Artificial neural networks for coordination and control: The portability of experiential representations

It is time to locate connectionist representation theory in the new wave of robotics research. The utility of representations developed in artificial neural networks (ANNs) during learning has been demonstrated in cognitive science research since the 1980s. The research reported here puts learned representations to work in a decentered control task, the disembodied arm problem, in which a mobile robot operates an arm fixed to a table to pick up objects. There is no physical linkage between the arm and the robot and so the robot's point of view must be decentered. This is done by developing a modular Artificial Neural Net system in three stages: (i) a classifier net is trained with laser scan data to output transformationally invariant position classes; (ii) an arm net is trained for picking up objects; (iii) an inter net is trained to communicate and coordinate the sensing and acting. The completed system is shown to create new nonsymbolic transformationally invariant representations in order to perform the effective generalization of decentered viewpoints.     

Response of peripheral nerve to cyclic compression in a laboratory rat model

Repetitive cyclic loading of a nerve has been proposed as a pathogenic factor in the development of occupational compression neuropathies. Little is known about the basic response of peripheral nerve to cyclic compression. We investigated the hypothesis that cyclic compression is more detrimental to nerve function than constant compression. We measured the amplitudes and velocities of distally evoked action potentials in the presence of constant or cyclic compression of the tibial nerve in rats. Seven groups were subjected to constant or cyclic compression for 6 h by a computer controlled, hydraulically activated compression chamber. Nerves were compressed with 0 (control group), 30, 60, or 90 mm Hg of constant pressure or 0-30, 20-50, or 30-60 mm Hg of cyclic compression for approximately 20,000 compression cycles. Action potentials were recorded every 15 min. The effects of cyclic compression on nerve conduction were equivalent to the effects of constant compression at the average applied pressure. Cyclic loading itself does not appear to be an important pathogenic factor in the development of nerve conduction block.     

Religiosity and emotional disturbance: A test of Ellis's thesis in his own counseling center.

Assessed the validity of A. Ellis's (1971, 1973, 1981) contention that being religious was negatively correlated with mental health. 28 very religious, 34 atheist, and 33 religiously-neutral individuals seeking consultation were compared on the frequency and number of 8 reported problems. Contrary to Ellis's prediction, there were no significant differences among the groups. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Effect of intracolonic benzalkonium chloride on trinitrobenzene sulphonic acid-induced colitis in the rat

BACKGROUND: We investigated the effects of benzalkonium chloride (BAC) on trinitrobenzene sulphonic acid (TNBS)-induced colitis in rats. METHODS: TNBS was administered intrarectally before and/or after BAC treatment. In the first study, the effects of treatment with BAC 6, 12 or 24 h after TNBS were examined. In the second study, animals were treated with BAC before, after or before and after TNBS, and were examined 7 days later. The severity of colitis was assessed by macroscopic and histological scoring of the colonic damage and by determination of colonic myeloperoxidase (MPO) activity. Macrophages and CD4+ and CD8+ T cells were examined by immunohistochemistry. RESULTS: When BAC was instilled into the colon 6, 12 or 24 h after TNBS, weight loss and macroscopic and histological features of the colon were similar to that of controls (TNBS alone). In contrast, MPO activity was significantly reduced in all three groups post-treated with BAC. In the groups examined 7 days after TNBS treatment, rats post-treated with BAC exhibited increased weight gain and significantly reduced macroscopic damage and MPO activity compared to the TNBS control group. Rats pre-treated with BAC exhibited less macroscopic damage of the colon than rats receiving only TNBS, but histological damage, MPO and weight gain were unchanged from TNBS controls. Immunohistochemistry revealed that BAC pre-treatment increased the numbers of macrophages and T cells in the colon. After TNBS treatment, macrophage accumulation was evident in the colon, but T cells were scarce. However, these cells were preserved or enhanced in the colonic mucosa in TNBS-treated rats that had been pre-treated with BAC. CONCLUSIONS: Treatment with BAC, particularly after induction of colitis, produces a significant reduction in the severity of tissue injury and inflammation through mechanisms that are not fully understood.     

Robust nonlinear model identification methods using forward regression

In this correspondence new robust nonlinear model construction algorithms for a large class of linear-in-the-parameters models are introduced to enhance model robustness via combined parameter regularization and new robust structural selective criteria. In parallel to parameter regularization, we use two classes of robust model selection criteria based on either experimental design criteria that optimizes model adequacy, or the predicted residual sums of squares (PRESS) statistic that optimizes model generalization capability, respectively. Three robust identification algorithms are introduced, i.e., combined A- and D-optimality with regularized orthogonal least squares algorithm, respectively; and combined PRESS statistic with regularized orthogonal least squares algorithm. A common characteristic of these algorithms is that the inherent computation efficiency associated with the orthogonalization scheme in orthogonal least squares or regularized orthogonal least squares has been extended such that the new algorithms are computationally efficient. Numerical examples are included to demonstrate effectiveness of the algorithms.     

