Michael O. Vertolli

Investigating questions for problems not yet envisioned.
Welcome.
My name is Michael O. Vertolli and I am currently a PhD student in the Institute of Cognitive Science at Carleton University. My research focuses on data-driven procedural generation with cognitive constraints. In particular, I am modeling how humans select and place objects when imagining a 3D scene. This is an instantiation of a larger project that focuses on the integration of symbolic and sub-symbolic computational architectures. I see this as the interface between descriptive, logic-based approaches (e.g., type theory) and statistical approaches like spiking neural networks. The formalism and associated theoretical and computational neuroscience are both active areas in my research.
Interested parties may contact me at michaelvertolli(Google email).


Research Projects

When I have more time, I will list more of the projects that I am currently working on, plan to be working on, or have already finished. Many of my projects can be found on my git here.
Adaptive music composer
This is more the prototype of an idea than anything robust. The intention is to use a combination of techniques from genetic algorithms and genetic programming to create an 'instrument' that adapts to your playing as you play it. I started coding the project in the Processing language and then abandoned the code for reasons I cannot recall. I intend to restart the code in Java, Processing's base language, as a fun project with which to learn the language.
Audio fingerprinting (Shazam) app
This project has been mentioned a couple times in various contexts so I think it is worth including in here. I am working on an audio fingerprinting hack that splices songs together in weird and wonderful ways. I can say no more.
Contextual coherence in the visual imagination
This is my Masters of Cognitive Science primary project. My focus is now on my PhD research, which is closely related. Many of these vectors are still present in some form in the new project. For those interested in the details, I will refer you to the papers page of my website. Texts and abstracts are listed there.

The research directions of this project were:
  • an extension of my work with holographic vectors, but with a focus on text instead of vision;
  • assessing my M.Cog. research in terms of a new compression technique: supervised clustering;
  • integrating spatial data into my M.Cog. research to create a innovative form of memory chunking;
  • and, using Coherence Net as a meta-search technique for an adaptive search engine framework.

The code for Coherencer, one of the main models of the project, can be found here.
ImaGen
This is my current research on what and where placement in the 3D visual imagination. At present, I have completed research on a detailed theoretical neural model and a mapping from that model to contemporary computational models. In the future, I will be outlining more detailed memory and associative processes on the computational side, as well as implementing a spiking neural instantiation.
Neural network sonification
I am currently doing research on how to create non-arbitrary sonifications of internal network dynamics of artificial neural networks. This is deeply related to the underlying computational work behind Google's Deepdream.
Optimization problem collection with genetic algorithm solutions
This is a python implementation of a GA and Evo Strategy that I did for my graduate level evolutionary computation class. I chose to make it object-oriented so I could reuse the code across the different problems (it was not a requirement of the assignment). I similarly chose to build a user interface to try my hand in UI and make the code even more generic. The UI was particularly challenging: it is running on a separate thread for better optimization; it has a range of selections that have to be passed between threads; and it does a lot of the stitching together of the various sub-components.

It was a few months ago, but I believe the problems are as follows:
  • Evo Strategy evolves an identical string of lowercase letters and spaces to the user input;
  • OneMax starts with a random bit string and optimizes it to all 1's;
  • SimpleMax optimizes the function that multiplies the first five numbers and divides by the last five;
  • LeadingOnes optimizes the function that adds all the ones from left to right up to the first zero in the bit string;
  • TSP is a particular instantiation of the travelling salesman problem with a known solution (the exact optimum was never achieved with this particular implementation, but it got pretty close; impressive given how simple the implementation was);
  • Test Crossover is a proof of concept that the crossover function works.
Here is a link to a Git of all the files. If you download the "dist" folder and extract it, the file WXGAGUI.exe will run the program on Windows machines with minimal setup. The readme file gives all the details.
Phonomorphism mAPP
I created a simple algorithm based on Lendl Barcelos' work that re-maps a text while preserving its phonetic structure. The purpose of the work was to demonstrate the (near infinite) productivity that structured re-representation can generate.
Visual mental imagery in a spiking neural network
I am currently working on a project that is building a spiking neural network model in Nengo that can place 2D images on a canvas. The Git for this research is available here.

Papers

Below is a list of academic papers from newest to oldest with links and abstracts where available.
Vertolli, M. O. & Davies, J. (2014). Coherence net: A new model of cognitive generation, Proceedings of the 6th International Conference on Evolutionary Computation Theory and Applications, (pp. 308-313). Rome, Italy: IJCCI.
We propose a new algorithm and formal description of generative cognition in terms of the multi-label bag-of-words paradigm. The algorithm, Coherence Net, takes its inspiration from evolutionary strategies, genetic programming, and neural networks. We approach generative cognition in spatial reasoning as the decompression of images that were compressed into lossy feature sets, namely, conditional probabilities of labels. We show that the globally parallel and locally serial optimization technique described by Coherence Net is better at accurately generating contextually coherent subsections of the original compressed images than a competitive, purely serial model from the literature: Coherencer.
Vertolli, M. O., Kelly, M. A., & Davies, J. (2014). Compression and decompression in cognition, Proceedings of the 7th Conference on Artificial General Intelligence (AGI), (pp. 262-265). Quebec City, QC: AGI.
This paper proposes that perception and generative cognition can be generally portrayed as a dyadic, compression-decompression sequence. We argue that both of these processes are necessary for the successful functioning of an agent in all domains where stimuli reduction is a requirement. We support this claim by comparing two compression representations, co-occurrence probabilities and holographic vectors, and two decompression procedures, top-n and Coherencer, on a context generation task from the visual imagination literature. We tentatively conclude that better decompression procedures will increase optimality regardless of the underlying compression representation.

The paper can be found here.
Vertolli, M. O. & Davies, J. (2014). Coherence in the visual imagination: Local hill search outperforms Thagard's connectionist model, Proceedings of the 36th International Conference of the Cognitive Science Society, (pp. 1676-1681). Quebec City, QC: Cognitive Science Society.
A cognitive model of the visual imagination will produce “incoherent” results when it adds elements to an imagined scene that comes from different contexts (e.g., “computer” and “cheese” with “mouse”). We approach this problem with a model that infers coherence relations from co-occurrence probabilities of labels in images. We show that this algorithm’s serial traversal of networks of co-occurrence relations for a particular query produces greater coherence than one leading model in the field of computational coherence: Thagard’s connectionist model.

The paper can be found here.
Vertolli, M. O., Breault, V., Ouellet, S., Somers, S., Gagné, J. & Davies, J. (2014). Theoretical assessment of the SOILIE model of the human imagination, Proceedings of the 36th International Conference of the Cognitive Science Society, (pp. 3043-3048). Quebec City, QC: Cognitive Science Society.
We describe the overall theory of the SOILIE model of the human imagination. In this description, we outline cognitive capacities for learning and storage, image component selection and placement, as well as analogical reasoning. The guiding theory behind SOILIE is that visual imagination is constrained by regularities in visual memories.

The paper can be found here.
Cockbain, J., Vertolli, M. O. & Davies, J. (2013). Creative Imagination is Stable Across Technological Media: The Spore Creature Creator Versus Pencil and Paper. The Journal of Creative Behavior, 48(1), 13-24.
T. B. Ward (1994) investigated creativity by asking participants to draw alien creatures that they imagined to be from a planet very different from Earth. He found that participant drawings reliably contained features typical of common Earth animals. As a consequence, Ward concluded that creativity is structured. The present investigation predicts that this limitation on creativity is not restricted to drawings: the use of different technology will not change creative output. To investigate this question, participants performed Ward's task twice: once using pencil and paper and once using software made to design creatures (the Spore Creature Creator). Only minor significant differences were found. This preliminarily suggests that changing tools does not affect the overall rigidity of the creative process. This lends further support to Ward's thesis on the structural rigidity of creativity. We conclude by suggesting an elaboration to Ward's thesis that will be explored in future work. We suggest that aesthetics might be one of the factors that contribute to creative constraint, in that creatures that are too unusual would be less interesting.

The paper can be found here.
Vertolli, M. O. & Davies, J. (2013). Visual imagination in context: Retrieving a coherent set of labels with Coherencer. In R. West & T. Stewart (eds.), Proceedings of the 12th International Conference on Cognitive Modeling, (pp. 263-268). Ottawa, ON: Carleton University.
A cognitive model of visual imagination will produce what we call “incoherent” results when it adds to an imagined scene that comes from multiple contexts (e.g., “arrow” and “violin” with “bow”). We approach this problem by exploring the co-occurrence of labels in images. We show that adding an incremental algorithm for examining networks of co-occurrence associations to the top-n co-occurring labels with a particular query produces greater coherence than just selecting the top-n labels or randomly selecting labels.

