Photo-Works is the first major exhibition in the United States by German artists Anna and Bernhard Blume. Organized by Dean Sobel, curator of contemporary art, and Tom Bamberger, adjunct curator at the Milwaukee Art Museum, the exhibition at the Milwaukee Art Museum was comprised of 10 large photography installations, 50 medium scale photographs and 40 Polaroid collages. Accompanied by the Blumes' first English language monograph, the catalog contains 200 black and white reproductions, 40 full color reproductions, essays by the exhibition curators, an interview with the artists and essays by German scholars Bazon Brock and Wolfgang Kasprzik. The 10 photo installations are monumental, billboard-size graphic novels that fill a cavernous, columned, long open hall of the Milwaukee Art Museum. The images document a performance of pathological disorientation and the ecstasy of the oblivious state of mind of tourists and the German lower-middle class. As both director and actor, the Blumes dress up in the stereotypical garb of the lower-middle class and photograph themselves. Like any performance it is best to encounter the Blumes' work in person. Though clearly necessary, since there is no English language documentation of the Blumes' work, the catalog is for the most part disappointing, visually and verbally. The images are reproduced with far more contrast than the original prints and the translation of the Blumes' artist statements is questionably arch and borders on the simple-minded. For example, Anna Blume's assertion about Home Sweet Home: Today's observations, mental and visual, are again what we call a photo-document. Fundamentally it was produced in my kitchen, and in space. Quite casually, of course! Since the last kitchen rage, potato chaos, French fries disaster we have been prepared for extraordinary events, in other words: the camera was there this time as well!. . . Putting aside the artists' statements for a moment, let us focus on the exhibition and the essays by Sobel, Bamberger, Brock and Kasprzik. In Waldeslust (Forest Pleasures, 1982-83) Anna wears a loudly patterned floral dress. Her hair is a hair-sprayed helmet. Bernhard wears a plaid jacket, dress shirt (buttoned all the way up without a tie) and dress slacks, and holds Anna's large white pocketbook. They smile a big toothy grin as they pose in front of an acid rain devastated forest for a snapshot record of their holiday. They seem almost proud and clearly unaware of the death that surrounds them. Their inappropriate dress clearly points at their ignorance of and inability to relate to nature. The absurdity of their pose and expressions is mind-boggling and disturbing. The work's humor is angry and cynical, evoking an echoing laughter from viewers. The monumental scale of the work (each photograph is 98-1/2x50[inches] and each installation is made up to six to eight photographs) points to the scale and weight of the subject matter, which is the loss of and disconnection from nature. The Blumes' tableaux seem to document a cult from a lost civilization; visitors to the museum come upon the images like murals in an ancient, awe-inspiring tomb. The grid system format of the installations orders the chaotic composition within individual frames, thus emphasizing the notion of disconnection. The apocalyptic message is sobering. The loss conveyed is profound. The scenes of the tourist in the images are real. Forests are being devastated. People are oblivious. Now the laughter at someone else's expense evokes guilt and embarrassment as the roles of the viewer and tourist are reversed. Kasprzik's point is right on target: We are forced to play this game as long as we have the strong feeling that those people do not really think whereas we, being able to judge about them, do. The absurd is also heightened by the Blumes' adept use of sight gags and physical humor. Hysteria crosses the line into neurosis most profoundly in Kuchenkoller (Kitchen Frenzy, 1985) where potatoes take flight of their own volition and Anna (as hausfrau) displays an expression - like a housewife encountering talking scrubbing bubbles in a television advertisement. …

