On The Development of Visual Object Memory: The Stay/Go Decision Problem (2008)
Clayton T. Morrison, Paul R. Cohen, Paola Sebastiani
A developing memory requires a mechanism for deciding how much information to gather, based on what is currently represented in memory. That is, we need to know when we have seen enough to say we...
Early Warnings of Plan Failure, Falso Positives and Envelopes: Experiments and a Model * (2008)
Paul R. Cohen, Robert St. Amant, David M. Hart
We analyze a tradeoff between early warnings of plan failures and false positives. In general, a decision rule that provides earlier warnings will also produce more false positives. Slack time...
ABSTRACT Measuring Confidence Intervals in Link Discovery: A Bootstrap Approach (2008)
Jafar Adibi, Paul R. Cohen, Clayton T. Morrison
Recently there has been great interest in the development of
Temporal Data Mining for Educational Applications (2008)
Intelligent tutoring systems (ITS) acquire rich data about students ’ behavior during learning; data mining techniques can help to describe, interpret and predict student behavior, and to evaluate...
Clayton T. Morrison, Paul R. Cohen
Abstract. We present an overview of our work in information fusion for intelligence analysis. This work includes the Hats Simulator and the COLAB system. The Hats Simulator is a parameterized model...
Professional Experience (2008)
Matthew D. Schmill, Advisor Paul, R. Cohen, Committee Neil Berthier, Paul R. Cohen, Rod Grupen, ...
Primary responsibility was design and development of the Meta Cognitive Loop, a domain-general metareasoning system intended to make intelligent systems more robust by allowing them to reason about...
Abstract A Randomized ANOVA Procedure for Comparing Performance Curves (2008)
Justus H. Piater, Paul R. Cohen
Three factors are related in analyses of performance curves such as learning curves: the amount of training, the learning algorithm, and performance. Often we want to know whether the algorithm...
Abstract A Randomized ANOVA Procedure for Comparing Performance Curves (2008)
Justus H. Piater, Paul R. Cohen
Three factors are related in analyses of performance curves such as learning curves: the amount of training, the learning algorithm, and performance. Often we want to know whether the algorithm...
Clustering Main Concepts from e-Mails (2008)
Paul R. Cohen, Josecristóbal Riquelme
E--mail is one of the most common ways to communicate, assuming, in some cases, up to 75% of a company's communication, in which every employee spends about 90 minutes a day in e--mail tasks...
Learning a Deterministic Finite Automaton With a Recurrent Neural Network (2008)
Laura Firoiu, Tim Oates, Paul R. Cohen
We consider the problem of learning a finite automaton with recurrent neural networks, given a training set of sentences in a language. We train Elman recurrent neural networks on the prediction task...
Paul R. Cohen, Roderic A. Grupen, Victor R. Lesser, Krithi Ramamritham, Edward M. Riseman, ...
NSF IRI-9208920
Dynamic Visualization of Battle Simulations (2007)
Paul R. Cohen, James A. Davis, John L. Warwick
We present a case study of visualization in understanding encounters between multiple agents in an adversarial environment. The information visualized consists of time series of attributes and...
Preliminary Design of an Empirical Discovery Agent (2007)
Dawn E. Gregory, Paul R. Cohen
We approach the problem of automated knowledge discovery from an empirical perspective. We are developing an artificial agent, called the Scientist's Empirical Assistant (SEA), that discovers...
Towards an Automated Empirical Assistant (2007)
Dawn E. Gregory, Paul R. Cohen
this paper, we look toward developing an intelligent agent which can support the scientist in each of these stages.
A Distributed Approach to Finding Complex Dependencies in Data (2007)
Matthew D. Schmill, Tim Oates, Paul R. Cohen, Kathryn S. McKinley
Learning complex dependencies from time series data is an important task; dependencies can be used to make predictions and characterize a source of data. We have developed Multi-Stream Dependency...
Empirical Methods, Paul R. Cohen, Reviewed Ron Kohavi
blem. Only 8% of the articles presented results for more than one problem using real world data. While Prechelt (1996) only looked at whether comparisons were done, Cohen went a step further and...
