It should be noted that the systems used in this experiment were synthetic, so the effects should generalise to a virtual world, as it is simply shapes and colours with physical properties. However, there is a history of visual search research with simple shapes not generalising to real images. This must be considered in any assumptions of efficacy in virtual world simulations.
The experiments involved the manipulation of 2D objects on the screen with a mouse pointer in placing and shaking tasks. The weight of the objects is aligned with categories and some of the categories are also based on colour, blinking and shape. The humans (20) were compared to neural network agents.
"The results indicated the discriminative features affecting the agent environment interaction such as weight facilitate the acquisition of the required categorisation abilities with respect to alternative features that are equally informative but that do not affect the outcome of the agent actions." This leads them to the conclusion that the categorisation for both humans and agents, not withstanding any other factors, is affected by the embodiment of the activity; weight required interaction, not just observation.
The results showed support for a model whereby the interaction with light vs heavy objects produces categories far more effectively than other factors. Embodied action thus has a great affect on categorisation, whether it affects every category is still uncertain, as the other visual effects (from grounded cognitive affects) still caused categories to form, just not as soon in the training.
They consider this to contribute to a STRONG position of embodiment being central to the creation of categories, and not just being a more peripheral contributor.
They also note a shape effect with humans, ie. they used a curvilinear path with circles, and a rectilinear path with square. Thus previous memories of the objects influenced their actions and thus the categories.
They also note that the categories are from an interaction of the agent with the environment, and not so from top-down or bottom-up processes exclusively, not overgeneralised or fine granularity categories, but as a dynamic process between agent and environment.
While this is categorisation, and not a memory task, one still has to wonder, for my work, if the memory of a process will be much more enhanced by embodied interactions, and not just visual interactions alone. One could hypothesise that if the category is more strongly created with embodied action, then the memory of that category (if it maps to say activity specifications) then should be stronger on acting it out. So an Occulus and Kinect space should measurably work better in process memory tasks than a pure visual space; with both working better than a simple interview.
Something to think about I guess.
Wednesday, May 21, 2014
Video of S-BPM based virtual world modelling tool developed by my Honours student Joel Harman with the financial assistance of Metasonic, AG - http://www.metasonic.de/
The tool allows a stakeholder to use a 3D representation of their work place to provide activity and message information to create an S-BPM model.
This was presented and tested in focus groups at the S-BPM One conference in Eichstaett over Easter.
Well done Joel! Thanks again to Metasonic for the provision of a scholarship for Joel and travel money to Eichstaett.
How Social Distance of Process Designers Affects the Process of Process Modeling:
Insights From a Controlled Experiment
Jens Kolb, Michael Zimoch, Barbara Weber, Manfred Reichert
In: 29th Symposium On Applied Computing (SAC 2014), Enterprise Engineering Track, 24-28 March, 2014, Gyeongju, South Korea.
The increasing adoption of process-aware information systems (PAISs) by enterprises has resulted in large process model collections. Usually, process models are created either by in-house domain experts or external consultants. Thereby, high model quality is crucial, i.e., process mod- els should be syntactically correct and sound, and also reflect the real business processes properly. While numerous guidelines exist for creating correct and sound process mod- els, there is only little work dealing with cognitive aspects affecting process modeling. This paper addresses this gap and presents a controlled experiment using construal level theory. We investigate the influence the social distance of a process designer to the modeled domain has on the creation of process models. In particular, we are able to show significant differences between high and low social distance in respect to model quality and granularity. The results may help enterprises to compose adequate teams for creating or optimizing business process models.
Dsiclaimer!!! I do work with these people, and have a virtual world now with Jens and Michael, as a further extension to this work. However, I am reviewing this paper as it has relevance to the situated cognition aspects that I am investigating, so I thought I would mull over it a little.
So their paper talks about a psychological distance concept, derived from Construal Level Theory (Todorov) which is a social psychology construct. While it makes sense, I wonder whether this distance is related to how grounded, embodied or situated the memory of the tasks has become, I think social distance is linked to straight forward memory of sequences in some way.
Is the social distance related or synonymous with theories of situated cognition and memory with task sequences. The context of their work is within the view-oriented visuals generated by their proView project implementation. However, that is for typical process models, not 3D visuals of a work place.
Does social distance creates abstraction? Or does it simply provide different memories of acting out the process, or, different simulations of that process from our own internal perspectives?!?!?
They do note spatial issues, but have not addressed this in their work.
Interestingly, their thoughts give support to my ideas of using virtual worlds to engage people at the level of operationalisation in a process. A first hypothesis could be: does the use of the virtual world undermine some of the social distance by the act of roleplaying the process in front of the analyst doing the modelling?!?
From their analysis: "In summary, hypotheses H1,1 and H1,3 can be accepted. In turn, hypothesis H1,4 is only partially supported and thus it cannot be accepted. Further, hypothesis H1,2 must be rejected. From this, we can conclude that low social distance has a positive impact on the granularity (H1,1) and semantic quality (H1,3) of resulting process models. We may also assume that low social distance has a positive impact on the perceived quality (H1,4); however, since it is partially supported, it cannot be generalized. Regarding syntactic quality (H1,2), no statistically significant difference is observed."
So the closer the social distance, the better are the semantic and granular aspects of the model. I find interesting the observation that the only change is the modelling brief at the start having a stranger vs intimate friend in the lunch acquisition process. It is intriguing that the social distance has an effect on the articulation of a process, and at a significant level on first analysis. Does, maybe, this even come down to motivation?!? Could it be that the participants were primed by the reference to friend to be more engaged in the task. I note no controls on this aspect in the paper, but maybe I missed some subtlety in my brief reading.
Also, could the emotional attachment priming have affected the acquisition of memories, memories that can be read by the person in their modelling task. So, is it social distance, or attachment levels that are measured here? Maybe some controls around the identity of the person they are lunching with are needed. Hey, if I imagined my lovely partner Ulrike, then I can imagine it affecting my motivation levels in some way. :-)
Also, the chances are that they have a memory of lunching with friends, and not with foreign strangers, thus the particpants are reading straight from memory.
My final thoughts are that there may be more experimentation needed to tease out the interactions of social distance vs actual memories vs motivation levels.
However, an intriguing paper, with lots of future analysis possibilities relevant to my work.
Construal Theory - Todorov, A., Goren, A., Trope, Y.: Probability as a Psychological Distance: Construal and Preferences. J Experimental Social Psychology 43 (2007) 473–482
Semiotic Theory and Conceptual Models - Lindland, O.I., Sindre, G., Solvberg, A.: Understanding Quality in Conceptual Modeling. IEEE Software 11 (1994) 42–49
Experimental Design - Wohlin, C., Runeson, P., Ho ̈st, M., Ohlsson, M.C., Regnell, B., Wesslen, A.: Experimentation in Software Engineering - An Introduction. Kluwer (2000)
Tuesday, May 13, 2014
Aspects of situated cognition in embodied numerosity: the case of ﬁnger counting
Mirjam Wasner, Korbinian Moeller, Martin H. Fischer, Hans-Christoph Nuerk
Cognitive Processing, Springer Verlag
Numerical cognitions such as spatial-numerical associations have been observed to be inﬂuenced by
grounded, embodied and situated factors. For the case of ﬁnger counting, grounded and embodied inﬂuences have been reported. However, situated inﬂuences, e.g., that reported counting habits change with perception and action within a given situation, have not been systematically examined. To pursue the issue of situatedness of reported ﬁnger-counting habits, 458 participants were tested in three separate groups: (1) spontaneous condition: counting with both hands available, (2) perceptual condition: counting with horizontal (left-to-right) perceptual arrangement of ﬁngers (3) perceptual and proprioceptive condition: counting with horizontal (left-to-right) perceptual arrangement of ﬁngers and with busy dominant hand. Report of typical counting habits differed strongly between the three conditions. 28 % reported to start counting with the left hand in the spontaneous counting condition (1),
54 % in the perceptual condition (2) and 62 % in the perceptual and proprioceptive condition (3). Additionally, all participants in the spontaneous counting group showed a symmetry-based counting pattern (with the thumb as number 6), while in the two other groups, a considerable Numerical cognitions such as spatial-numerical associations have been observed to be inﬂuenced by
grounded, embodied and situated factors. For the case of ﬁnger counting, grounded and embodied inﬂuences have been reported. However, situated inﬂuences, e.g., that reported counting habits change with perception and action within a given situation, have not been systematically examined. To pursue the issue of situatedness of reported ﬁnger-counting habits, 458 participants were tested in three