By pointing to deep philosophical confusions endemic to cognitive science, Wittgenstein might seem an enemy of computational approaches. We agree (with Mills 1993 ) that while Wittgenstein would reject the classicist’s symbols and rules approach, his observations align well with connectionist or neural network approaches. While many connectionisms that dominated the later twentieth century could fall prey to criticisms of biological, pedagogical, and linguistic implausibility, current connectionist approaches can resolve those problems in a Wittgenstein-friendly manner. We (a) present the basics of a Vector Symbolic Architecture formalism, inspired by Smolensky ( 1990 ), and indicate how high-dimensional vectors can operate in a context-sensitive and object-independent manner in biologically plausible time scales, reflecting Wittgenstein’s notions of language-games and family resemblance; we (b) show how “soft” symbols for such a formalism can be formed with plausible learning cycles using Sparse Distributed Memory, resolving disputes surrounding Wittgenstein’s private language argument; and (c) show how connectionist networks can extrapolate meaningful patterns to solve problems, providing “ways to go on” without explicit rules, which indicates linguistic plausibility. Connectionism thus provides a systematicity and productivity that is more than a mere implementation of a classical approach, and provides Wittgenstein-friendly and Wittgenstein-illuminating models of mind and language for cognitive science.

In attempting to build neurorobotic systems based on flying animals, engineers have come to rely on existing firmware and simulation tools designed for miniature aerial vehicles (MAVs). Although they provide a valuable platform for the collection of data for Deep Learning and related AI approaches, such tools are deliberately designed to be general (supporting air, ground, and water vehicles) and feature-rich. The sheer amount of code required to support such broad capabilities can make it a daunting task to adapt these tools to building neurorobotic systems for flight. In this paper we present a complementary pair of simple, object-oriented software tools (multirotor flight-control firmware and simulation platform), each consisting of a core of a few thousand lines of C++ code, that we offer as a candidate solution to this challenge. By providing a minimalist application programming interface (API) for sensors and PID controllers, our software tools make it relatively painless for engineers to prototype neuromorphic approaches to MAV sensing and navigation. We conclude our discussion by presenting a simple PID controller we built using the popular Nengo neural simulator in conjunction with our flight-simulation platform.

We present work-in-progress on a novel, open-source software platform supporting Deep Reinforcement Learning (DRL) of flight behaviors for Miniature Aerial Vehicles (MAVs). By using a physically realistic model of flight dynamics and a simple simulator for high-frequency visual events, our platform avoids some of the shortcomings associated with traditional MAV simulators. Implemented as an OpenAI Gym environment, our simulator makes it easy to investigate the use of DRL for acquiring common behaviors like hovering and predation. We present preliminary experimental results on two such tasks, and discuss our current research directions. Our code, available as a public github repository, enables replication of our results on ordinary computer hardware.

We advocate for a novel connectionist modeling framework as an answer to a set of challenges to AGI and cognitive science put forth by classical formal systems approaches. We show how this framework, which we call Vector Symbolic Architectures, or VSAs, is also the kind of model of mental activity that we arrive at by taking Ludwig Wittgenstein’s critiques of the philosophy of mind and language seriously. We conclude by describing how VSA and related architectures provide a compelling solution to three central problems raised by Wittgenstein in the Philosophical Investigations regarding rule-following, aspect-seeing, and the development of a “private” language.

We propose a knowledge-representation architecture allowing a robot to learn arbitrarily complex, hierarchical / symbolic relationships between sensors and actuators. These relationships are encoded in high-dimensional, low-precision vectors that are very robust to noise. Low-dimensional (single-bit) sensor values are projected onto the high-dimensional representation space using low-precision random weights, and the appropriate actions are then computed using elementwise vector multiplication in this space. The high-dimensional action representations are then projected back down to low-dimensional actuator signals via a simple vector operation like dot product. As a proof-of-concept for our architecture, we use it to implement a behavior-based controller for a simulated robot with three sensors (touch sensor, left/right light sensor) and two actuators (wheels). We conclude by discussing the prospects for deriving such representations automatically.