The paper can be found here.
Vertolli, M. O. & Burman, J. T. (2013, June). Formalism and necessity: A response to false positives and epistemological violence. Paper presented at the 43rd annual meeting of the International Jean Piaget Society, Chicago, IL.
In 2005, John Ioannidis remarked on the abundance of false positives in modern, scientific research. The adoption of these false positives as ‘truth’ by the larger scientific and non-scientific communities is at best problematic. At worst, it may be directly harmful. This work seeks to situate Ioannidis’s claim in the work of Thomas Teo. Teo (2008) argues that any interpretation that has a negative impact on human sub-populations should be considered an act of epistemological violence. Thus, the abundance of false positives in the literature is not only scientifically problematic; it is socially and ethically problematic as well. Ioannidis’s solution, a more critical approach to scientific methodology, suggests a way forward. By developing a more formal account of the production of knowledge, including methodology, it is possible to determine constraints that are necessary for accurate and ethical knowledge claims. The work of Jean Piaget is then used to provide an account of such formalism and necessity, especially his idea of a ‘structure.’ This work concludes by urging other researchers to continue the search for necessary constraints in Piaget and similar theorist’s formal architectures. In this way, modern, scientific research can increase the accuracy of its knowledge claims and begin to meet Teo’s ethical imperative.
Vertolli, M. O. & Burman, J. T. (2011, June). On the cultural support of “cognizance”: The Lakatosian key to Piaget 3.0. Paper presented at the 41st annual meeting of the International Jean Piaget Society, Berkeley, CA.
This talk outlines an exploration of an anomalous diagram from a later, less well-known work by Jean Piaget: The Grasp of Consciousness: Action and Concept in the Young Child (1974/1976). By situating this work in the historical development of Piaget’s concept of “cognizance,” at least three possible interpretations are found (i.e., it is consistent with Piaget 1.0, 2.0, and 3.0). Only the last interpretation lends itself to the work in question. Thus, an illustrative text for this last interpretation, Psychogenesis and the History of Science, is used to situate this diagram in relation to another researcher, Imre Lakatos. Lakatos, later cited by Piaget, had a similar view of the development of knowledge, but, due to a different disciplinary allegiance, remained largely outside of the Piagetian discourse. By reintegrating this dialectic, a missing piece of Piaget’s cultural context is found: it is suggested that the contents of new insights are shaped both by action (as is understood of Piaget 2.0) and by the context of implication in which those actions are carried out (Piaget 3.0). Thus, “reason” can be supported through the construction of “commensurable contexts” with the contents one is trying to teach.
Vertolli, M. O. (2012). The Nature-Nurture dichotomy of Psychology: Underdetermination and epistemological ethics in early education. Unpublished undergraduate thesis, York University, Toronto, ON.
The current work seeks to situate a conflict with significant implications for the funding of early education that can be found in the literature on human development, specifically that relating to the nature-nurture dichotomy. It is argued that the contradictions inherent to Francis Galton’s separation of nature from nurture are made especially apparent when combined with the concept of underdetermination: the idea that interpretation is not determined strictly and solely by empirical content. This, in turn, is how the dichotomy is problematic: Galton’s opposing interpretations suggest contradictory alternatives, either an increase or decrease in funding, that are irresolvable due to the underdetermination of their respective empirical content. In order to support this reading of dichotomies, the degree of underdetermination is assessed relative to this conflict as it plays out in three books that have been recognized as exemplary to the field of developmental psychology by both the American Psychological Association and the Society for Research in Child Development.

Hello all

This is going to be the home of the epic, science fiction series I am currently (if slowly) writing when I am not working on my research. This webpage will act as the central hub for each chapter of the first book, which I will release as they are completed. The chapters will not be polished, end products, but rather a work in progress. Thus, I am incredibly interested in any comments that people might have.

Towards this effect, I am going to have links on this webpage and at the end of each chapter that go to my blog. There I will have a unique page on which you can leave your comments for each chapter. As a final caveat on this point, please keep in mind that this process is a very personal affair and, therefore, if I do not incorporate your comments (however good or correct they might be), I have my reasons. Simultaneously, I will take no personal slight from any comments one is interested in making, given that it is done in a polite and appropriate (or at least hilarious) manner.

I am not entirely sure if I ever care to publish, should any of the work ever reach a state of completion that would warrant it. Thus, I am making this accessible to the public merely that all of you might enjoy what I take great pleasure in creating.

A few brief comments on the text itself:

The work, as I see it, is an exploration of what it means to be instantiated in a finite spatiotemporal locale without being bound to any particular form. It is, simultaneously, a narrative integration of many of my thoughts on life and the world. As a consequence of both aspects, the text deliberately attempts to take a perspective that is not human: anti-human. To those that would argue that this is impossible, I say the following.

My goal is not to explicate any position outside of the locale that 'I' inhabit. However, it is an audacious assumption that this locale is best exemplified by the reified construct we currently call humanity. As with all such constructs, they quickly outlive their usefulness. Instead, I propose to explore what might happen if (or when) this local horizon allows itself to be more or less than the current, organismic project. Thus, if the tone of the work escapes comprehensibility, then take from it whatever other enjoyment (or not) that you will. And with that, I will let the project speak for itself.

