Companion Cognitive Systems: Design Goals and Lessons Learned So Far

The Companion cognitive architecture supports experiments in achieving human-level intelligence. This article describes seven key design goals of Companions, relating them to properties of human reasoning and learning, and to engineering concerns raised by attempting to build large-scale cognitive systems. We summarize our experiences with Companions in two domains: test taking and game playing.     

Mediating Power: Distance Learning Interfaces, Classroom Epistemology, and the Gaze

Distance learning's interfaces—from corresponding through the postal service to the televised talking head—have traditionally been designed from the top down, supporting banking models of learning or, in writing instruction, current-traditional rhetoric pedagogies. Due to temporal and spatial constraints, these interface designs often support (or encourage) one-way communication from the instructor to the student. Students mostly interact with the instructor by asking questions or submitting work, and they tend to have little correspondence with other peers. These methods clearly privilege the instructor's knowledge and evaluation. Furthermore, these interface designs empower the instructor to gaze upon the students and assess them—often not as a corporeal body but as a corpus of texts. Thus, each interface adopted for distance learning sets up a power dynamic in which the capability to share the roles of creating knowledge is juxtaposed with the instructor's capability to normalize the students and reify their own authority through their gaze. In this article we examine the traditional classroom interface through the correspondence course interface, the simulated classroom interface, and the synchronous video interface to raise questions about the infrastructures of distance learning and their implications for student learning.     

On the estimation of adequate test set size using fault failure rates

Test set size in terms of the number of test cases is an important consideration when testing software systems. Using too few test cases might result in poor fault detection and using too many might be very expensive and suffer from redundancy. We define the failure rate of a program as the fraction of test cases in an available test pool that result in execution failure on that program. This paper investigates the relationship between failure rates and the number of test cases required to detect the faults. Our experiments based on 11 sets of C programs suggest that an accurate estimation of failure rates of potential fault(s) in a program can provide a reliable estimate of adequate test set size with respect to fault detection and should therefore be one of the factors kept in mind during test set construction. Furthermore, the model proposed herein is fairly robust to incorrect estimations in failure rates and can still provide good predictive quality. Experiments are also performed to observe the relationship between multiple faults present in the same program using the concept of a failure rate. When predicting the effectiveness against a program with multiple faults, results indicate that not knowing the number of faults in the program is not a significant concern, as the predictive quality is typically not affected adversely.     

Detection of environmental changes from network structure

We studied the networks of the temperature record in the atmosphere. They are made by the strength of the synchronization, including the delay between temperatures at locations on Earth. We consider these locations as nodes, and we consider a pair of locations as a link if the synchronization between them is stronger than a threshold. The network is scale-free, which is thought to contribute to the stability of the climate. This work was presented in part at the 12th International Symposium on Artificial Life and Robotics, Oita, Japan, January 25–27, 2007     

Survival as a Digital Ghost

You can survive after death in various kinds of artifacts. You can survive in diaries, photographs, sound recordings, and movies. But these artifacts record only superficial features of yourself. We are already close to the construction of programs that partially and approximately replicate entire human lives (by storing their memories and duplicating their personalities). A digital ghost is an artificially intelligent program that knows all about your life. It is an animated auto-biography. It replicates your patterns of belief and desire. You can survive after death in a digital ghost. We discuss a series of digital ghosts over the next 50 years. As time goes by and technology advances, they are progressively more perfect replicas of the lives of their original authors.

Compositional pattern producing networks: A novel abstraction of development

Natural DNA can encode complexity on an enormous scale. Researchers are attempting to achieve the same representational efficiency in computers by implementing developmental encodings, i.e. encodings that map the genotype to the phenotype through a process of growth from a small starting point to a mature form. A major challenge in in this effort is to find the right level of abstraction of biological development to capture its essential properties without introducing unnecessary inefficiencies. In this paper, a novel abstraction of natural development, called Compositional Pattern Producing Networks (CPPNs), is proposed. Unlike currently accepted abstractions such as iterative rewrite systems and cellular growth simulations, CPPNs map to the phenotype without local interaction, that is, each individual component of the phenotype is determined independently of every other component. Results produced with CPPNs through interactive evolution of two-dimensional images show that such an encoding can nevertheless produce structural motifs often attributed to more conventional developmental abstractions, suggesting that local interaction may not be essential to the desirable properties of natural encoding in the way that is usually assumed.

Identification of statistical patterns in complex systems via symbolic time series analysis

Identification of statistical patterns from observed time series of spatially distributed sensor data is critical for performance monitoring and decision making in human-engineered complex systems, such as electric power generation, petrochemical, and networked transportation. This paper presents an information-theoretic approach to identification of statistical patterns in such systems, where the main objective is to enhance structural integrity and operation reliability. The core concept of pattern identification is built upon the principles of Symbolic Dynamics, Automata Theory, and Information Theory. To this end, a symbolic time series analysis method has been formulated and experimentally validated on a special-purpose test apparatus that is designed for data acquisition and real-time analysis of fatigue damage in polycrystalline alloys.