separate groups: (1) spontaneous condition: counting with both hands available, (2) perceptual condition: counting with horizontal (left-to-right) perceptual arrangement of ﬁngers (3) perceptual and proprioceptive condition: counting with horizontal (left-to-right) perceptual arrangement of ﬁngers and with busy dominant hand. Report of typical counting habits differed strongly between the three conditions. 28 % reported to start counting with the left hand in the spontaneous counting condition (1),
54 % in the perceptual condition (2) and 62 % in the perceptual and proprioceptive condition (3). Additionally, all participants in the spontaneous counting group showed a symmetry-based counting pattern (with the thumb as number 6), while in the two other groups, a considerable
A key insight comes from the statement that all or at least some of our knowledge representations have a situated component drawn from the circumstances of knowledge acquisition.
They suggest that cognitive reps are grounded on the physical properties in the real world - grounded cognitions, as compared to embodied cognition, which is based upon a proprioceptive feedback from our sensor motor systems as they engage with the physical world. Finally, situated cognition is associated with the tasks being performed, from contextual information presented.
Thus counting has been associated with grounded cognition; the number of objects in the world, but not so far with a situated form of cognition. Note that large numbers above us in space is consistent from culture to culture, but left to right effects in grounded cognition occur from the written language rules of the culture; left to right or right to left effects due to numerical reading order.
This differentiation is a key issue for my work in virtual worlds, as we can see effects coming from the visual perception of the worlds, in a form of grounded cognition, but in the case of desktop virtual worlds, we can only have a mediated experience of the world, so key situated cognition paths in our brains are not engaged due to lack of sensory motor feedback.
This is evidence of a layering of the cognitive effects of VWs on people, due to levels of immersion. Thus even with Kinect based interfaces etc., unless there is haptic feedback, we will lack a lot of the information required to make a full recall of information available from previous experiences of the task.
So does this indicate that the knowledge is compartmentalised into the various senses, with some relationships in between? What would change in our cognition when moving from a desktop world, to an embodied world. The idea is that the sensory motor component is strongly related to the task information; ie. you have access to this memory when you perform the task with your body.
This may explain why we cannot articulate verbally our work practices without the use of our bodies to prompt recall, even with a visual representation, we need embodiment and task context to fully activate all recall; ie. a physical role-play.
Is there another level, the differentiation between real-context and roleplaying context, the so-called suspension of disbelief, where we think the representation is real, and not a synthetic representation?
Thus the major contribution of the paper is to note the influence of task context effects on the finger counting process. So for me, the question is whether this effect continues into complex work environment tasks in processes.
So, in an experiment, we can hypothesise that for a memory test, the use of an immersive system (Occulus) will produce different results to a desktop screen virtual world. We expect that the task specified will be influenced by the part of the body used in the experiment. This could be an important concept to explore with memory and immersive virtual worlds.
Loetscher T, Schwarz U, Schubiger M, Brugger P (2008) Head turns
bias the brain’s random number generator. Curr Biol 18(2):R60–
Friday, May 9, 2014
2nd CFP: TAProViz 2014 : 3rd International Workshop on Theory and Applications of Process Visualization
2nd Call For Papers
3rd International Workshop on Theory and Applications of Process Visualization, Haifa, Israel - 08 September 2014
In conjunction with the 12th International Conference on Business Process Management BPM2014 - http://bpm2014.haifa.ac.il/
Call for Papers
Visualizations can make the structure and dependencies between elements in processes accessible in order to support users who need to analyze process models and their instances. However, effectively visualizing processes in a user-friendly way is often a big challenge, especially for complex process models which can consist of hundreds of process components (e.g., process activities, data flows, and resources) and thousands of running process instances in different execution states. Many challenges remain to be addressed within the broad area of process visualization such as: scalability, human-computer interaction, cognitive aspects, applicability of different approaches, collaboration, process evolution, run-time requirements of process instances and applications, etc.