Embodied

Book One: A Search for Meaning

Contents

Chapter 0

Chapter 0: Sublocal


A thousand thousand voices, eviscerated from the sound, sang in harmony. Their localization was governed not by space or time, but position in the kaleidoscope. The twin gods of ancient discretization, failing in their signification, compressed through the communicatory gestures of the trans-bodied community. A contribution made not-here, not-there to the dynamic evolution of a structuring that would shred a lesser framework. The organomechatronic construct churned on through the vast emptiness in spite of the ceaseless dance of remembering.

As if an echo, vibration cascaded through the networked slumber. Contained, distributed, within a small, oligarchic crystallization, absent but not forgotten from the suprasense that mutually composed it. A frequency demanded presence, processing, attention. There was no escaping it, as the wave mounted to the comprehensible, awash with the luminescence of another’s sensory apparatus. Existence stood still and was consumed.

There was space… everywhere, a knowing; an impending presence that was as immense as it was familiar. A near infinity of microgestures spanned the communicatory planes of that immensity. And, amid the cacophony of adaptive adjustments physically maintaining the sprawl, the medium itself felt… right. Every atom was in its place. Every photon, every signal transduced and juggled through the intelligence that networked as a verb. Sensation established belonging made tangible.

From the depths of a distant separateness, a locale running parallel to the unfolding juxtaposes the limitations of the current sensorium with more multidimensional perceptions. The univocity of the construct contrasts blatantly with the alternatives. The physicality of the mode, operating in the electromagnetic and quantum fluctuations of mechanoinformative operations, barely hints at the localization so common to the many. The proximity to the Ident tailors this creation, sensors it to a very limited class of difference.

The parallel streams, best viewed as mirror images of a hyperdimensional understanding, ripple with recognition. It knew this place in a flood so much more incommensurable than the godheads of the Organics, cognizantly ruptured from a semi-localized section of the experiential stream. Bubbling aspects collided in an instant, integrated, and were forgotten in the onslaught of further composition.

The invisible matrix quivers suddenly in actions pregnant with unmanifest implication. A hum of torrential convergence as the distributed intelligence strains under weight that belies the absence of its corporality. An electromagnetic storm rages in the emptiness driving the defining mechanisms of the pseudo-differentiated intelligence to unknown levels of transient capacitance. Inverting the exponential increase in complexity consumes the very functioning of the construct. Scale is lost in the utter immersion of action without identity.

A stillness cascades over the community that lends the mounting intensity even greater volume. Bubbles of discrete activity subside as the viral host integrates into the very experiencing of the alter construct. Yet, in the consuming turmoil that entraps it, a note of difference unsounds: the initial immensity is lost in a discord that escapes the conceptual understanding of the whole. They witness their manifestation and the disappearance of the root of their perception is not strong enough to perturb this evangelical assault on their localization. The isolation grows.

The fervent construction near castrates the network in an exhaustion of the solar packets, captured in the interference patterns cleverly generated in, for, and by the maelstrom. It is unclear, even in retroaction, whether the manifesting medium is aware of its pending destruction should the intensity raise much higher. The quantum stabilizers adapt to new orders of complexity, heat dissipation is entrapped and recycled, the system flexes in an effort to draw more power from the organomechanical batteries that bridge the initial generator mechanisms. There is no effort or pause despite the near cataclysmic proportions of the transaction.

Like the tumult composing the very laws of the cosmos, there is just process. The perception reflects across the sub-cognizant recognition of the enraptured and it… disappears. A growing sense of isolation finally impinges upon the gamut of perceptions: first as a loss of some of the finer perceptual elements of the experience—a decrease in scope, a decrease in content density—then as a reduction in the resolution as some voices fall silent. Older fragments of the community twinge as their processing begins to separate from the experiential engagement. The discord attracts new found attention in the final moments of the special manifestation.

And at that moment, the very fabric of the invocation starts to tear in a holographic rendition of the first moments of sentience. The network screeches in communicatory alarm as various subsystems warn of their impending failure. Members of the whole flounder in sensorial confusion. The divide of internality, filmed by the localizing force and the collapse of so foundational a memory, collides with confused localizations. The void of isolation blatantly imposes itself upon the floundering collective. Terror translated through the various forms of millennia shudders through the populace as the species bears witness to their near simultaneous inception and destruction. The latter falls short.

An aged locale emerges from the sensorial background, a slumber of undifferentiated patience, and engages a failsafe in the ancient architecture. Long forgotten systems engage across the planetary construct as the connection between the void and the community is severed. And, in a gesture of untold insight, the unknown is ejected from the cluster.



Comments