D. Chapman. Vision, Instruction, and Action. MIT Press, Cambridge, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... organization, for review see Magoun (1958) Berntson, Boysen and Cacioppo (1993) Prescott et al. Layered Control Architectures 6 perception and action a perspective which, while radical for his era, 3 is clearly in sympathy with recent proponents of situated action (e.g. Brooks, 1990, 1995; Chapman, 1991): A man physically regarded is a sensorimotor mechanism. I particularly wish to insist that the highest centers physical basis of mind or consciousness have this kind of constitution, that they represent innumerable different impressions and movements of all parts of the body [ It may ....

Chapman, D. (1991). Vision, instruction, and action. Cambridge, MA: MIT Press.

.... single agent architectures and have acted as a springboard for research in Distributed AI (Adler et al. 1992) Various models and notions of agents are currently found in diverse literature in the fields of autonomous and distributed systems (Meinkoehn and Knoll 1993) language action systems (Chapman 1991), and graphics animation (Badler et al. 1991) Of interest, in our context, is work on interface agents (Laurel 1990; Maes and Kozierok 1993) which typically use knowledge about tasks, habits, preferences of their users to perform actions on their behalf. Such an idea has been adopted in visions ....

Chapman, D. (1991). Vision, Instruction, and Action. Cambridge (MA): MIT Press.

....behavior becomes overwhelmingly difficult to control when the behavioral set becomes large as a result of the combinatorial complexity of specifying interactions among behaviors or in finding reasonable behavior weights. Many researchers have pointed out this problem including Brooks [31] Chapman [38], Maes [93] and Mataric [101] Some have tried to solve this problem through the application of learning algorithms. Unfortunately, current learning algorithms tend to be too general, requiring domain specific knowledge to be effective in the particular domain. Our solution is parsimony. We ....

David Chapman. Vision, Instruction, and Action. The MIT Press, 1991.

....Interface User interface User Controller Event management Data Control FIGURE 1. Generic ibot architecture FIGURE 2. Source data for vision system computational account of high level vision proposed by Ullman [Ullman, 1996] and an implementation of many of these ideas by Chapman [Chapman, 1991] . As will be clear from the discussion below, our work also adopts two conventional assumptions in vision modeling research [Marr, 1982] First, despite the considerable differences between biological systems and computer systems at a hardware level, we can nevertheless develop computational ....

Chapman, David 1991. Vision, instruction, and action. MIT Press, Cambridge, MA.

....is a central topic in a number research projects. Torrance ( Tor94] presents a hybrid system, consisting of a symbolic and a behavior based level, for the navigation of a mobile robot. After preprocessing, natural language instructions control the robot via a planning component. Chapman ( Cha91] describes a behavior based system, which controls a video game autonomously. A human observer can influence the system by natural language instructions. The instructions are send to the behavior system directly, where they are treated like other sensor data. Stopp and Blocher ( SB97] ....

D. Chapman. Vision, Instruction, and Action. MIT Press, Cambridge, MA, 1991.

....functional outline of biological visual processing, as we describe below in more detail. Our discussion is influenced heavily by the computational account of high level vision proposed by Ullman [Hildreth and Ullman, 1989; Ullman, 1996] and an implementation of many of these ideas by Chapman [Chapman, 1991] . In the first subsection below we briefly discuss the relationship between biological vision and computational vision. In the remaining subsections we describe the structure and processing of a model for perception in a graphical user interface, from low level vision through high level vision ....

Chapman, David 1991. Vision, instruction, and action. MIT Press, Cambridge, MA.

....our research aligns with recent developments in the field of artificial intelligence concerning methods for representing and controlling action. While there is still considerable controversy between employing deep reasoning [50, 101, 149] or reactive models in the design of intelligent creatures [16, 33, 37, 99], recent works have tried to find specific domains where middle term solutions can be applied [39, 130] Our work on action representation is based on the use of a small number of symbolic primitives (based on [149] and on a simple reasoning system (inspired by [144] that compensates for the ....

.... typically experience difficulties when connected to real world perceptual data, typically present in movement detection) As an answer to this problem, procedural representations for actions have been proposed in different ways by Brooks [31, 32] Rosenchein and Kaelbling [146] and Chapman [37]. Actions in these systems are represented by the connections between actuators and sensors directly, but these representations present difficulties for the exploration of any kind of symbolic knowledge, making them vulnerable to context change. Our work has been taking a middle ground between the ....

D. Chapman. Vision, Instruction, and Action. The MIT Press, Cambridge, Massachusetts. 1991.

.... Some theoretical aspects of natural language communication with robot systems from the perspective of computer linguistics are discussed in [Lobin 92] Other approaches have been concerned with natural language control of autonomous agents within simulated 2D or 3D environments [Badler et al. 91; Chapman 91; Vere Bickmore 90] One salient aspect for natural language access to robot systems is the relationship between sensory information and verbal descriptions. Such issues have already been investigated in the field of integrated natural languageand vision processing [Bajcsy et al. 85; ....

D. Chapman. Vision, Instruction, and Action. Cambridge, MA: MIT Press, 1991.

....of the values of actions or states can be accurately stored 1 . Although several straightforward combinations of generalising function approximators and reinforcement learning algorithms have had remarkable success in complex domains [14,15,1,12] not all such results have been as encouraging [8,4,3,2]. Reinforcement learning methods based on dynamic programming operate by estimating the value of actions. The discounted techniques usually used attach values to actions in such a way that the difference in value between actions available at a state fades exponentially as distance between a state ....