Topics of interest include (but are not limited to):
* Visual Metaphors in Processes
* Visual Design and Aesthetics for Processes
* Visualization of Dynamic Data in Processes
* Change Visualization for Processes
* Interface and Interaction Techniques for Process Visualization
* Visualization Techniques for Collaboration and Distributed Processes
* Visualization of Large-scale Processes
* Cognition and Perception in Process Visualization
* Evaluation and User Studies of Process Visualization
* Visual Modeling Languages
* Analysis Techniques and Visualization for Processes
* Process Visualization of Large Screens
* Mobile Process Visualization
* Visualization Tools and Systems for Processes
* Visualization Techniques for Processes
* Process Visualization and Sonification
* Virtual World Process Visualization
* Immersive Process Modeling Approaches
* 3D Process Visualization Approaches
Format of the Workshop
The half day workshop will comprise accepted papers and tool evaluations. Papers should be submitted in advance and will be reviewed by at least three members of the program committee.
This year will also include a new innovation in the programme. Part of the workshop time (depending on the number of prototype submissions) will be set aside for focus group assessments of tools. We will be requesting tool report authors, successful workshop paper authors and panel members attending BPM, to assist in the assessment of demonstration visualization techniques and software. This evaluation process will be a service to attendees, as these heuristic assessments can be written up later as separate papers, or by the workshop chairs as an aggregated workshop outcome. Such evaluations will be an exciting addition to the workshop, as people experienced in Information Visualization, BPM, HCI and related fields, will provide detailed feedback on your prototypes. The evaluation approach is largely in the hands of the tool report writers, but at a minimum, should involve direct interaction with your software and some form of validation via a questionnaire.
All accepted papers will appear in the workshop proceedings published by Springer in the Lecture Notes in Business Information Processing (LNBIP) series. There will be a single LNBIP volume dedicated to the proceedings of all BPM workshops. As this volume will appear after the conference, there will be informal proceedings during the workshop. At least one author for each accepted paper should register for the workshop and present the paper.
* Deadline for workshop paper submissions: 1 June 2014
* Notification of Acceptance: 1 July 2014
* Camera-ready version: 23 July 2014
* TAProViz Workshop: 8 September 2014
Prospective authors are invited to submit papers for presentation in any of the areas listed above.
Three types of submissions are possible:
* (1) full papers (12 pages long) reporting mature research results
* (2) position papers reporting research that may be in preliminary stage that has not yet been evaluated
* (3) tool reports, to be evaluated at the workshop
Position papers and tool reports should be no longer than 6 pages. Tool reports should include a brief evaluation plan as an appendix, for the evaluation session at the workshop on the day.
Papers must be in English and must present original research contributions not concurrently submitted elsewhere. Papers should be submitted in the LNBIP format. The title page must contain a short abstract, a classification of the topics covered, preferably using the list of topics above, and an indication of the submission category (regular paper/position paper/tool report).
All accepted workshop papers will be published by Springer as a post-workshop proceedings volume in the series Lecture Notes in Business Information Processing (LNBIP). Hard copies of these proceedings will be shipped to all registered participants approximately four months after the workshops, while preliminary proceedings will be distributed during the workshop.
Submitted papers will be evaluated, in a double blind manner, on the basis of significance, originality, technical quality, and exposition. Papers should clearly establish their research contribution and the relation to the theory and application of process visualization.
Accepted papers imply that at least one of the authors will register for BPM2014 and present the paper at the TAProViz workshop.
Further workshop information is available from the website:http://www.wst.univie.ac.at/topics/taproviz14/
Hope to see you at TAProViz'14!
Thanks and best regards,
TAProViz Organising Committee