Medium-sized computers: bringing computers into the lab.
Explore the source record for details and available documents.
SEARCH · PubMed Health
Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Explore the source record for details and available documents.
This paper describes an effort at UC Berkeley and the International Computer Science Institute to develop a supercomputer for artificial neural network applications. Our perspective has been strongly influenced by earlier experiences with the construction and use of a simpler machine. In particular, we have observed Amdahl's Law in action in our designs and those of others. These observations inspire attention to many factors beyond fast multiply-accumulate arithmetic. We describe a number of these factors along with rough expressions for their influence and then give the applications targets, machine goals and the system architecture for the machine we are currently designing.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
By implementing analog computer techniques using digital circuits, an instantaneous rate meter was built with approximately 100 times the accuracy of commercially available rate meters. The circuit is accurate to within +/- 0.2 events/min over a range of rates of 0.2-900 epm. Modifications can be made to provide a digital display of rate. The circuit design techniques used in developing the rate meter may be used to generate a wide variety of functions of time with very high accuracy.
An interrupter technique for the measurement of airway resistance in spontaneously breathing, unanesthetized small experimental animals is described. The method makes use of an automatic 35-mm camera body to provide the necessary speed for interruption in these rapidly breathing animals and an analog computer to monitor breathing. The system was designed as a screening technique for assessing the irritancy of air pollutants. Initial results using rabbits is presented.
It has been shown that if a recurrent neural network (RNN) learns to process a regular language, one can extract a finite-state machine (FSM) by treating regions of phase-space as FSM states. However, it has also been shown that one can construct an RNN to implement Turing machines by using RNN dynamics as counters. But how does a network learn languages that require counting? Rodriguez, Wiles, and Elman (1999) showed that a simple recurrent network (SRN) can learn to process a simple context-free language (CFL) by counting up and down. This article extends that to show a range of language tasks in which an SRN develops solutions that not only count but also copy and store counting information. In one case, the network stores information like an explicit storage mechanism. In other cases, the network stores information more indirectly in trajectories that are sensitive to slight displacements that depend on context. In this sense, an SRN can learn analog computation as a set of interdependent counters. This demonstrates how SRNs may be an alternative psychological model of language or sequence processing.
Currently there is considerable debate as to the nature of the pathways that are responsible for the perception and motor performance. We have studied the relationship between perceived speed, which is the experiential representation of a moving stimulus, and the speed of smooth pursuit eye movements, the motor action. We determined psychophysical thresholds for detecting small perturbations in the speed of moving patterns, and then by an ideal observer analysis computed analogous "oculometric" thresholds from the eye movement traces elicited by the same stimuli on the same trials. Our results confirm those of previous studies that show a remarkable agreement between perceptual judgments for speed discrimination and the fine gradations in eye movement speed. We analyzed the initial pursuit period of long duration (1000 ms) and short (200 ms) duration perturbations. When we compared the errors for perception and pursuit on a trial-by-trial basis there was no correlation between perceptual errors and eye movement errors. The observation that both oculometric and psychometric performance were similar, with Weber fractions in the normal range, but that there is no correlation in the errors suggests that the motor system and perception share the same constraints in their analysis of motion signals, but act independently and have different noise sources. We simulated noise in two models of perceptual and eye movement performance. In the first model we postulate an initial common source for the perceptual and eye movement signals. In that case about ten times the observed noise is required to produce no correlation in trial-by-trial performance. In the second model we postulate that the perceptual signal is a combination of a reafferent eye velocity signal plus the perturbation signal while the pursuit signal is derived from the oculomotor plant plus the perturbation signal. In this model about three times the noise level in the independent signals will mask any correlation due to the common perturbation signal.
The influence of fever on the pharmacokinetics of cefazolin was investigated in patients with acute febrile diseases. Nine patients were included in the study. Antibiotic serum concentrations were determined using high performance liquid chromatograpy (HPLC). An analog computer and the SIMULINK software package were used to identify the pharmacokinetic model and PCNONLIN software package to obtain the secondary parameters. In 6 patients a two-compartment pharmacokinetic model of cefazolin was observed during fever and after defervescence. In 2 patients a two-compartment model changed to a one-compartment after defervescence, and a one-compartment model was observed in one patient during both periods. Cefazolin-treated patients with a two-compartment model (6/9) had higher Cmax, mean steady state serum concentrations (Css), and area under the plasma concentration-time curve (AUC(0-->infinity)), smaller central compartment volume (V1), and lower clearance (Cl) during fever. The varying distribution of antibiotics during fever probably reflects different hemodynamic responses to fever.
