331 research outputs found

    (Coarse Coding of Shape Fragments) + (Retinotopy) = Representation of Structure

    No full text
    st as tens of thousands of spoken words can be generated using a small number of phonemes as components (Biederman, 1987). The symbolic/structural approach to representation can be challenged as a model of human performance, which falls short of viewpoint invariance and is limited in other ways (Edelman, 1999). So far, it also proved to be a poor blueprint for computer vision, where no full-fledged system based on structural representations was ever implemented. This short note outlines an alternative to conventional structural representations, based on an established model of recognition and categorization: the Chorus of Prototypes (Edelman, 1998). The extended model, which I call the Chorus of Fragments (CoF), is based on the idea of combining "what" and "where" information within the same computational units. CoF aims at supporting all four core recognition-related tasks listed above, without recourse to problematic symbol-and-structure techniques. Present address:

    Psychophysical support for a 2D view interpolation theory of object recognition

    No full text
    Does the human brain represent objects for recognition by storing a series of twodimensional snapshots, or are the object models, in some sense, three-dimensional analogs of the objects they represent? One way to address this question is to explore the ability of the human visual system to generalize recognition from familiar to novel views of three-dimensional objects. Three recently proposed theories of object recognition --- viewpoint normalization or alignment of 3D models [Ullman, S. (1989) Cognition, 32, 193-254], linear combination of 2D views [Ullman, S. & Basri, R. (1990)], and view approximation [Poggio, T. & Edelman, S. (1990) Nature, 343, 263-266] --- predict different patterns of generalization to novel views. We have exploited the conflicting predictions to test the three theories directly, in a psychophysical experiment involving computer-generated 3D objects. Our results suggest that the human visual system is better described as recognizing these objects by 2D view in..

    REPRESENTATION IS SPACE-VARIANT

    No full text
    ABSTRACT: Under shift, caused for example by eye movement, or by relative movement of the subject or object of perception, the cortical representation undergoes very large changes in “size ” and “shape”. Space-variance of cortical representation rules out models which fundamentally require linear interpolation between shifted patterns (e.g. Shimon Edleman’s model), or rigid shift of an invariant retinal stimulus corresponding to shift at the cortex (e.g. the shifter theory of VanEssen). Recently, a computational solution to “quasi-shift ” invariance for space-variant mappings has been constructed [Bonmassar and Schwartz, 1997a, Bonmassar and Schwartz, 1997b]. Shimon Edelman’s (SE,from here on) work addresses an important gap in the computational discus-sion of neural representation, which to date has largely been carried out on a verbal level. His position is that representation is a record of similarities to stored prototypes, rather than direct representation in the form of templates, or feature vectors. Rather than learn all possible prototypes (similarities), a ”small” number are stored, with interpolation of new stimuli providing generalization. SE uses a particular form of cluster analysis (multi-dimensional scaling) to effect classification. No neurally plausible means of implementing multi-dimensional scaling in the brain is provided, and no comparison to other similar forms of clustering, or indeed, of statistical pattern recognition in general, is supplied. It seems to us to there is a basic mathematical equivalence between clustering based on “similarities ” and clustering based on direct feature vector representation. We will instead focus on the issue of linear interpolation of learned prototypes, which we identify as the key contribution of this model. Representation in the brain is expressed, we believe, in a wide variety of cortical loci. The majority of cortical visual area’s are topographically organized, and we assert that spatial representation in the brain (in the form of topography and columnar spatial patterns) are themselves a form of representation, and one which obviously does not depend on “similarities ” between prototypes, but which is an example o

    LANGUAGE MODELS FOR ROBOT TASK PLANNING WITH HUMAN DEMONSTRATIONS

    No full text
    78 pagesTwo major challenges exist with high-level robot task planning when the goal is under specified: humans have implicit assumptions and preferences they may not articulate when specifying a goal/reward, and visual demonstrations arehard to ground in a form robots can understand. This thesis addresses these challenges leveraging Large Language Models (LLMs) and Vision-Language Models (VLMs) to convert task demonstrations to robot code offline, then adapt to changes planning online. This work is divided into three connected components. Firstly, we developed DEMO2CODE, that generates robot code from demonstrations assumed to be grounded in text form. We focus here on its quality in capturing preferences from real-world demonstrations. Secondly, we delve into grounding visual input in text form with VIDEO2DEMO, focusing on open-vocabulary predicate and action recognition in kitchen tasks. Lastly, we build and deploy an AI task planner to allow collaborative cooking in our MOSAIC framework. We focus here on evaluating the task planner on its safety violations while interacting with participants in our user study. In summary, this thesis focuses on robot task planning and human-robot interaction by using demonstrations for effective code generation, grounding visual information in text, capturing and adhering to preferences of the human, all using generative language models

    The Textual Space Najati Sidky’s Novel of Shimon Bouzaglo as a Model

    No full text
    The study aims at highlighting the text’s public sphere of the shot story “Shimon Bouzaglou” by Najati Sidki, which depicts several artistic dimensions that include both the open sphere of the text and the place. It also shows how the text’s sphere converges with the political, psychological, economical and socio-cultural spheres. Although short, the story’s text comprises a fertile ground for several interpretations. It resembles a portrait that carries several questions related to Shimon the beggar and to the character of the Sephardic Jews. Also, it has to do with a proportion of people who go through a double personality ruptured between psychological greed and societal oppression. The study begins by analyzing the title, which does not refer to an attribute or a name of a certain author. However, mentioning and reading texts with their titles represent a cultural status for listeners, and an important state of communication , especially if a considerable number of listeners come to an agreement in recognizing a certain title. Thus, the title Shimon Buzaglo provides the opportunity towards a multiple visions and interpretations vis-à-vis the Jewish situation and how it is viewed by Palestinians. The study includes different approaches about the beginning of the text that is related to the place. It refers the reader to a destitute neighborhood known as the Tanak neighborhood whose houses are built of wood, old shingles, and interlocking houses to protect residents from rain and wind in winter, and to bring them shade and cold in summer. These of course are signs of misery and destitution. The beginning of the text plays a bridge to cross between the title “Shimon Buzaglo” and the bulk of the text in terms of both the layout and context. The study also addresses other narrative elements of the novel in terms of time, place and characters, as well as their contextual, cultural and interpretive implications pertaining to history, thought and daily events

    Representation is Representation of Similarities

    No full text
    Intelligent systems are faced with the problem of securing a principled (ideally, veridical) relationship between the world and its internal representation. I propose a unified approach to visual representation, addressing both the needs of superordinate and basic-level categorization and of identification of specific instances of familiar categories. According to the proposed theory, a shape is represented by its similarity to a number of reference shapes, measured in a high-dimensional space of elementary features. This amounts to embedding the stimulus in a low-dimensional proximal shape space. That space turns out to support representation of distal shape similarities which is veridical in the sense of Shepard's (1968) notion of second-order isomorphism (i.e., correspondence between distal and proximal similarities among shapes, rather than between distal shapes and their proximal representations). Furthermore, a general expression for similarity between two stimuli, based on comparisons to reference shapes, can be used to derive models of perceived similarity ranging from continuous, symmetric, and hierarchical, as in the multidimensional scaling models (Shepard, 1980), to discrete and non-hierarchical, as in the general contrast models (Tversky, 1977; Shepard and Arabie, 1979)

    How Representation Works Is More Important Than What Representations Are.

    No full text
    A theory of representation is incomplete if it states "representations are X" where X can be symbols, cell assemblies, functional states, or the flock of birds from Theaetetus, without explaining the nature of the link between the universe of X's and the world. Amit's thesis, equating representations with reverberations in Hebbian cell assemblies, will only be considered a solution to the problem of representation when it is complemented by a theory of how a reverberation in the brain can be a representation of anything

    Receptive Fields for Vision: from Hyperacuity to Object Recognition

    No full text
    Many of the lower-level areas in the mammalian visual system are organized retinotopically, that is, as maps which preserve to a certain degree the topography of the retina. A unit that is a part of such a retinotopic map normally responds selectively to stimulation in a well-delimited part of the visual field, referred to as its receptive field (RF). Receptive fields are probably the most prominent and ubiquitous computational mechanism employed by biological information processing systems. This paper surveys some of the possible computational reasons behind the ubiquity of RFs, by discussing examples of RF-based solutions to problems in vision, from spatial acuity, through sensory coding, to object recognition

    Representation, Similarity, and the Chorus of Prototypes

    No full text
    It is proposed to conceive of representation as an emergent phenomenon that is supervenient on patterns of activity of coarsely tuned and highly redundant feature detectors. The computational underpinnings of the outlined concept of representation are (1) the properties of collections of overlapping graded receptive fields, as in the biological perceptual systems that exhibit hyperacuity-level performance, and (2) the sufficiency of a set of proximal distances between stimulus representations for the recovery of the corresponding distal contrasts between stimuli, as in multidimensional scaling. The present preliminary study appears to indicate that this concept of representation is computationally viable, and is compatible with psychological and neurobiological data
    corecore