AI models are often categorized in terms of the connectionist vs. symbolic distinction. In addition to being descriptively unhelpful, these terms are also typically conflated with a host of issues that may have nothing to do with the commitments entailed by a particular model. A more useful distinction among cognitive representations asks whether they are local or distributed (van Gelder 1999). Traditional symbol systems (grammar, predicate calculus) use local representations: a given symbol has no internal content and is located at a particular address in memory. Although well understood and successful in a number of domains, traditional representations suffer from brittleness. The number of possible items to be represented is fixed at some arbitrary hard limit, and a single corrupt memory location or broken pointer can wreck an entire structure. In a distributed representation, on the other hand, each entity is represented by a pattern of activity distributed over many computing elements, and each computing element is involved in representing many different entities (Hinton 1984). Such representations have a number of properties that make them attractive for knowledge representation (McClelland, Rumelhart, & Hinton 1986): they are robust to noise, degrade gracefully, and support graded comparison through distance metrics. These properties enable fast associative memory and efficient comparison of entire structures without unpacking the structures into their component parts. This article provides an overview of distributed representations, setting the approach in its historical context. The two essential operations necessary for building distributed representation of structures – binding and bundling – are described. We present example applications of each model, and conclude by discussing the current state of the art.

In October 2004, approximately 30 connectionist and nonconnectionist researchers gathered at a AAAI symposium to discuss and debate a topic of central concern in artificial intelligence and cogniti...

After remaining more or less in the background for several decades, the issue of recursion in human languages has returned to the fore. The current, vigorous debate is about not just recursion, but the very idea of linguistic universals as a coherent scientific notion, and the nature of language itself. The prevailing view is that recursion – the ability to produce an unbounded number of well-formed sentences through the combinatorial operations of syntax – is a fundamental property of the “faculty of language in the narrow sense” (FLN), i.e., that recursion is what differentiates human language from other forms of animal communication (Hauser, Chomsky, and Fitch 2002). As an explicit challenge this view, Everett (2005) has documented the syntax of Piraha, a language that appears to have no embedded clauses, recursive possessives, or other syntactic structures suggesting this capacity. Everett offers Piraha as a fatal counterexample to the universality of recursion, and questions the utility of searching for other sorts of linguistic universals as well. Although this debate has involved linguistics, and sometimes politics, we suggest that it might also be fruitfully informed by neuroscience. In particular, we will present a model of event semantics based on neurally plausible role/filler bindings, and describe how recursion, like other syntactic processes, can be viewed as the by-product of “unpacking” and serializing the contents of these representations. A rough outline of our argument: we start by enumerating some basic assumptions or ground rules about what any model of human language and cognition should look like. We then provide a review of simple vector arithmetic for the benefit of readers unfamiliar with this branch of mathematics, and we show how high-dimensional vectors can be used as a plausible model of role/filler semantics and other mental representations without violating any of our stated assumptions. Next we sketch the rudiments of a model of language that treats grammar as a process of transduction between these mental representations and sequences of symbols. We present a brief account of how this language model can be related to traditional syntactic categories like agent and theme, and conclude by setting the model in the context of current research on the sources of linguistic universals.

In this position paper we argue that BICA must simultaneously be compatible with the explanation of human cognition and support the human design of artificial cognitive systems. Most cognitive neuroscience models fail to provide a basis for implementation because they neglect necessary levels of functional organisation in jumping directly from physical phenomena to cognitive behaviour. Of those models that do attempt to include the intervening levels, most either fail to implement the required cognitive functionality or do not scale adequately. We argue that these problems of functionality and scaling arise because of identifying computational entities with physical resources such as neurons and synapses. This issue can be avoided by introducing appropriate virtual machines. We propose a tool stack that introduces such virtual machines and supports design of cognitive architectures by simplifying the design task through vertical modularity.

We are concerned with the practical feasibility of the neural basis of analogical mapping. All existing connectionist models of analogical mapping rely to some degree on localist representation (each concept or relation is represented by a dedicated unit/neuron). These localist solutions are implausible because they need too many units for human-level competence or require the dynamic re-wiring of networks on a sub-second time-scale. Analogical mapping can be formalised as finding an approximate isomorphism between graphs representing the source and target conceptual structures. Connectionist models of analogical mapping implement continuous heuristic processes for finding graph isomorphisms. We present a novel connectionist mechanism for finding graph isomorphisms that relies on distributed, high-dimensional representations of structure and mappings. Consequently, it does not suffer from the problems of the number of units scaling combinatorially with the number of concepts or requiring dynamic network re-wiring.