Pharmacokinetic analysis on glucose tolerance curves in genetically diabetic KK-CAy mice treated with hypoglycemics was investigated using an analog computer. The glucose tolerance curves in KK-CAy mice showed humping phenomena. They could not be simulated by a simple compartmental model, but might be fitted to the model containing the non-linear function. "Compartment-H" model was contrived by adding the H-factor between glucose and insulin compartments. All the glucose tolerance curves in the KK-CAy mice could be simulated reasonable by this model and some suggestions on the humping effect were obtained from the analysis of H-factor. The H-factor was observed to be influenced by adrenergic alpha-blocker. This pharmacokinetic approach was indicated to be useful for the pharmacological research.
We describe an optoelectronic incoherent multichannel processor that is able to segment an object in a real image. The process is based on an active contour algorithm that has been transposed to optical signal processing to accelerate image processing. This implementation requires exact-valued correlations and thus opens attractive perspectives in terms of optical analog computation. Furthermore, this optical multichannel processor setup encourages incoherent processing with high-resolution images.
We develop and experimentally test a method for three-dimensional imaging of hidden objects in a scattering medium. In our scheme, objects hidden between two biological tissues at different depths from the viewing system are recovered, and their three-dimensional locations are computed. Analogous to a fly's two eyes, two microlens arrays are used to observe the hidden objects from different perspectives. At the output of each lens array we construct the objects from several sets of many speckled images with a previously suggested technique that uses a reference point. The differences of the reconstructed images in both arrays with respect to the reference point yield the information regarding the relative depth among the various objects.
A pharmacokinetic study of an intentional pentachlorophenol ingestion by an elderly human has been undertaken. Information on the presence or absence of forced diuresis either continuously or for a short restricted period indicates that such treatment would materially reduce the body burden of pentachlorophenol. It is suggested that forced diuresis is the treatment of preference for such intoxication at this time.
14C-labelled coproporphyrin I and III and uroporphyrin I were injected intravenously into healthy human subjects. Two experiments were performed with each porphyrin. Counting of blood, urine and faeces was made as a function of time. The coproporphyrin data were simulated by means of an analogy computer by using a four-compartment model with a delay pool. The physiological counterparts to the compartments were considered to be liver, blood and gut. From the blood radioactivity disappearance curve of uroporphyrin three exponential components could be worked out. The least square method was used for the solution of the plasma and urine radioactivity curves. A three-compartment model was constructed to illustrate the kinetics of uroporphyrin I. Coproporphyrin I and III behaved essentially similarly, whilst the kinetic behaviour of uroporphyrin I was completely different from that of coproporphyrins. A rapid hepatic uptake, preferential faecal excretion and a significant enterohepatic circulation were features of coproporphyrin kinetics. Over 70 per cent of the tracer activity of uroporphyrin I was found to accumulate to a rather large pool (volume 16--28 times the plasma volume) from where the back flow to plasma was rather slow. Practically all the uroporphyrin was excreted via urine.
In a series of publications, the author used the correlation hypothesis to explain the main mechanisms of functioning of bat and dolphin echo locators. A good fit of calculated parameters to experimental data was shown. In this work, by the example of generalization of the recognition problem, the use by the brain of correlation methods for solving more general problems independently on the modality of sensor signals was analyzed. In favor of the hypothesis that correlation methods are widely used in brain functioning are the data presented in the paper, which prove that an associative neuron is a suitable analog computer. It is suited for rapid "computation" of the intercorrelation function of discrete input and reference signals. The set of weighting coefficients of neuron synaptic inputs serve as such signals. The pool of associative neurons determines the values on the correlation function in the required range by changing the "numbers" of inputs from neuron to neuron at which discrete signals arrive.
Abnormalities of cardiac rhythm are due either to disorders of impulse propagation (delay or block of conduction, re-entry, circus movement, etc.), or to disturbances of impulse formation (dysfunction of ordinary pacemakers and induction of focal activity). In most instances certain disturbances of transmembrane movements of Na or K ions may be considered to be involved basically in the genesis of dysrhythmic cardiac activity. By means of intracellular recordings as well as of voltage clamp measurements and of analogical computations it is demonstrated that characteristic changes of the membrane potential might result from entirely different ionic mechanisms. Furthermore, an attempt is made to present a detailed analysis of permeability changes underlying normal pacemaker activity in Purkinje fibers as compared with the sinus node or with focal activity induced in myocardial fibers by different influences (aconitine, barium ions, stretch, and strong currents).
To evaluate human workload during manual operation, 9 subjects participated in the attitude maintaining experiment with simulated earth scenery as the display and a stick as the controller. Mean-rectified EMG was chosen as the index. Surface EMG was recorded in a couple of agonist and antagonist. Kinematics equation was simulated on a analog computer, and three difficulty of work were controlled by changing the amplifier's coefficients. 486 Computer was used to record the quantity of subject's operation and the data of Mean EMG from amplifier. Relevant software was compiled. The results showed that there was a linear relationship between mean-rectified EMG and difficulty of work. The effects of training on workload were also discussed. Mean EMG can be used as an index for evaluating human workload.
Explore the source record for details and available documents.