Learning from Innate Behaviors: A Quantitative Evaluation of Neural Network Controllers

The aim was to investigate a method of developing mobile robot controllers based on ideas about how plastic neural systems adapt to their environment by extracting regularities from the amalgamated behavior of inflexible (nonplastic) innate subsystems interacting with the world. Incremental bootstrapping of neural network controllers was examined. The objective was twofold. First, to develop and evaluate the use of prewired or innate robot controllers to bootstrap backpropagation learning for Multilayer Perceptron (MLP) controllers. Second, to develop and evaluate a new MLP controller trained on the back of another bootstrapped controller. The experimental hypothesis was that MLPs would improve on the performance of controllers used to train them. The performances of the innate and bootstrapped MLP controllers were compared in eight experiments on the tasks of avoiding obstacles and finding goals. Four quantitative measures were employed: the number of sensorimotor loops required to complete a task; the distance traveled; the mean distance from walls and obstacles; the smoothness of travel. The overall pattern of results from statistical analyses of these quantities supported the hypothesis; the MLP controllers completed the tasks faster, smoother, and steered further from obstacles and walls than their innate teachers. In particular, a single MLP controller incrementally bootstrapped by a MLP subsumption controller was superior to the others.     

It’s not all written on the robot’s face

Past work on creating robots that can make convincing emotional expressions has concentrated on the quality of those expressions, and on assessing people’s ability to recognize them in neutral contexts, without any strong emotional valence. It would be interesting to find out whether observers’ judgments of the facial cues of a robot would be affected by a surrounding emotional context. This paper takes its inspiration from the contextual effects found on our interpretation of the expressions on human faces and computer avatars, and looks at the extent to which they also apply to the interpretation of the facial expressions of a mechanical robot head. The kinds of contexts that affect the recognition of robot emotional expressions, the circumstances under which such contextual effects occur, and the relationship between emotions and the surrounding situation, are observed and analyzed. Design implications for believable emotional robots are drawn.     

Artificial intelligence and natural magic

Robotics with AI is part of a long tradition that has run from ancient times that treated the precursors of robots, the automata, as part of Natural Magic or conjury. Deception is an integral part of AI and robotics; in some ways they form a science of illusion. There are many robot tasks, such as caring for the elderly, minding children, doing domestic chores and being companionable, that involve working closely with humans and so require some illusion of animacy and thought. We discuss how the natural magic of robotics is assisted by the cultural myth of AI together with innate human predispositions such as zoomorphism, the willing suspension of disbelief and a tendency to interpret AI devices as part of the social world. This approach provides a justifiable way of meeting the goals of AI and robotics provided that researchers do not allow themselves to be deceived by their own illusions.     

Reduction of renal uptake of monoclonal antibody fragments by amino acid infusion

The renal uptake of radiolabeled antibody fragments and peptides presents a problem in radioimmunodetection and therapy, compromising lesion sensitivity, especially with intracellularly-retained isotopes. Previously, we showed that cationic amino acids and their derivatives are capable of significantly reducing kidney uptake in animals. We report our initial clinical results of successful renal uptake reduction in five patients who underwent cancer radioimmunodetection with 99mTc-anti-CEA Fab' fragments. METHODS: The patients were infused with two liters of a commercially-available nutritive amino acid solution (containing approximately 2.25 g/liter lysine-glutamate and 2.50 g/liter arginine), whereas 75 control patients received the same volume of saline (quantification of organ and tumor kinetics from conjugate whole-body views by ROI technique). RESULTS: The renal uptake in the amino acid group was significantly lower (p<0.05) than in the control group (11.1 +/- 2.0% injected dose versus 17.7 +/- 7.0% injected dose at 24 hr postinjection), whereas the uptake of all other organs remained unaffected. Gel filtration chromatography of the urine taken from amino-acid-treated patients showed that a significantly higher amount of excreted activity was bound to intact Fab' (53% of excreted activity) in contrast to only less than 10% in the control group. CONCLUSION: The renal uptake of monoclonal antibody fragments in patients can be reduced significantly by amino acid infusion, even at considerably lower doses than those that were safe and effective in animals. As was found in animals, the mechanism seems to rely on an inhibition of the re-absorption of tubularly-filtered proteins by the proximal tubule cells. These results encourage further clinical trials to lower the renal uptake experienced in radioimmunodetection, as well as in therapeutic trials with antibody fragments and peptides.     

A Multi-Net System for the Fault Diagnosis of a Diesel Engine

A multi-net fault diagnosis system designed to provide an early warning of combustion-related faults in a diesel engine is presented. Two faults (a leaking exhaust valve and a leaking fuel injector nozzle) were physically induced (at separate times) in the engine. A pressure transducer was used to sense the in-cylinder pressure changes during engine cycles under both of these conditions, and during normal operation. Data corresponding to these measurements were used to train artificial neural nets to recognise the faults, and to discriminate between them and normal operation. Individually trained nets, some of which were trained on subtasks, were combined to form a multi-net system. The multi-net system is shown to be effective when compared with the performance of the component nets from which it was assembled. The system is also shown to outperform a decision-tree algorithm (C5.0), and a human expert; comparisons which show the complexity of the required discrimination. The results illustrate the improvements in performance that can come about from the effective use of both problem decomposition and redundancy in the construction of multi-net systems.