Segregating Planners and Their Environments (2007)
Scott D. Anderson, Paul R. Cohen
By implementing agents and environments using a domain-independent, extensible simulation substrate, described in this paper, agents will have clean interfaces to their environments. This makes it...
Learning a Deterministic Finite Automaton With a Recurrent Neural Network (2007)
Laura Firoiu, Tim Oates, Paul R. Cohen
We consider the problem of learning a finite automaton with recurrent neural networks, given a training set of sentences in a language. We train Elman recurrent neural networks on the prediction task...
Segregating Planners and Their Environments (2007)
By implementing agents and environments using a domain-independent, extensible simulation substrate, described in this paper, agents will have clean interfaces to their environments. These makes it...
Monitoring in Embedded Agents (2007)
Finding good monitoring strategies is an important process in the design of any embedded agent. We describe the nature of the monitoring problem, point out what makes it difficult, and show that...
Finding Structure in Streams (2007)
Finding structure in streams (series of categorical data) is an important task. Consider a patient in an intensive care unit, where monitors record different aspects of the patient's condition....
Charles Sutton, Brendan Burns, Clayton Morrison, Paul R. Cohen
Because uncertain reasoning is often intractable, it is hard to reason with a large amount of knowledge. One solution to this problem is to specify a set of possible models, some simple and some...
Clayton T. Morrison, Erin N. Cannon, Paul R. Cohen, Richard S. Bogartz, Carole R. Beal
The maps-for-verbs framework predicts that our use of verbs to describe simple whole-body interactions is influenced by the characteristics of the physical dynamics in the before, during and after...
Yoshitaka Kuwata, David M. Hart, Paul R. Cohen
When a traffic management system involves many thousands of vehicles using hundreds of streets and highways, it can be difficult or impossible to tell whether the network is flowing smoothly and to...
Abstract Finding Structure in Streams (2007)
Finding structure in streams (series of categorical data) is an important task. Consider a patient in an intensive care unit, where monitors record di erent aspects of the patient's condition....
Learning Predictive Generalizations for Multiple Streams: An Incremental Algorithm (2007)
Matthew D. Schmill, Paul R. Cohen
We present an approach to learning complex dependencies among multiple streams of time-series data incrementally. Given a set of input streams that contain categorical values that change over time,...
Segregating Planners and Their Environments (2007)
Scott D. Anderson, Paul R. Cohen
To be published in the proceedings of the
Abstract Over tting arises when model components are evaluated against the wrong reference distribution. Most modeling algorithms iteratively nd the best of several components and then test whether...
AToolbox for Analyzing Programs (2007)
Scott D. Anderson, David M. Hart, David L. Westbrook, Paul R. Cohen
The paper describes two separate but synergistic tools for running experiments on large Lisp programs. The rst tool, called Clip (Common Lisp Instrumentation Package), allows the researcher to de ne...
Integrating Many Techniques for Discovering Structure in Data (2007)
Dawn E. Gregory, Paul R. Cohen
Abstract. This paper describes a formal representation of the discovery process that e ciently integrates of any number of data analysis strategies, regardless of their similarities and di erences....
Conceptual structures or ontologies are usually built by hand by skilled knowledge engineers. This paper presents a theory of how conceptual structure may be acquired by an intelligent agent...
Identifying Qualitatively Di erent Outcomes of Actions: Experiments with a Mobile Robot (2007)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
We present an unsupervised learning method that allows a situated embodied agent to identify and represent qualitatively di erent outcomes of actions. The initiation of a particular action triggers...
Informatics and Cybernetics. A Case Study of Planning for Exploratory Data Analysis (2007)
Robert St. Amant, Paul R. Cohen
To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be...
Very Predictive Ngrams for Space-Limited Probabilistic Models (2007)
Paul R. Cohen, Charles A. Sutton
Abstract. In sequential prediction tasks, one repeatedly tries to predict the next element in a sequence. A classical way to solve these problems is to fit an order-n Markov model to the data, but...
