Machine Learning is the study of methods for programming computers to learn. Computers are applied to a wide range of tasks, and for most of these it is relatively easy for programmers to design and implement the necessary software. However, there are many tasks for which this is difficult or impossible. These can be divided into four general categories. First, there are problems for which there exist no human experts. For example, in modern automated manufacturing facilities, there is a need to predict machine failures before they occur by analyzing sensor readings. Because the machines are new, there are no human experts who can be interviewed by a programmer to provide the knowledge necessary to build a computer system. A machine learning system can study recorded data and subsequent machine failures and learn prediction rules. Second, there are problems where human experts exist, but where they are unable to explain their expertise. This is the case in many perceptual tasks, such as speech recognition, hand-writing recognition, and natural language understanding. Virtually all humans exhibit expert-level abilities on these tasks, but none of them can describe the detailed steps that they follow as they perform them. Fortunately, humans can provide machines with examples of the inputs and correct outputs for these tasks, so machine learning algorithms can learn to map the inputs to the outputs. Third, there are problems where phenomena are changing rapidly. In finance, for example, people would like to predict the future behavior of the stock market, of consumer purchases, or of exchange rates. These behaviors change frequently, so that even if a programmer could construct a good predictive computer program, it would need to be rewritten frequently. A learning program can relieve the programmer of this burden by constantly modifying and tuning a set of learned prediction rules. Fourth, there are applications that need to be customized for each computer user separately. Consider, for example, a program to filter unwanted electronic mail messages. Different users will need different filters. It is unreasonable to expect each user to program his or her own rules, and it is infeasible to provide every user with a software engineer to keep the rules up-to-date. A machine learning system can learn which mail messages the user rejects and maintain the filtering rules automatically. Machine learning addresses many of the same research questions as the fields of statistics, data mining, and psychology, but with differences of emphasis. Statistics focuses on understanding the phenomena that have generated the data, often with the goal of testing different hypotheses about those phenomena. Data mining seeks to find patterns in the data that are understandable by people. Psychological studies of human learning aspire to understand the mechanisms underlying the various learning behaviors exhibited by people (concept learning, skill acquisition, strategy change, etc.).

Machine learning

Human-Robot Interaction (HRI) has recently received considerable attention in the academic community, in labs, in technology companies, and through the media. Because of this attention, it is desirable to present a survey of HRI to serve as a tutorial to people outside the field and to promote discussion of a unified vision of HRI within the field. The goal of this review is to present a unified treatment of HRI-related problems, to identify key themes, and discuss challenge problems that are likely to shape the field in the near future. Although the review follows a survey structure, the goal of presenting a coherent "story" of HRI means that there are necessarily some well-written, intriguing, and influential papers that are not referenced. Instead of trying to survey every paper, we describe the HRI story from multiple perspectives with an eye toward identifying themes that cross applications. The survey attempts to include papers that represent a fair cross section of the universities, government efforts, industry labs, and countries that contribute to HRI, and a cross section of the disciplines that contribute to the field, such as human, factors, robotics, cognitive psychology, and design.

/pdf/human-robot-interaction-a-survey-4q8yi606a9.pdf

Human-Robot Interaction: A Survey

Remote Agent: to boldly go where no AI system has gone before

Multiagent Systems is the title of a collection of papers dedicated to surveying specific themes of Multiagent Systems (MAS) and Distributed Artificial Intelligence (DAI). All of them authored by leading researchers of this dynamic multidisciplinary field.

/pdf/multiagent-systems-a-modern-approach-to-distributed-5fon0donq1.pdf

Multiagent systems: a modern approach to distributed artificial intelligence

On May 17th 1999, NASA activated for the first time an AI-based planner/scheduler running on the flight processor of a spacecraft. This was part of the Remote Agent Experiment (RAX), a demonstration of closed-loop planning and execution, and model-based state inference and failure recovery. This paper describes the RAX Planner/Scheduler (RAX-PS), both in terms of the underlying planning framework and in terms of the fielded planner. RAX-PS plans are networks of constraints, built incrementally by consulting a model of the dynamics of the spacecraft. The RAX-PS planning procedure is formally well defined and can be proved to be complete. RAX-PS generates plans that are temporally flexible, allowing the execution system to adjust to actual plan execution conditions without breaking the plan. The practical aspect, developing a mission critical application, required paying attention to important engineering issues such as the design of methods for programmable search control, knowledge acquisition and planner validation. The result was a system capable of building concurrent plans with over a hundred tasks within the performance requirements of operational, mission-critical software.

/pdf/planning-in-interplanetary-space-theory-and-practice-1qpgdxdkv9.pdf

Planning in interplanetary space: theory and practice

Several successful autonomous systems are separated into technologically diverse functional layers operating at different levels of abstraction. This diversity makes them difficult to implement and validate. In this paper, we present IDEA (Intelligent Distributed Execution Architecture), a unified planning and execution framework. In IDEA a layered system can be implemented as separate agents, one per layer, each representing its interactions with the world in a model. At all levels, the model representation primitives and their semantics is the same. Moreover, each agent relies on a single model, plan database, plan runner and on a variety of planners, both reactive and deliberative. The framework allows the specification of agents that operate, within a guaranteed reaction time and supports flexible specification of reactive vs. deliberative agent behavior. Within the IDEA framework we are working to fully duplicate the functionalities of the DS1 Remote Agent and extend it to domains of higher complexity than autonomous spacecraft control.