We are concerned with the practical feasibility of the neural basis of analogical mapping. All existing connectionist models of analogical mapping rely to some degree on localist representation (each concept or relation is represented by a dedicated unit/neuron). These localist solutions are implausible because they need too many units for human-level competence or require the dynamic re-wiring of networks on a sub-second time-scale. Analogical mapping can be formalised as finding an approximate isomorphism between graphs representing the source and target conceptual structures. Connectionist models of analogical mapping implement continuous heuristic processes for finding graph isomorphisms. We present a novel connectionist mechanism for finding graph isomorphisms that relies on distributed, high-dimensional representations of structure and mappings. Consequently, it does not suffer from the problems of the number of units scaling combinatorially with the number of concepts or requiring dynamic network re-wiring.

We present a fully distributed connectionist architecture sup- porting lateral inhibition / winner-takes all competition. All items (individuals, relations, and structures) are represented by high-dimensional distributed vectors, and (multi)sets of items as the sum of such vectors. The architecture uses a neurally plausible permutation circuit to support a multiset intersec- tion operation without decomposing the summed vector into its constituent items or requiring more hardware for more com- plex representations. Iterating this operation produces a vector in which an initially slightly favored item comes to dominate the others. This result (1) challenges the view that lateral in- hibition calls for localist representation; and (2) points toward a neural implementation where more complex representations do not require more complex hardware. In this article we will argue that localist representations are not necessary to support winner-takes-all competition or lateral inhibition in general. We will present a fully dis- tributed connectionist architecture supporting lateral inhibi- tion / winner-takes all behavior, in which all items (indi- viduals, relations, and structures) are represented by high- dimensional distributed vectors, and (multi)sets of items as the sum of such vectors. Unlike a localist representation, such representations are based on a fixed neural architecture that does not need to grow as new representational categories are added.

We provide an overview of Vector Symbolic Architectures (VSA), a class of structured associative memory models that offers a number of desirable features for artificial general intelligence. By directly encoding structure using familiar, computationally efficient algorithms, VSA bypasses many of the problems that have consumed unnecessary effort and attention in previous connectionist work. Example applications from opposite ends of the AI spectrum --visual map-seeking circuits and structured analogy processing --attest to the generality and power of the VSA approach in building new solutions for AI.

This undergraduate honors thesis focuses on the problem that machine translators face in choosing the correct translation of a polysemous English noun in a foreign language. For instance, general-purpose internet translators such as Google Translate cannot distinguish between the meanings of the English noun bat, generating the French word for the baseball implement rather than that for the flying mammal - even when provided with such contextual hints as "wings" and "cave." Latent Semantic Analysis (LSA; Foltz & Laham 1998) offers a solution for this issue. LSA is a well-developed technique and theory for relating words and meanings by analyzing text corpora. A multidimensional vector represents each word and each sentence or other contextual block, and similarity or disparity of meaning can then be calculated by the relative angles of these vectors. LSA uses singular value decomposition (Golub & vanLohn 1996) on matrices constructed from words and passages in the learning texts to reduce the number of dimensions in these vectors. Such a dimensional reduction has been shown to be capable of producing simulations of human contextual associations significantly better than simple proximity frequencies.

Semantic roles describe “who did what to whom” and as such are central to many subfields of AI and cognitive science. Each subfield or application tends to use its own “flavor” of roles. For analogy processing, logical deduction, and related tasks, roles are usually specific to each predicate: for loves there is a LOVER and a BELOVED, for eats an EATER and an EATEN, etc. Language modeling, on the other hand, requires more general roles like AGENT and PATIENT in order to relate form to meaning in a parsimonious way. Commitment to a particular type of role makes it difficult to model processes of change, for example the change from specific to general roles that seems to take place in language learning. The use of semantic features helps solve this problem, but still limits the nature and number of changes that can take place. This paper presents a new model of semantic role change that addresses this problem. The model uses an existing technique, Holographic Reduced Representation (HRR) for representing roles and their fillers. Starting with specific roles, the model learns to generalize roles through exposure to language data. The learning mechanism is simple and efficient, and is scaling properties are well-understood. The model is able to learn and exploit new representations without losing the information from existing ones. We present experimental data illustrating these principles, and conclude with by discussing some implications of the model for the issues of changing representations as a whole.

Semantic roles describe "who did what to whom" and as such are central to analogy processing and other cognitive pro- cesses. For analogy processing, roles are usually specific to each predicate: for loves there is a LOVER and a BELOVED, for eats an EATER and an EATEN, etc. Language modeling, on the other hand, requires more general roles like AGENT and PATIENT in order to relate form to meaning in a parsimonious way. This paper presents a new model of semantic roles that addresses this dichotomy. The model uses a distributed repre- sentation scheme called Vector Symbolic Architectures (VSA) for representing roles and their fillers. Starting with specific roles, the model learns to generalize roles through exposure to language data, through a process that is itself analogical. The learning mechanism is simple and efficient, and its scal- ing properties are well-understood. The model is able to learn and exploit new representations without losing the informa- tion from existing ones. The contribution of the model to the study of analogy is thus twofold: it shows how representations needed for analogy processing can be accommodated within a more general theory of semantic roles, and suggests how im- portant analogy may be to language learning. We present ex- perimental data illustrating these principles, and conclude by discussing some implications for the relation between analog- ical processing and language.

We present a neural-competitive learning model of language evolution in which several symbol sequences compete to signify a given propositional meaning. Both symbol sequences and propositional meanings are represented by high-dimensional vectors of real numbers. A neural network learns to map between the distributed representations of the symbol sequences and the distributed representations of the propositions. Unlike previous neural network models of language evolution, our model uses a Kohonen Self-Organizing Map with unsupervised learning, thereby avoiding the computational slowdown and biological implausibility of back-propagation networks and the lack of scalability associated with Hebbian-learning networks. After several evolutionary generations, the network develops systematically regular mappings between meanings and sequences, of the sort traditionally associated with symbolic grammars. Because of the potential of neural-like representations for addressing the symbol-grounding problem, this sort of model holds a good deal of promise as a new explanatory mechanism for both language evolution and acquisition.

We demonstrate how a first-person shooter (FPS) video game can be made more fun and challenging by replacing the hard-wired behavior of opponents with behaviors evolved via an evolutionary algorithm. Using the open-source FPS game Cube as a platform, we replaced the agents' (opponents) hard-wired behavior with binary "DNA" supporting a much richer variety of agent responses. Survival-of-the-fittest ensured that only those agents whose DNA allowed them to avoid being killed by the human player would continue on to the next "generation" (game). Mutating the DNA of the survivors provided enough variability in behavior to make the agent's actions unpredictable. Our demo will show how this approach produces an increasingly challenging level of play, more fine-tuned to the skills of an individual human player than the traditional approach using pre-programmed levels of difficulty or simply adding more opponents.

Matlab is the most popular platform for rapid prototyping and development of scientific and engineering applications. A typical university computing lab will have Matlab installed on a set of networked Linux workstations. With the growing availability of distributed computing networks, many third-party software libraries have been developed to support parallel execution of Matlab programs in such a setting. These libraries typically run on top of a message-passing library, which can lead to a variety of complications and difficulties. One alternative, a distributed-computing toolkit from the makers of Matlab, is prohibitively expensive for many users. As a third alternative, we present PECON, a very small, easy-to-use Matlab class library that simplifies the task of parallelizing existing Matlab programs. PECON exploits Matlab’s built-in Java Virtual Machine to pass data structures between a central client and several “compute servers” using sockets, thereby avoiding reliance on lower-level messagepassing software or disk i/o. PECON is free, open-source software than runs ”out of the box” without any additional installation or modification of system parameters. This arrangement makes it trivial to parallelize and run existing applications in which time is mainly spent on computing results from small amounts of data. We show how using PECON for one such application – a genetic algorithm for evolving cellular automata – leads to linear reduction in execution time. Finally, we show an application – computing the Mandelbrot set – in which element-wise matrix computations can be performed in parallel, resulting in dramatic speedup.