Abstract Learning What Is Relevant to the E ects of Actions for a Mobile Robot (2007)
Matthew D. Schmill, Michael T. Rosenstein, Paul R. Cohen, Paul Utgo
We have developed a learning mechanism that allows robots to discover the conditional e ects of their actions. Based on sensorimotor experience, this mechanism permits a robot to explore its...
Scott D. Anderson, Adam Carlson, David Westbrook, David M. Hart, Paul R. Cohen
The paper describes two separate but synergistic tools for running experiments on large Lisp systems such as Arti cial Intelligence planning systems, by which we mean systems that produce plans and...
Monitoring in Embedded Agents (2007)
Finding good monitoring strategies is an important process in the design of any embedded agent. We describe the nature of the monitoring problem, point out what makes it di cult, and show that while...
A Dynamical Basis for the Semantic Content of Verbs Much (2007)
ofwhatwe know and say refers to the dynamics of our world. Here we include our physical world, as well as our mental world, the world of social interactions, and other not-entirely-physical...
Towards an Automated Empirical Assistant (2007)
Dawn E. Gregory, Paul R. Cohen
It is not suprising to realize that all sciences have much the same goal: given some kind of system, develop an understanding of its behavior in terms of its task, environment, and architecture[2]....
Abstract Contentful Mental States for Robot Baby (2007)
In this paper we claim that meaningful representations can be learned by programs, although today they are almost always designed by skilled engineers. We discuss several kinds of meaning that...
Early Warnings of Plan Failure, Falso Positives and Envelopes: Experiments and a Model * (2007)
Paul R. Cohen, Robert St. Amant, David M. Hart
We analyze a tradeoff between early warnings of plan failures and false positives. In general, a decision rule that provides earlier warnings will also produce more false positives. Slack time...
Genetic Programming to Learn an Agentis Monitoring Strategy* (2007)
Many tasks require an agent to monitor its environment, but little is known about appropriate monitoring strategies to use in particular situations. Our approach is to learn good monitoring...
J. A. Joines, R. R. Barton, K. Kang, P. A. Fishwick, Marc S. Atkin, Paul R. Cohen
Many artificial intelligence techniques rely on the notion of a “state ” as an abstraction of the actual state of the world, and an “operator ” as an abstraction of the actions that take you...
Proceedings of the 2000 Winter Simulation Conference (2007)
J. A. Joines, R. R. Barton, K. Kang, P. A. Fishwick, Marc S. Atkin, David L. Westbrook, ...
DOMAIN-GENERAL SIMULATION AND PLANNING WITH PHYSICAL SCHEMAS Physical schemas are representations of simple physically grounded relationships and interactions such as “move,” “push, ” and...
Identifying Qualitatively Dierent Outcomes of Actions: Gaining Autonomy Through Learning (2007)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
If robotic agents are to act autonomously they must have the ability to construct and reason about models of their physical environment. In all but the simplest, static domains, such models must...
�Ò ÓÖÔÓÖ�Ø� × �Ò �Ò� � �ØÓÖ Ó � �ÜØ�ÖÒ�Ð ÓÒ��Ø�ÓÒ × �Ò � � � Ø��
1 Detecting Complex Dependencies in Categorical Data (2007)
Tim Oates, Matthew D. Schmill, Dawn E. Gregory, Paul R. Cohen
ABSTRACT Locating and evaluating relationships among values in multiple streams of data is a di cult and important task. Consider the data owing from monitors in an intensive care unit. Readings from...
Evaluation of a Mixed-Initiative Approach to Schedule Maintenance (2007)
Plans and schedules formulated to run in the real world will often fail due to the complexity and unpredictability of the environment. An unexpected, high priority order may arrive at the job shop or...
Learning and transferring action schemas (2007)
Paul R. Cohen, Yu-han Chang, Clayton T. Morrison
Jean is a model of early cognitive development based loosely on Piaget’s theory of sensori-motor and pre-operational thought. Like an infant, Jean repeatedly executes schemas, gradually...
Experimental State Splitting for Transfer Learning (2006)
Clayton T. Morrison, Yu-han Chang, Paul R. Cohen, Joshua Moody
Jean is a model of early cognitive development based loosely on Piaget’s theory of sensori-motor and pre-operational thought (Piaget, 1954). Like an infant, Jean repeatedly executes schemas,...
The Colab Mixed-initiative Analysis Enviroment (2006)
Clayton T. Morrison, Paul R. Cohen
COLAB is an analysis environment in which multiple human analysts in different physical locations can collaborate to build hypotheses of unfolding scenarios. COLAB consists of two components: an...
An image schema language (2006)
Robert St. Amant, Clayton T. Morrison, Yu-han Chang, Paul R. Cohen, Carole Beal
This paper introduces ISL, a language for representing and manipulating image schemas. ISL supports the representation of symbolic as well as quantitative dynamic properties of objects and...
Maps for verbs: The relation between interaction dynamics and verb use (2005)
Paul R. Cohen, Clayton T. Morrison
We report a study of word meaning that tests whether dynamical aspects of movies predict word use. The movies were based on a novel representation of verb semantics called maps for verbs. We asked...
COLAB: A Laboratory Environment for Studying Analyst Sensemaking and Collaboration (2005)
Clayton T. Morrison, Paul R. Cohen
COLAB is a laboratory for studying tools that facilitate collaboration and sensemaking among groups of human analysts as they build interpretations of unfolding situations based on accruing...
Maps for verbs: The relation between interaction dynamics and verb use (2005)
Paul R. Cohen, Clayton T. Morrison
We report a study of word meaning that tests whether dynamical aspects of movies predict word use. The movies were based on a novel representation of verb semantics called maps for verbs. We asked...
A bayesian blackboard for information fusion (2004)
Charles Sutton, Clayton Morrison, Paul R. Cohen, Joshua Moody, Jafar Adibi
Abstract – A Bayesian blackboard is just a conventional, knowledge-based blackboard system in which knowledge sources modify Bayesian networks on the blackboard. As an architecture for intelligence...
Measuring confidence intervals in link discovery: a bootstrap approach (2004)
Jafar Adibi, Paul R. Cohen, Clayton T. Morrison
Recently there has been great interest in the development of information technology for Link Discovery (LD). LD is relevant to a wide range of research topics, including social network analysis,...
When Push Comes to Shove: A Study of the Relation Between Interaction Dynamics and Verb Use (2004)
Clayton T. Morrison, Erin N. Cannon, Paul R. Cohen
A significant portion of our language is devoted to referring to, expressing, and representing the temporally extended dynamics of our world. Following Tomasello (1992), we label the class of words...
Guided incremental construction of belief networks (2003)
Charles Sutton, Brendan Burns, Clayton Morrison, Paul R. Cohen
Abstract. Because uncertain reasoning is often intractable, it is hard to reason with a large amount of knowledge. One solution to this problem is to specify a set of possible models, some simple and...
Tapir: The Evolution of an Agent Control Language (2002)
Gary W. King, Marc S. Atkin, David L. Westbrook, Paul R. Cohen
Tapir is a general purpose, semi-declarative agent control language that extends and enhances the Hierarchical Agent Control (HAC) architecture (Atkin, Westbrook, & Cohen 2000a). Tapir...
Segmenting Time Series with a Hybrid Neural Networks (2002)
This paper describes work on a hybrid HMM/ANN system for finding patterns in a time series, where a pattern is a function that can be approximated by a recurrent neural network embedded in the state...
Using dynamic time warping to bootstrap HMM-based clustering of time series (2001)
Tim Oates, Laura Firoiu, Paul R. Cohen
Abstract. Given a source of time series data, there is often utility in determining whether there are qualitatively dierent regimes in the data and in characterizing those regimes. Hidden Markov...
Some issues in AI engine design (2000)
Marc S. Atkin, Gary W. King, David L. Westbrook, Paul R. Cohen
Our goal is to build an "AI Engine " akin to the graphics engines that have revolutionized some parts of the game industry. A central issue is finding a general way to specify and...
Identifying qualitatively different outcomes of actions: Gaining autonomy through learning (2000)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
oates,schmill,cohen) cs.umass.edu If robotic agents are to act autonomously they must have the ability to construct and reason about models of their physical environment. In all but the simplest,...
Tim Oates, Zachary Eyler-walker, Paul R. Cohen
{ oat es, z walker, co hen} cs. ureas s. edu Even highly autonomous agents must be told what we want them to do. Ideally, we could communicate our goals to agents by talking with them, rather than...
Some issues in AI engine design (2000)
Marc S. Atkin, Gary W. King, David L. Westbrook, Paul R. Cohen
Our goal is to build an \AI Engine " akin to the graphics engines that have revolutionized some parts of the game industry. A central issue is nding a general way to specify and control an...
Learning planning operators in real-world, partially observable environments (2000)
Matthew D. Schmill, Tim Oates, Paul R. Cohen
We are interested in the development of activities in situated, embodied agents such as mobile robots. Central to our theory of development is means-ends analysis planning, and as such, we must rely...
Tim Oates, Zachary Eyler-walker, Paul R. Cohen
Even highly autonomous agents must be told what we want them to do. Ideally, we could communicate our goals to agents by talking with them, rather than encoding our goals in agent-specic knowledge...
A method for clustering the experiences of a mobile robot that accords with human judgements (2000)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
If robotic agents are to act autonomously they must have the ability to construct and reason about models of their physical environment. For example, planning to achieve goals requires knowledge of...
Learning Concepts by Interaction (2000)
This paper presents a theory of how robots may learn concepts by interacting with their environment in an unsupervised way. First, categories of activities are learned, then abstractions over those...
Identifying Qualitatively Different Outcomes of Actions: Experiments with a Mobile Robot (2000)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
We present an unsupervised learning method that allows a situated embodied agent to identify and represent qualitatively different outcomes of actions. The initiation of a particular action triggers...
Identifying Qualitatively Different Experiences: Experiments with a Mobile Robot (2000)
Tim Oates, Matthew Schmill, Paul R. Cohen
We present an unsupervised learning method that allows a situated embodied agent to identify and represent qualitatively different experiences. The occurrence of events, such as the initiation of a...
Probabilistic Dynamic Maps to Visualize the Dynamics of Monte Carlo Simulations (2000)
Paul R. Cohen, John Warwick, Gary W. King, Clayton Morrison, Marco Ramoni, Paola Sebastiani
Probabilistic dynamic maps visualize the dynamics of state variables in Monte Carlo simulations. Each trial in a batch of simulations produces a time series of state variables, and it can be...
Some Issues in AI Engine Design (2000)
Marc S. Atkin, Gary W. King, David L. Westbrook, Paul R. Cohen
Our goal is to build an \AI Engine" akin to the graphics engines that have revolutionized some parts of the game industry. Key to that is nding a general way to specify and control agent's...
Multiple comparisons in induction algorithms (2000)
David D. Jensen, Paul R. Cohen
Abstract. A single mechanism is responsible for three pathologies of induction algorithms: attribute selection errors, overfitting, and oversearching. In each pathology, induction algorithms compare...
A randomized ANOVA procedure for comparing performance curves (1999)
Justus H. Piater, Paul R. Cohen
Three factors enter into analyses of performance curves such as learning curves: the amount of training, the learning algorithm, and performance. Often we want to know whether the algorithm a ects...
Capture the Flag: Military Simulation Meets Computer Games (1999)
Marc S. Atkin, David L. Westbrook, Paul R. Cohen
Some of the more complex AI testbeds are not that different from computer games. It seems that both sides, AI and game design, could pro t from each other's technology. We go a rst step in this...
E cient mining of statistical dependencies (1999)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
The Multi-Stream Dependency Detection algorithm nds rules that capture statistical dependencies between patterns in multivariate time series of categorical data (Oates & Cohen 1996c). Rule...
Learned models for continuous planning (1999)
Matthew D. Schmill, Tim Oates, Paul R. Cohen
We are interested in the nature of activity { structured behavior of nontrivial duration { in intelligent agents. We believe that the development of activity is a continual process in which simpler...
Clustering time series with hidden Markov models and dynamic time warping (1999)
Tim Oates, Laura Firoiu, Paul R. Cohen
Given a source of time series data, suchasthestockmarket or the monitors in an intensive care unit, there is often utility in determining whether there are qualitatively di erent regimes in the data...
Multiple Comparisons in Induction Algorithms (1999)
David D. Jensen, PAUL R. COHEN
. A single mechanism is responsible for three pathologies of induction algorithms: attribute selection errors, overfitting, and oversearching. In each pathology, induction algorithms compare multiple...
Continuous Categories For a Mobile Robot (1999)
Michael T. Rosenstein, Paul R. Cohen
Autonomous agents make frequent use of knowledge in the form of categories | categories of objects, human gestures, web pages, and so on. This paper describes a way for agents to learn such...
Efficient Mining of Statistical Dependencies (1999)
Tim Oates Matthew, Matthew D. Schmill, Paul R. Cohen
The Multi-Stream Dependency Detection algorithm finds rules that capture statistical dependencies between patterns in multivariate time series of categorical data [ Oates and Cohen, 1996c ] . Rule...
A Randomized ANOVA Procedure for Comparing Performance Curves (1999)
Justus Piater, Paul R. Cohen, Xiaoqin Zhang, Michael Atighetchi
Three factors are related in analyses of performance curves such as learning curves: the amount of training, the learning algorithm, and performance. Often we want to know whether the algorithm...
Using Syntax to Learn Semantics: An Experiment in Language Acquisition with a Mobile Robot (1999)
Tim Oates Zachary, Tim Oates, Zachary Eyler-walker, Paul R. Cohen
Children learn natural languages by hearing utterances while interacting with their physical environment. We investigate one aspect of language acquisition by similarly situated, embodied artificial...
Clustering Time Series with Hidden Markov Models and Dynamic Time Warping (1999)
Tim Oates Laura, Laura Firoiu, Paul R. Cohen
this paper presents a method for automatically determining K, the number of generating HMMs, and for learning the parameters of those HMMs.
Clustering Time Series with Hidden Markov Models and Dynamic Time Warping (1999)
Tim Oates, Laura Firoiu, Paul R. Cohen
This paper presents a method for automatically determining K, the number of generating HMMs, and for learning the parameters of those HMMs. An initial estimate of K is obtained by unsupervised...
Efficient Mining of Statistical Dependencies (1999)
Tim Oates, Paul R. Cohen, Casey Durfee
this paper, we describe three methods for reducing the size of the search space that msdd considers and empirically evaluate their utility. The remainder of this section discusses the core msdd...
Efficient Mining of Statistical Dependencies (1999)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
The Multi-Stream Dependency Detection algorithm finds rules that capture statistical dependencies between patterns in multivariate time series of categorical data (Oates & Cohen 1996c). Rule...
Continuous Categories For a Mobile Robot (1999)
Michael T. Rosenstein, Paul R. Cohen
Autonomous agents make frequent use of knowledge in the form of categories --- categories of objects, human gestures, web pages, and so on. This paper describes a way for agents to learn such...
Learned Models for Continuous Planning (1999)
Matthew D. Schmill, Tim Oates, Paul R. Cohen
We are interested in the nature of activity -- structured behavior of nontrivial duration -- in intelligent agents. We believe that the development of activity is a continual process in which simpler...
Capture the Flag: Military Simulation Meets Computer Games (1999)
Marc Atkin, David L. Westbrook, Paul R. Cohen
Some of the more complex AI testbeds are not that different from computer games. It seems that both sides, AI and game design, could profit from each other's technology. We go a first step in...
Clustering Time Series with Hidden Markov Models and Dynamic Time Warping (1999)
Tim Oates, Laura Firoiu, Paul R. Cohen
this paper presents a method for automatically determining K, the number of generating HMMs, and for learning the parameters of those HMMs. An initial estimate of K is obtained by unsupervised...