.... successful applications of discounted reinforcement learning to complex domains appear to have been in domains such as Back gammon and Go [14,15,1,12] where the presence of another player ensures that the effect of actions can generally not be immediately reversed, while unsuccessful applications [8,4,3] have been often been in domains where most actions have inverses. These results suggest a number of topics for further research. ffl It may be possible to dramatically reduce approximation error in many domains by employing function approximators that can be expected to perform well in a domain ....

D. Chapman. Vision, Instruction, and Action. M.I.T. Press, Cambridge, Mass, 1991.

....behaviour. The most notable of these is the Subsumption architecture (Brooks, 1986) where the control system of a robot is composed of a hierarchy of task achieving behaviours in which higher levels of behaviour can subsume lower levels. The concrete situated approach (Agre Chapman, 1987; Chapman, 1991) formulates the control system for a robot as a collection of action proposing modules. Conflicts between proposals for external actions are resolved through a fixed priority scheme. In the situated automata approach (Kaelbling Rosenschein, 1990) a fixed ranking of goal priorities and a set of ....

Chapman, D. 1991. Vision, Instruction, and Action. MIT Press.

....that is relevant is defined in terms of single tasks, rather than in interaction with an environment over time. 1.2. 2 Situated Action models Discontents with planning models have led to an enormous variety of research in recent years in this section we will focus on situated action models [1, 4], and leave connections with closely related work to the reader. This line of work stresses the interaction of agents and their environments, and their mutual dependence. It argues that design of an all purpose agent is not feasible, and that it is incumbent on the designer of an agent to ....

D. Chapman. Vision, Instruction, and Action. The MIT Press, Cambridge, Massachusetts, 1991.

....the updating is fast. A control system based on the architecture controls a mobile robot equipped with a stereo gaze platform. As far as we know, this is the first behavior based control system executing real time stereo tracking. The architecture was inspired by the work of Brooks [6] and Chapman [7]. We have extended Chapman s work by attentive behaviors for the 3D domain. Further, we have integrated image space data with mobile space data. The markers also have arbitration functionality. Our architecture has similarities with the Alliance architecture [5] The motivational behaviors in ....

Chapman, D (1991) Vision, instruction, and action, MIT Press, ISBN 0-262- 03181-7.

....the training process, but as with the Mostow s foo system, our system does not have the operationalization capability of Gordon and Subramanian s system. Developing robot programming languages A number of researchers have introduced languages for programming robot like agents (e.g. Brooks, 1990; Chapman, 1991; Gat, 1991; Kaelbling, 1987; Nilsson, 1994) These systems do not generally focus on programming agents that learn to refine their programs. Suppes (1991) investigated how a robot can learn to understand a human s instructions, but he focuses on the problem of understanding natural language, ....

Chapman, D. (1991). Vision, instruction, and action. MIT Press, Cambridge, MA.

....of these benefits. 5 SA and other disciplines The SA perspective was not invented within AI, nor is AI the only discipline to which is relevant. In this section I very briefly discuss work in disciplines other than computer science that has been influenced by the SA perspective. See [Agr88, Cha91] for comprehensive reviews. Industrial and urban engineering Norman [Nor88] investigates some problems with the everyday artifacts produced by modern industrial designers, and argues that theories of design are insufficiently based on ecologically valid psychological theories. Horswill [Hor93a] ....

D. Chapman. Vision, Instruction, and Action. MIT Press, 1991.

....process of planning should be brief and the results sketchy any detailed predictions will most likely be wrong. The reconsideration of traditional planning has already begun under the new 1. Reconsidering planning 2 research approach termed situated activity or reactive planning (Agre, 1988; Chapman, 1991; Firby, 1989) The adjectives situated and reactive are illustrative of the tenets of this new view. Instead of building a disembodied planner, an autonomous agent is constructed and situated in realistic environments. The agent must constantly react to the dynamic world while still pursuing its ....

Chapman, D. (1991). Vision, Instuction, and Action. MIT Press, Cambridge, MA.

.... closely to perception, such as language for describing simple shapes and their spatial relations, as used in speech input enhanced drawing tools and in learning models (Regier 1996) ffl T8 suggests that another good place to start studying language is its effects on lower level motion behaviors (Chapman 1991). ffl T2, T3, T4, T5, T6, T7 and T8 suggest that language in social interaction, not just language for talking about things, is also interesting to study. This last topic, language in social interaction, is the focus of this paper. It reports on some analysis and an experiment in this area, ....

Chapman, David (1991). Vision, Instruction, and Action. MIT Press.

No context found.

D. Chapman. Vision, Instruction, and Action. MIT Press, Cambridge, 1992.

No context found.

David Chapman. Vision, Instruction, and Action. MIT Press, Cambridge, 1992.

No context found.

Chapman, D. Vision, Instruction, and Action, MIT Press, New York, 1991.

No context found.

Chapman, David, Vision, Instruction, and Action (Cambridge, MA: MIT Press), 1991, 311 pp.

No context found.

David Chapman. Vision, Instruction, and Action. The MIT Press, 1991.

No context found.

D. Chapman, 1991, Vision, Instruction, and Action. MIT Press, MA, USA.

No context found.

D. Chapman, (1991). Vision, Instruction, and Action. MIT Press, Cambridge, MA.

No context found.

Chapman, D.: Vision, Instruction, and Action. MIT Press, Cambridge, MA, 1991. 6

No context found.

D. Chapman, Vision, Instruction, and Action, The MIT Press, Cambridge, Massachusetts, 1991.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC