Abstract—We describe an embodied cognitive system based on a three-level architecture that includ... more Abstract—We describe an embodied cognitive system based on a three-level architecture that includes a sensorimotor layer, a mid-level layer that stores and reasons about object-action episodes, and a high-level symbolic planner that creates abstract action plans to be realised and possibly further specified by the lower levels. The system works in two modes, exploration and plan execution, that both make use of the same architecture. We give results of different sub-processes as well as their interaction. In particular, we describe the ...
... Author(s) Morten Kjærgaard, Alex Bierbaum, Dirk Kraft, Sinan Kalkan, Norbert Krüger, Tamim As... more ... Author(s) Morten Kjærgaard, Alex Bierbaum, Dirk Kraft, Sinan Kalkan, Norbert Krüger, Tamim Asfour, Rüdiger Dillmann ... The sensor technology used for force sensors relies either on a variant of the FSR (Force Sensing Resistor) principle [3][4][1] or on the capacitive effect [2 ...
ECCV 2010 Workshop on “Vision for Cognitive Tasks, 2010
Abstract. We discuss the need of an elaborated in-between stage bridging early vision and cogniti... more Abstract. We discuss the need of an elaborated in-between stage bridging early vision and cognitive vision which we call 'Early Cognitive Vision'(ECV). This stage provides semantically rich, disambiguated and largely task independent scene representations which can be used in many contexts. In addition, the ECV stage is important for generalization processes across objects and actions. We exemplify this at a concrete realisation of an ECV system that has already been used in variety of application domains.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000
Computational modeling of the primate visual system yields insights of potential relevance to som... more Computational modeling of the primate visual system yields insights of potential relevance to some of the challenges that computer vision is facing, such as object recognition and categorization, motion detection and activity recognition, or vision-based navigation and manipulation. This paper reviews some functional principles and structures that are generally thought to underlie the primate visual cortex, and attempts to extract biological principles that could further advance computer vision research. Organized for a computer vision audience, we present functional principles of the processing hierarchies present in the primate visual system considering recent discoveries in neurophysiology. The hierarchical processing in the primate visual system is characterized by a sequence of different levels of processing (on the order of 10) that constitute a deep hierarchy in contrast to the flat vision architectures predominantly used in today's mainstream computer vision. We hope that the functional description of the deep hierarchies realized in the primate visual system provides valuable insights for the design of computer vision algorithms, fostering increasingly productive interaction between biological and computer vision research.
IEEE Transactions on Autonomous Mental Development, 2015
ABSTRACT It is now widely accepted that concepts and conceptualization are key elements towards a... more ABSTRACT It is now widely accepted that concepts and conceptualization are key elements towards achieving cognition on a humanoid robot. An important problem on this path is the grounded representation of individual concepts and the relationships between them. In this article, we propose a probabilistic method based on Markov Random Fields to model a concept web on a humanoid robot where individual concepts and the relations between them are captured. In this web, each individual concept is represented using a prototype-based conceptualization method that we proposed in our earlier work. Relations between concepts are linked to the cooccurrences of concepts in interactions. By conveying input from perception, action, and language, the concept web forms rich, structured, grounded information about objects, their affordances, words, etc. We demonstrate that, given an interaction, a word, or the perceptual information from an object, the corresponding concepts in the web are activated, much the same way as they are in humans. Moreover, we show that the robot can use these activations in its concept web for several tasks to disambiguate its understanding of the scene.
ABSTRACT In this chapter the second ingredient of a program namely “actions” that programmers use... more ABSTRACT In this chapter the second ingredient of a program namely “actions” that programmers use to manipulate data for solving real-world problems are introduced. Actions are transformed into machine instructions, which are then decoded and executed by the CPU. After a brief discussion on the purpose and the scope of actions, different action types are introduced. An important part of Chap. 3 is devoted to expressions and their evaluation: regarding this, Church–Rosser property, side-effects and Dikstra’s “Shunting Yard Algorithm”, which is the most common expression evaluation method, are presented. The chapter also talks about the different operators that a programmer can use in expressions, the actions that can be used for controlling the flow of the program (i.e., conditionals) and the actions that can be reused again (i.e., functions). Like other chapters, the discussions and the presentations are supported by the case studies and examples in Python.
ABSTRACT Chapter 2, on top of the basic knowledge provided in Chap. 1, gets the reader familiar w... more ABSTRACT Chapter 2, on top of the basic knowledge provided in Chap. 1, gets the reader familiar with how “information” is represented by a computer in the form of “data”. The chapter starts by introducing why data is essential and how it relates to the representation of a real-world problem on a computer. Afterwards data is classified into two types based on their representational complexity and internals: “basic data” and “structured data”. After presenting the basic data types in details, the string, list, and tuple containers are introduced as structured data. While doing so, the chapter discusses how the different basic data types are represented internally by the computer and how the structured data is organized in the memory. As case studies, the corresponding data types in Python are introduced. Lastly, the chapter covers the concept of “variables”, i.e., names that can be used to store and access data in memory.
ABSTRACT This chapter pertains to how a programmer can deal with problems that involve actions on... more ABSTRACT This chapter pertains to how a programmer can deal with problems that involve actions on ‘bulky’ data. For such problems, a programmer has two choices: “recursion” or “iteration”. The chapter presents a meta-algorithm, a novel set of rules for the construction of a recursive solution to a world problem, and illustrates it with clear examples. As for iteration, the chapter covers all the basic types of iterative constructs (e.g., Premod-Posttest, etc.) that are widely used in the world of programming. Tips for the construction of an efficient iterative solution are also given. Finally, the advantages and disadvantages of recursion and iteration are discussed in depth. While covering recursion and iteration, examples and case studies are provided in Python.
Abstract—We describe an embodied cognitive system based on a three-level architecture that includ... more Abstract—We describe an embodied cognitive system based on a three-level architecture that includes a sensorimotor layer, a mid-level layer that stores and reasons about object-action episodes, and a high-level symbolic planner that creates abstract action plans to be realised and possibly further specified by the lower levels. The system works in two modes, exploration and plan execution, that both make use of the same architecture. We give results of different sub-processes as well as their interaction. In particular, we describe the ...
... Author(s) Morten Kjærgaard, Alex Bierbaum, Dirk Kraft, Sinan Kalkan, Norbert Krüger, Tamim As... more ... Author(s) Morten Kjærgaard, Alex Bierbaum, Dirk Kraft, Sinan Kalkan, Norbert Krüger, Tamim Asfour, Rüdiger Dillmann ... The sensor technology used for force sensors relies either on a variant of the FSR (Force Sensing Resistor) principle [3][4][1] or on the capacitive effect [2 ...
ECCV 2010 Workshop on “Vision for Cognitive Tasks, 2010
Abstract. We discuss the need of an elaborated in-between stage bridging early vision and cogniti... more Abstract. We discuss the need of an elaborated in-between stage bridging early vision and cognitive vision which we call 'Early Cognitive Vision'(ECV). This stage provides semantically rich, disambiguated and largely task independent scene representations which can be used in many contexts. In addition, the ECV stage is important for generalization processes across objects and actions. We exemplify this at a concrete realisation of an ECV system that has already been used in variety of application domains.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000
Computational modeling of the primate visual system yields insights of potential relevance to som... more Computational modeling of the primate visual system yields insights of potential relevance to some of the challenges that computer vision is facing, such as object recognition and categorization, motion detection and activity recognition, or vision-based navigation and manipulation. This paper reviews some functional principles and structures that are generally thought to underlie the primate visual cortex, and attempts to extract biological principles that could further advance computer vision research. Organized for a computer vision audience, we present functional principles of the processing hierarchies present in the primate visual system considering recent discoveries in neurophysiology. The hierarchical processing in the primate visual system is characterized by a sequence of different levels of processing (on the order of 10) that constitute a deep hierarchy in contrast to the flat vision architectures predominantly used in today's mainstream computer vision. We hope that the functional description of the deep hierarchies realized in the primate visual system provides valuable insights for the design of computer vision algorithms, fostering increasingly productive interaction between biological and computer vision research.
IEEE Transactions on Autonomous Mental Development, 2015
ABSTRACT It is now widely accepted that concepts and conceptualization are key elements towards a... more ABSTRACT It is now widely accepted that concepts and conceptualization are key elements towards achieving cognition on a humanoid robot. An important problem on this path is the grounded representation of individual concepts and the relationships between them. In this article, we propose a probabilistic method based on Markov Random Fields to model a concept web on a humanoid robot where individual concepts and the relations between them are captured. In this web, each individual concept is represented using a prototype-based conceptualization method that we proposed in our earlier work. Relations between concepts are linked to the cooccurrences of concepts in interactions. By conveying input from perception, action, and language, the concept web forms rich, structured, grounded information about objects, their affordances, words, etc. We demonstrate that, given an interaction, a word, or the perceptual information from an object, the corresponding concepts in the web are activated, much the same way as they are in humans. Moreover, we show that the robot can use these activations in its concept web for several tasks to disambiguate its understanding of the scene.
ABSTRACT In this chapter the second ingredient of a program namely “actions” that programmers use... more ABSTRACT In this chapter the second ingredient of a program namely “actions” that programmers use to manipulate data for solving real-world problems are introduced. Actions are transformed into machine instructions, which are then decoded and executed by the CPU. After a brief discussion on the purpose and the scope of actions, different action types are introduced. An important part of Chap. 3 is devoted to expressions and their evaluation: regarding this, Church–Rosser property, side-effects and Dikstra’s “Shunting Yard Algorithm”, which is the most common expression evaluation method, are presented. The chapter also talks about the different operators that a programmer can use in expressions, the actions that can be used for controlling the flow of the program (i.e., conditionals) and the actions that can be reused again (i.e., functions). Like other chapters, the discussions and the presentations are supported by the case studies and examples in Python.
ABSTRACT Chapter 2, on top of the basic knowledge provided in Chap. 1, gets the reader familiar w... more ABSTRACT Chapter 2, on top of the basic knowledge provided in Chap. 1, gets the reader familiar with how “information” is represented by a computer in the form of “data”. The chapter starts by introducing why data is essential and how it relates to the representation of a real-world problem on a computer. Afterwards data is classified into two types based on their representational complexity and internals: “basic data” and “structured data”. After presenting the basic data types in details, the string, list, and tuple containers are introduced as structured data. While doing so, the chapter discusses how the different basic data types are represented internally by the computer and how the structured data is organized in the memory. As case studies, the corresponding data types in Python are introduced. Lastly, the chapter covers the concept of “variables”, i.e., names that can be used to store and access data in memory.
ABSTRACT This chapter pertains to how a programmer can deal with problems that involve actions on... more ABSTRACT This chapter pertains to how a programmer can deal with problems that involve actions on ‘bulky’ data. For such problems, a programmer has two choices: “recursion” or “iteration”. The chapter presents a meta-algorithm, a novel set of rules for the construction of a recursive solution to a world problem, and illustrates it with clear examples. As for iteration, the chapter covers all the basic types of iterative constructs (e.g., Premod-Posttest, etc.) that are widely used in the world of programming. Tips for the construction of an efficient iterative solution are also given. Finally, the advantages and disadvantages of recursion and iteration are discussed in depth. While covering recursion and iteration, examples and case studies are provided in Python.
Uploads
Papers