Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “collective intelligence”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2
All sourcesNASA NTRSDOE OSTI
NASA NTRS2000

Optimal Reward Functions in Distributed Reinforcement Learning

We consider the design of multi-agent systems so as to optimize an overall world utility function when (1) those systems lack centralized communication and control, and (2) each agents runs a distinct Reinforcement Learning (RL) algorithm. A crucial issue in such design problems is to initialize/update each agent's private utility function, so as to induce best possible world utility. Traditional 'team game' solutions to this problem sidestep this issue and simply assign to each agent the world utility as its private utility function. In previous work we used the 'Collective Intelligence' framework to derive a better choice of private utility functions, one that results in world utility performance up to orders of magnitude superior to that ensuing from use of the team game utility. In this paper we extend these results. We derive the general class of private utility functions that both are easy for the individual agents to learn and that, if learned well, result in high world utility. We demonstrate experimentally that using these new utility functions can result in significantly improved performance over that of our previously proposed utility, over and above that previous utility's superiority to the conventional team game utility.

Wolpert, David H.↗
NASA NTRS2000

Optimal Wonderful Life Utility Functions in Multi-Agent Systems

The mathematics of Collective Intelligence (COINs) is concerned with the design of multi-agent systems so as to optimize an overall global utility function when those systems lack centralized communication and control. Typically in COINs each agent runs a distinct Reinforcement Learning (RL) algorithm, so that much of the design problem reduces to how best to initialize/update each agent's private utility function, as far as the ensuing value of the global utility is concerned. Traditional team game solutions to this problem assign to each agent the global utility as its private utility function. In previous work we used the COIN framework to derive the alternative Wonderful Life Utility (WLU), and experimentally established that having the agents use it induces global utility performance up to orders of magnitude superior to that induced by use of the team game utility. The WLU has a free parameter (the clamping parameter) which we simply set to zero in that previous work. Here we derive the optimal value of the clamping parameter, and demonstrate experimentally that using that optimal value can result in significantly improved performance over that of clamping to zero, over and above the improvement beyond traditional approaches.

Wolpert, David H.↗
NASA NTRS2001

Learning Sequences of Actions in Collectives of Autonomous Agents

In this paper we focus on the problem of designing a collective of autonomous agents that individually learn sequences of actions such that the resultant sequence of joint actions achieves a predetermined global objective. We are particularly interested in instances of this problem where centralized control is either impossible or impractical. For single agent systems in similar domains, machine learning methods (e.g., reinforcement learners) have been successfully used. However, applying such solutions directly to multi-agent systems often proves problematic, as agents may work at cross-purposes, or have difficulty in evaluating their contribution to achievement of the global objective, or both. Accordingly, the crucial design step in multiagent systems centers on determining the private objectives of each agent so that as the agents strive for those objectives, the system reaches a good global solution. In this work we consider a version of this problem involving multiple autonomous agents in a grid world. We use concepts from collective intelligence to design goals for the agents that are 'aligned' with the global goal, and are 'learnable' in that agents can readily see how their behavior affects their utility. We show that reinforcement learning agents using those goals outperform both 'natural' extensions of single agent algorithms and global reinforcement, learning solutions based on 'team games'.

Turner, Kagan↗
NASA NTRS2003

Providing Effective Access to Shared Resources: A COIN Approach

Managers of systems of shared resources typically have many separate goals. Examples are efficient utilization of the resources among its users and ensuring no user's satisfaction in the system falls below a preset minimal level. Since such goals will usually conflict with one another, either implicitly or explicitly the manager must determine the relative importance of the goals, encapsulating that into an overall utility function rating the possible behaviors of the entire system. Here we demonstrate a distributed, robust, and adaptive way to optimize that overall function. Our approach is to interpose adaptive agents between each user and the system, where each such agent is working to maximize its own private utility function. In turn, each such agent's function should be both relatively easy for the agent to learn to optimize, and 'aligned' with the overall utility function of the system manager - an overall function that is based on but in general different from the satisfaction functions of the individual users. To ensure this we enhance the COllective INtelligence (COIN) framework to incorporate user satisfaction functions in the overall utility function of the system manager and accordingly in the associated private utility functions assigned to the users agents. We present experimental evaluations of different COIN-based private utility functions and demonstrate that those COIN-based functions outperform some natural alternatives.

