A simulation-based tutorial system for exploring parallel distributed processing

This article presents a simulation-based tutorial system for exploring parallel distributed processing (PDP) models of information processing. The system consists of software and an accompanying handbook. The intent of the package is to make the ideas underlying PDP accessible and to disseminate some of the main simulation programs that we have developed. This article presents excerpts from the handbook that describe the approach taken, the organization of the handbook, and the software that comes with it. An example is given that illustrates the approach we have taken to teaching PDP, which involves presentation of relevant mathematical background, together with tutorial exercises that make use of the simulation programs.     

Distributed memory and the representation of general and specific information

We describe a distributed model of information processing and memory and apply it to the representation of general and specific information. The model consists of a large number of simple processing elements which send excitatory and inhibitory signals to each other via modifiable connections. Information processing is thought of as the process whereby patterns of activation are formed over the units in the model through their excitatory and inhibitory interactions. The memory trace of a processing event is the change or increment to the strengths of the interconnections that results from the processing event. The traces of separate events are superimposed on each other in the values of the connection strengths that result from the entire set of traces stored in the memory. The model is applied to a number of findings related to the question of whether we store abstract representations or an enumeration of specific experiences in memory. The model simulates the results of a number of important experiments which have been taken as evidence for the enumeration of specific experiences. At the same time, it shows how the functional equivalent of abstract representations--prototypes, logogens, and even rules--can emerge from the superposition of traces of specific experiences, when the conditions are right for this to happen. In essence, the model captures the structure present in a set of input patterns; thus, it behaves as though it had learned prototypes or rules, to the extent that the structure of the environment it has learned about can be captured by describing it in terms of these abstractions.     

The relationship of affiliative arousal to dopamine release

Higher levels of affiliation motivation after exposure to affiliative films are significantly related to higher dopamine concentrations in saliva in one study of 61 college students of both sexes, and in plasma in another study of 47 adults of both sexes. Individuals high in dispositionaln Affiliation, if they also reported high life stress, showed high gains in dopamine concentration in plasma after exposure to an affiliative film significantly more often than other individuals. Since aroused affiliation motivation was not associated with higher concentrations of norepinephrine, epinephrine, or cortisol in either study, it appears likely that dopamine is especially associated with arousal of affiliation motivation, just as norepinephrine has been found to be associated with arousal of power motivation. Thus, different motives may be subserved by different hormones, making it unlikely that all motives lead to the same state of physiological arousal.This research was carried out when all the authors were connected with the Department of Psychology and Social Relations at Harvard University. We acknowledge with thanks the support of the John D. and Catherine T. MacArthur Foundation through a grant to the senior author.     

Perceptual interactions in two-word displays: familiarity and similarity effects

Previous studies have demonstrated the existence of perceptual interactions in the processing of two-word displays such as SAND LANE. When postcued to report one of the two words, subjects often make migration errors, in that the report of the specified word includes a letter of the other word (e.g., LAND or SANE instead of SAND). We find that migrations depend on the abstract, structural similarity of the strings, but not on the physical similarity; on whether the strings are words; and on whether the possible migration responses are words. We also rule out an interpretation of migration errors that attributes them to a guessing strategy. Our findings are interpreted in terms of models in which both strings simultaneously access high-level structural knowledge, that is, knowledge about what sequences of letters fit together to form familiar wholes.     

Achievement motivation in relation to achievement-related recall,performance, and urine flow,a marker associated with release of vasopressin

Administration of the pituitary hormone, arginine vasopressin (AVP), enhances memory in animals and humans. Since AVP is the antidiuretic hormone, its release can be inferred from reductions in urine flow. Forty-six male adults recalled the contents of a complex story and then either were exposed to achievement arousal or were not. Significantly more of the subjects in the aroused than the nonaroused condition showed a decrease in urine flow, and the greater the decrease in urine flow among the achievement-aroused subjects, the better the recall of achievement-related story content 24 hours later. Subjects high inn Achievement, in the achievement-oriented condition, showed lower urine output at the outset of the second day and greater improvement in learning to unscramble words. Since lower urine output was also related to improvement in unscrambling words, achievement arousal may have released AVP which in turn may have facilitated achievement-related retention and performance.I am greatly indebted to Sean Connelly for assistance in carrying out this research and to Dr. Franklin Epstein of the Beth Israel Hospital, Boston, for helping me to understand how the kidney functions in producing urine.     

Promoting Energy Conservation in Master-Metered Apartments through Group Financial Incentives1

An energy conservation contest among four groups of apartments (44 to 70 units each) in a University of Colorado family housing complex was conducted February-April, 1977. The winning group in each of six 2-week contests won $80 to use however it determined. The program produced a drop of about 10% in use of natural gas during the first contest; energy savings were smaller but still statistically significant throughout the first 8 weeks. Over a 12-week period, savings averaged 6.6%. Some changes in energy-using behaviors were reported by residents surveyed at the contest's end; however, knowledge of contest results by residents was minimal. The results are compared to those of other contest and rebate programs, the practical problems of costs exceeding savings and inaccurate estimates of energy savings are discussed, and implications for future research are suggested.     

Learning Continuous Probability Distributions with Symmetric Diffusion Networks

In this article we present symmetric diffusion networks, a family of networks that instantiate the principles of continuous, stochastic, adaptive and interactive propagation of information. Using methods of Markovion diffusion theory, we formalize the activation dynamics of these networks and then show that they can be trained to reproduce entire multivariate probability distributions on their outputs using the contrastive Hebbion learning rule (CHL). We show that CHL performs gradient descent on an error function that captures differences between desired and obtained continuous multivariate probability distributions. This allows the learning algorithm to go beyond expected values of output units and to approximate complete probability distributions on continuous multivariate activation spaces. We argue that learning continuous distributions is an important task underlying a variety of real-life situations that were beyond the scope of previous connectionist networks. Deterministic networks, like back propagation, cannot learn this task because they are limited to learning average values of independent output units. Previous stochastic connectionist networks could learn probability distributions but they were limited to discrete variables. Simulations show that symmetric diffusion networks can be trained with the CHL rule to approximate discrete and continuous probability distributions of various types.     

The immunoenhancing effects of humor on secretory IgA and resistance to respiratory infections

Abstract

In three different studies involving exposun to two differcot humorous and comparison videotapes, it was demonstrated that humor arousal signiticantly increases S-IgA in college students. Furthermore significantly more of the subjects exposed to the humor films showed an increese in S-IgA than the subjects exposed to the comparison films. A good sense of humor, measured as a trait, was associated with higher baseline concentrations of S-IgA and with greater increases in S-IgA in response to a humor arousing film. Higher bascline kvels and gains in S-IgA in respoase to humor were associated with lower reported severity of colds in the past twelve months and in the three months following assessment. However, a good sense of humor was related to lesser cold incidence and severity only if it involved appreciation rather than production of humor. While those who actively tell jokes also show higher S-IgA levels, they show greater susceptibility to colds than those who only appreciate humor probably because they sodalize more often under conditions that expose them more to cold infections.     

Some reflections on the two psychologies of love

The mainstream view of love is that it is a state that arises from people mutually reinforcing each other or providing benefits to each other Yet careful quantitative analyses of imaginative thought patterns of people in love or in a state of affiliative arousal reveal themes of dialogue, commitment, and harmony rather than of mutual benefits It is suggested that asking people about love leads them to give causal explanations for the state which psychologists have elaborated into the theory that love is a response to having one's needs met Imaginative experiences characteristic of the state of being in love on the other hand may be different because they are primarily right brain mediated and do not share in the causal, instrumental type of reasoning characteristic of left brain conscious thinking Such conceivably right brain mediated experiences in TATs are more closely associated than presumably left brain mediated self-reports are with such physiological measures as evoked scalp potentials, neurohormone concentrations, and immune functions A vew of love that takes both self-report (left brain) and fantasy (right brain) measures into account does a better job of predicting affiliative behavior than does a view that relies solely on one or the other of the two types of measures     

Mental search processes in problem solving

A new approach to problem solving was applied to multisolution problems in a memory search task. Subjects memorized a list of eight four-letter foods, and then searched mentally through the list for answers to questions. The times between successive answers (IRTs) were recorded along with the answers themselves. This allowed a comparison of two possible memory search strategies: (1) sampling with replacement, and (2) sampling without replacement. The results were largely in agreement with the sampling-without-replacement strategy. However, a more detailed breakdown of the data revealed that most subjects searched through the list in a rigid serial order. Further, an analysis of questions with identical answers showed that the IRTs were very nearly additive. This led to an additive time component model based on the independent summation of (a) read-in time, (b) memory-search time, (c) decision-making time, and (d)response-output time. This approach appeared generally more satisfactory than previous attempts to account for problem-solving behavior.