/pdf/idea-planning-at-the-core-of-autonomous-reactive-agents-2ysg9qka69.pdf

IDEA: Planning at the Core of Autonomous Reactive Agents

We expect a variety of autonomous systems, from rovers to life-support systems, to play a critical role in the success of manned Mars missions. The crew and ground support personnel will want to control and be informed by these systems at varying levels of detail depending on the situation. Moreover, these systems will need to operate safely in the presence of people and cooperate with them effectively. We call such autonomous systems human-centered in contrast with traditional Oblack-boxO autonomous systems. Our goal is to design a framework for human-centered autonomous systems that enables users to interact with these systems at whatever level of control is most appropriate whenever they so choose, but minimize the necessity for such interaction. This paper discusses on-going research at the NASA Ames Research Center and the Johnson Space Center in developing human-centered autonomous systems that can be used for a manned Mars mission.

/pdf/adjustable-autonomy-for-human-centered-autonomous-systems-on-1yep8j6juo.pdf

Adjustable Autonomy for Human-Centered Autonomous Systems on Mars

This paper describes the Remote Agent flight experiment for spacecraft commanding and control. In the Remote Agent approach, the operational rules and constraints are encoded in the flight software. The software may be considered to be an autonomous "remote agent" of the spacecraft operators in the sense that the operators rely on the agent to achieve particular goals. The experiment will be executed during the flight of NASA's Deep Space One technology validation mission. During the experiment, the spacecraft will not be given the usual detailed sequence of commands to execute. Instead, the spacecraft will be given a list of goals to achieve during the experiment. In flight, the Remote Agent flight software will generate a plan to accomplish the goals and then execute the plan in a robust manner while keeping track of how well the plan is being accomplished. During plan execution, the Remote Agent stays on the lookout for any hardware faults that might require recovery actions or replanning. In addition to describing the design of the remote agent, this paper discusses technology-insertion challenges and the approach used in the Remote Agent approach to address these challenges. The experiment integrates several spacecraft autonomy technologies developed at NASA Ames and the Jet Propulsion Laboratory: on-board planning, a robust multi threaded executive, and model-based failure diagnosis and recovery.

/pdf/design-of-the-remote-agent-experiment-for-spacecraft-4zmaza9dzv.pdf

Design of the Remote Agent experiment for spacecraft autonomy

ion is also critical to the success of planning and scheduling activities. In our scenarios, the crew will often have to deal with planning and scheduling at a very high level (e.g., what crops do I need to plant now so they can be harvested in six months) and planning and scheduling at a detailed level (e.g., what is my next task). The autonomous system must be able to move between various time scales and levels of abstraction, presenting the correct level of information to the user at the correct time. Model-based diagnosis and recovery When something goes wrong, a robust autonomous should figure out what went wrong and recover as best as it can. A model-based diagnosis and recovery system, such as Livingstone [Williams and Nayak, 96], does this. It is analogous to the autonomic and immune systems of a living creature. If the autonomous system has a model of the system it controls, it can use this to figure out what is the most likely cause that explains the observed symptoms as well as how can the system recover given this diagnosis so its mission can continue. For example, if the pressure of a tank is low, it could be because the tank has a leak, the pump blew a fuse, a valve is not open to fill the tank or not closed to keep the tank from draining. However, it could be that the tank pressure is not low and the pressure sensor is defective. By analyzing the system from other sensors, it may say the pressure is normal or suggest closing a valve, resetting the pump circuit breaker, or requesting a crewmember to check the tank for a leak.

https://ti.arc.nasa.gov/m/pub-archive/archive/0099.pdf

Adjustable Autonomy for Human-Centered Autonomous Systems

On May 17th 1999, the Remote Agent (RA) became the first Artificial Intelligence based closed loop autonomous control system to take control of a spacecraft. The RA commanded NASA's New Millennium Deep Space One spacecraft when it was 65 million miles away from earth. For a period of one week this system commanded DS1's Ion Propulsion System, its camera, its attitude control and navigation systems. A primary goal of this experiment was to provide an on-board demonstration of spacecraft autonomy. This demonstration included both nominal operations with goal-oriented commanding and closed-loop plan execution, and fault protection capabilities with failure diagnosis and recovery, on-board replanning following unrecoverable failures, and system-level fault protection.

This paper describes the Remote Agent Experiment and the model based approaches to Planning and Scheduling, Plan Execution and Fault Diagnosis and Recovery technologies developed at NASA Ames Research Center and the Jet Propulsion Laboratory.

/pdf/remote-agent-an-autonomous-control-system-for-the-new-4tw0bv05rb.pdf

Gregory A. Dorais

Papers

IDEA: Planning at the Core of Autonomous Reactive Agents

Adjustable Autonomy for Human-Centered Autonomous Systems on Mars

Design of the Remote Agent experiment for spacecraft autonomy

Adjustable Autonomy for Human-Centered Autonomous Systems

Remote agent: an autonomous control system for the New Millennium