Airiau, Stephane↗
NASA NTRS2004

Providing Effective Access to Shared Resources: A COIN Approach

Managers of systems of shared resources typically have many separate goals. Examples are efficient utilization of the resources among its users and ensuring no user s satisfaction in the system falls below a preset minimal level. Since such goals will usually conflict with one another, either implicitly or explicitly the manager must determine the relative importance of the goals, encapsulating that into an overall utility function rating the possible behaviors of the entire system. Here we demonstrate a distributed, robust, and adaptive way to optimize that overall function. Our approach is to interpose adaptive agents between each user and the system, where each such agent is working to maximize its own private utility function. In turn, each such agent's function should be both relatively easy for the agent to learn to optimize, and "aligned" with the overall utility function of the system manager - an overall function that is based on but in general different from the satisfaction functions of the individual users. To ensure this we enhance the Collective INtelligence (COIN) framework to incorporate user satisfaction functions in the overall utility function of the system manager and accordingly in the associated private utility functions assigned to the users agents. We present experimental evaluations of different COIN-based private utility functions and demonstrate that those COIN-based functions outperform some natural alternatives.

Airiau, Stephane↗
NASA NTRS2004

Product Distributions for Distributed Optimization

With connections to bounded rational game theory, information theory and statistical mechanics, Product Distribution (PD) theory provides a new framework for performing distributed optimization. Furthermore, PD theory extends and formalizes Collective Intelligence, thus connecting distributed optimization to distributed Reinforcement Learning (FU). This paper provides an overview of PD theory and details an algorithm for performing optimization derived from it. The approach is demonstrated on two unconstrained optimization problems, one with discrete variables and one with continuous variables. To highlight the connections between PD theory and distributed FU, the results are compared with those obtained using distributed reinforcement learning inspired optimization approaches. The inter-relationship of the techniques is discussed.

Bieniawski, Stefan R.↗
NASA NTRS2001

Design and Control of Large Collections of Learning Agents

The intelligent control of multiple autonomous agents is an important yet difficult task. Previous methods used to address this problem have proved to be either too brittle, too hard to use, or not scalable to large systems. The 'Collective Intelligence' project at NASA/Ames provides an elegant, machine-learning approach to address these problems. This approach mathematically defines some essential properties that a reward system should have to promote coordinated behavior among reinforcement learners. This work has focused on creating additional key properties and algorithms within the mathematics of the Collective Intelligence framework. One of the additions will allow agents to learn more quickly, in a more coordinated manner. The other will let agents learn with less knowledge of their environment. These additions will allow the framework to be applied more easily, to a much larger domain of multi-agent problems.

Agogino, Adrian↗
NASA NTRS1990

Intelligent perturbation algorithms for space scheduling optimization

The optimization of space operations is examined in the light of optimization heuristics for computer algorithms and iterative search techniques. Specific attention is given to the search concepts known collectively as intelligent perturbation algorithms (IPAs) and their application to crew/resource allocation problems. IPAs iteratively examine successive schedules which become progressively more efficient, and the characteristics of good perturbation operators are listed. IPAs can be applied to aerospace systems to efficiently utilize crews, payloads, and resources in the context of systems such as Space-Station scheduling. A program is presented called the MFIVE Space Station Scheduling Worksheet which generates task assignments and resource usage structures. The IPAs can be used to develop flexible manifesting and scheduling for the Industrial Space Facility.

Kurtzman, Clifford R.↗
NASA NTRS2004

Space-Based Sensor Web for Earth Science Applications: An Integrated Architecture for Providing Societal Benefits

There is a significant interest in the Earth Science research and user remote sensing community to substantially increase the number of useful observations relative to the current frequency of collection. The obvious reason for such a push is to improve the temporal, spectral, and spatial coverage of the area(s) under investigation. However, there is little analysis available in terms of the benefits, costs and the optimal set of sensors needed to make the necessary observations. Classic observing system solutions may no longer be applicable because of their point design philosophy. Instead, a new intelligent data collection system paradigm employing both reactive and proactive measurement strategies with adaptability to the dynamics of the phenomena should be developed. This is a complex problem that should be carefully studied and balanced across various boundaries including: science, modeling, applications, and technology. Modeling plays a crucial role in making useful predictions about naturally occurring or human-induced phenomena In particular, modeling can serve to mitigate the potentially deleterious impacts a phenomenon may have on human life, property, and the economy. This is especially significant when one is interested in learning about the dynamics of, for example, the spread of forest fires, regional to large-scale air quality issues, the spread of the harmful invasive species, or the atmospheric transport of volcanic plumes and ash. This paper identifies and examines these challenging issues and presents architectural alternatives for an integrated sensor web to provide observing scenarios driving the requisite dynamic spatial, spectral, and temporal characteristics to address these key application areas. A special emphasis is placed on the observing systems and its operational aspects in serving the multiple users and stakeholders in providing societal benefits. We also address how such systems will take advantage of technological advancement in small spacecraft and emerging information technologies, and how sensor web options may be realized and made affordable. Specialized detector subsystems and precision flying techniques may still require substantial innovation, development time and cost: we have presented the considerations for these issues. Finally, data and information gathering and compression techniques are also briefly described.

Habib, Shahid↗
NASA NTRS1991

Design and Analysis Tools for Concurrent Blackboard Systems

A blackboard system consists of a set of knowledge sources, a blackboard data structure, and a control strategy used to activate the knowledge sources. The blackboard model of problem solving is best described by Dr. H. Penny Nii of the Stanford University AI Laboratory: "A Blackboard System can be viewed as a collection of intelligent agents who are gathered around a blackboard, looking at pieces of information written on it, thinking about the current state of the solution, and writing their conclusions on the blackboard as they generate them. " The blackboard is a centralized global data structure, often partitioned in a hierarchical manner, used to represent the problem domain. The blackboard is also used to allow inter-knowledge source communication and acts as a shared memory visible to all of the knowledge sources. A knowledge source is a highly specialized, highly independent process that takes inputs from the blackboard data structure, performs a computation, and places the results of the computation in the blackboard data structure. This design allows for an opportunistic control strategy. The opportunistic problem-solving technique allows a knowledge source to contribute towards the solution of the current problem without knowing which of the other knowledge sources will use the information. The use of opportunistic problem-solving allows the data transfers on the blackboard to determine which processes are active at a given time. Designing and developing blackboard systems is a difficult process. The designer is trying to balance several conflicting goals and achieve a high degree of concurrent knowledge source execution while maintaining both knowledge and semantic consistency on the blackboard. Blackboard systems have not attained their apparent potential because there are no established tools or methods to guide in their construction or analyze their performance.

McManus, John W.↗
NASA NTRS2016

Weathering the Storm: Unmanned Aircraft Systems in the Maritime, Atmospheric and Polar Environments

Unmanned systems have the potential to efficiently, effectively, economically, and safely bridge critical observation requirements in an environmentally friendly manner. As the United States Atmospheric, Marine and Arctic areas of interest expand and include hard-to-reach regions of the Earth (such as the Arctic and remote oceanic areas) optimizing unmanned capabilities will be needed to advance the United States science, technology and security efforts. Through increased multi-mission and multi-agency operations using improved inter-operable and autonomous unmanned systems, the research and operations communities will better collect environmental intelligence and better protect our country against hazardous weather, environmental, marine and polar hazards. This presentation will examine UAS (Unmanned Aircraft Systems) operations which includes developing a coordinated effort to maximize the efficiency and capabilities of unmanned systems across the federal government and research partners. Numerous intra- and inter-agency operational demonstrations and assessments have been made to verify and validated these UAS observations. The NASA-specific elements will consist of an overview of NASA UAVs (Unmanned Aerial Vehicles) available to science and recent missions.

Atmospheric↗
NASA NTRS2022

Principles for Architecting Autonomous Systems

This paper distills principles for developing autonomous systems based on experience and lessons learned from past efforts. The purpose of these principles is to establish a common understanding and knowledge of architectural elements to guide the development of next-generation multi-mission autonomous systems and ensure the safe and productive operation of space assets. An attempt has been made to ground these principles in fundamentals that should withstand the test of time while allowing for and enabling the advancement of technologies. They are not intended to prescribe a design nor a software representation. There may be multiple designs that can honor these principles. These principles are focused on autonomy for robotic assets. As such, they do not address autonomy for crewed assets nor autonomy that can collectively generate intelligent behavior without top-level system cognizance (e.g., intelligent swarm behavior). These areas would be a subject of future efforts.

Day, John↗
NASA NTRS1999

Autonomous Science Analyses of Digital Images for Mars Sample Return and Beyond

To adequately explore high priority landing sites, scientists require rovers with greater mobility. Therefore, future Mars missions will involve rovers capable of traversing tens of kilometers (vs. tens of meters traversed by Mars Pathfinder's Sojourner). However, the current process by which scientists interact with a rover does not scale to such distances. A single science objective is achieved through many iterations of a basic command cycle: (1) all data must be transmitted to Earth and analyzed; (2) from this data, new targets are selected and the necessary information from the appropriate instruments are requested; (3) new commands are then uplinked and executed by the spacecraft and (4) the resulting data are returned to Earth, starting the process again. Experience with rover tests on Earth shows that this time intensive process cannot be substantially shortened given the limited data downlink bandwidth and command cycle opportunities of real missions. Sending complete multicolor panoramas at several waypoints, for example, is out of the question for a single downlink opportunity. As a result, long traverses requiring many science command cycles would likely require many weeks, months or even years, perhaps exceeding rover design life or other constraints. Autonomous onboard science analyses can address these problems in two ways. First, it will allow the rover to transmit only "interesting" images, defined as those likely to have higher science content. Second, the rover will be able to anticipate future commands, for example acquiring and returning spectra of "interesting" rocks along with the images in which they were detected. Such approaches, coupled with appropriate navigational software, address both the data volume and command cycle bottlenecks that limit both rover mobility and science yield. We are developing algorithms to enable such intelligent decision making by autonomous spacecraft. Reflecting the ultimate level of ability we aim for, this program has been dubbed the "Grad Student on Mars Project". We envision, for example, an appropriately intelligent Athena-like rover at the Pathfinder landing site might be able to traverse over the ridge towards "Twin Peaks" to obtain better information on the stratigraphy of these "streamlined islands" or of the size, composition and morphology of boulders located on them. Along the traverse, the intelligent rover would collect and analyze images and obtain spectra of geologically interesting features or regions. The intelligent rover might also traverse further up Arcs Vallis, and find additional paleoflood stage indicators such as slackwater deposits. Recognizing additional regions where boulders are imbricated, noting changes in their size, distribution, morphology, composition and the associated changes in channel geometry would yield important information on the outflow channel's paleoflood history, Representative images and associated supporting data from these locations could be downlinked to Earth along with the data requested by scientists from the previous uplink opportunity. Our initial work has focused on recognizing geologically interesting portions of images. Here we summarize some of the algorithms to date.

Gulick, V. C.↗
NASA NTRS1993

Intelligent Agent Architectures: Reactive Planning Testbed

An Integrated Agent Architecture (IAA) is a framework or paradigm for constructing intelligent agents. Intelligent agents are collections of sensors, computers, and effectors that interact with their environments in real time in goal-directed ways. Because of the complexity involved in designing intelligent agents, it has been found useful to approach the construction of agents with some organizing principle, theory, or paradigm that gives shape to the agent's components and structures their relationships. Given the wide variety of approaches being taken in the field, the question naturally arises: Is there a way to compare and evaluate these approaches? The purpose of the present work is to develop common benchmark tasks and evaluation metrics to which intelligent agents, including complex robotic agents, constructed using various architectural approaches can be subjected.

Rosenschein, Stanley J.↗
NASA NTRS1987

Mixed-initiative control of intelligent systems

Mixed-initiative user interfaces provide a means by which a human operator and an intelligent system may collectively share the task of deciding what to do next. Such interfaces are important to the effective utilization of real-time expert systems as assistants in the execution of critical tasks. Presented here is the Incremental Inference algorithm, a symbolic reasoning mechanism based on propositional logic and suited to the construction of mixed-initiative interfaces. The algorithm is similar in some respects to the Truth Maintenance System, but replaces the notion of 'justifications' with a notion of recency, allowing newer values to override older values yet permitting various interested parties to refresh these values as they become older and thus more vulnerable to change. A simple example is given of the use of the Incremental Inference algorithm plus an overview of the integration of this mechanism within the SPECTRUM expert system for geological interpretation of imaging spectrometer data.

Borchardt, G. C.↗
NASA NTRS2003

Exploring with PAM: Prospecting ANTS Missions for Solar System Surveys

ANTS (Autonomous Nano-Technology Swarm), a large (1000 member) swarm of nano to picoclass (10 to 1 kg) totally autonomous spacecraft, are being developed as a NASA advanced mission concept. ANTS, based on a hierarchical insect social order, use an evolvable, self-similar, hierarchical neural system in which individual spacecraft represent the highest level nodes. ANTS uses swarm intelligence attained through collective, cooperative interactions of the nodes at all levels of the system. At the highest levels this can take the form of cooperative, collective behavior among the individual spacecraft in a very large constellation. The ANTS neural architecture is designed for totally autonomous operation of complex systems including spacecraft constellations. The ANTS (Autonomous Nano Technology Swarm) concept has a number of possible applications. A version of ANTS designed for surveying and determining the resource potential of the asteroid belt, called PAM (Prospecting ANTS Mission), is examined here.

Clark, P. E.↗
NASA NTRS2004

Complexity of life via collective mind

e mind is introduced as a set of simple intelligent units (say, neurons, or interacting agents), which can communicate by exchange of information without explicit global control. Incomplete information is compensated by a sequence of random guesses symmetrically distributed around expectations with prescribed variances. Both the expectations and variances are the invariants characterizing the whole class of agents. These invariants are stored as parameters of the collective mind, while they contribute into dynamical formalism of the agents' evolution, and in particular, into the reflective chains of their nested abstract images of the selves and non-selves. The proposed model consists of the system of stochastic differential equations in the Langevin form representing the motor dynamics, and the corresponding Fokker-Planck equation representing the mental dynamics (Motor dynamics describes the motion in physical space, while mental dynamics simulates the evolution of initial errors in terms of the probability density). The main departure of this model from Newtonian and statistical physics is due to a feedback from the mental to the motor dynamics which makes the Fokker-Planck equation nonlinear. Interpretation of this model from mathematical and physical viewpoints, as well as possible interpretation from biological, psychological, and social viewpoints are discussed. The model is illustrated by the dynamics of a dialog.

collective intelligence↗
NASA NTRSDate not supplied

Quantitative Phenotypic Analysis of Arabidopsis Thaliana Grown in Microgravity Using Soap, an Applied Artificial Intelligence

Phenotypic analysis is an essential step in studying the gravitropic responses and gravitational stress experienced by plants grown in microgravity. Many of the phenotypic traits analyzed in gravitropism studies, such as root length, leaf area, secondary root count, and number of root hairs, currently rely upon manual measurement methods for quantification. However, new advances in data analysis technology using artificial intelligence offer an opportunity for more efficient phenotypic analysis and a reduction of time spent in the data collection phase. In this project, the ability of a new artificially intelligent data collection software, SOAP (Simple Object Access Protocol), to collect and quantify phenotypic traits of Arabidopsis thaliana will be assessed. This project will test the measurements taken by an initial draft of the software. SOAP will take measurements of shoot length, a key phenotype used to assess A. thaliana stress response when grown in microgravity conditions. Shoot length measurements made by SOAP will be compared against a series of manual shoot length measurements. The comparison between the two methods of data measurement will provide valuable insight into the relative accuracy of SOAP and the margin of human error when conducting lab measurements.

Arabidopsis↗