Efficient Mining of Statistical Dependencies (1999)
Tim Oates, Matthew D. Schmill, Paul R. Cohen, Casey Durfee
The Multi-Stream Dependency Detection algorithm finds rules that capture statistical dependencies between patterns in multivariate time series of categorical data (Oates & Cohen 1996c). Rule...
A Randomized ANOVA Procedure for Comparing Performance Curves (1999)
Three factors enter into analyses of performance curves such as learning curves: the amount of training, the learning algorithm, and performance. Often we want to know whether the algorithm affects...
Continuous Categories For a Mobile Robot (1999)
Michael Rosenstein, Paul R. Cohen
Autonomous agents make frequent use of knowledge in the form of categories --- categories of objects, human gestures, web pages, and so on. This paper describes a way for agents to learn such...
Using Syntax to Learn Semantics: An Experiment in Language Acquisition with a Mobile Robot (1999)
Tim Oates, Zachary Eyler-walker, Paul R. Cohen
Children learn natural languages by hearing utterances while interacting with their physical environment. We investigate one aspect of language acquisition by similarly situated, embodied artificial...
Discovering rules for clustering and predicting asynchronous events (1998)
Tim Oates, David Jensen, Paul R. Cohen
A wide variety of complex systems generate asynchronous events, including nuclear power plants, computer networks, governments, relational database systems and operating systems. We present...
AFS and HAC: Domain-general agent simulation and control (1998)
Marc S. Atkin, David L. Westbrook, Paul R. Cohen, Gregory D. Jorstad
We present two systems for simulating and designing agents, respectively. The rst, the Abstract Force Simulator (AFS), is a domain-general simulator of agents applying forces � many domains can be...
Symbol Grounding with delay coordinates (1998)
Michael T. Rosenstein, Paul R. Cohen
\There is nothing more basic than categorization to our thought, perception, action, and speech " (Lako 1984). Moreover, categories of sensory experience provide the semantic glue between...
Physical Planning and Dynamics (1998)
and continuous. It poses several hard planning problems. We have developed a planner that attempts to exploit Capture the Flag's inherent dynamics, instead of being stymied by them. In...
Concepts From Time Series (1998)
Michael T. Rosenstein, Paul R. Cohen
This paper describes a way of extracting concepts from streams of sensor readings. In particular, we demonstrate the value of attractor reconstruction techniques for transforming time series into...
Symbol Grounding With Delay Coordinates (1998)
Michael T. Rosenstein, Paul R. Cohen
tactic classes in sensorimotor interaction with the environment? We answer this question, not with recurrent neural networks, which require a great deal of training, but with the method of delays,...
Concepts From Time Series (1998)
Michael Rosenstein And, Michael T. Rosenstein, Paul R. Cohen
This paper describes a way of extracting concepts from streams of sensor readings. In particular, we demonstrate the value of attractor reconstruction techniques for transforming time series into...
Learning What Is Relevant to the Effects of Actions for a Mobile Robot (1998)
Matthew D. Schmill, Michael T. Rosenstein, Paul R. Cohen, Paul Utgoff
We have developed a learning mechanism that allows robots to discover the conditional effects of their actions. Based on sensorimotor experience, this mechanism permits a robot to explore its...
Intelligent Support for Exploratory Data Analysis (1998)
Robert St. Amant, Paul R. Cohen
Exploratory data analysis (EDA) is as much a matter of strategy as of selecting specific statistical operations. We have developed a knowledge-based planning system, called Aide, to help users with...
Conceptual structures or ontologies are usually built by hand by skilled knowledge engineers. This paper presents a theory of how conceptual structure may be acquired by an intelligent agent...
Conceptual structures or ontologies are usually built by hand by skilled knowledge engineers. This paper presents a theory of how conceptual structure may be acquired by an intelligent agent...
Interaction with a Mixed-Initiative System for Exploratory Data Analysis (1998)
Robert St, Robert St. Amant, Paul R. Cohen
Exploratory data analysis (EDA) plays an increasingly important role in statistical analysis. EDA is difficult, however, even with the help of modern statistical software. We have developed an...
Learning Regular Languages from Positive Evidence (1998)
Laura Firoiu, Tim Oates, Paul R. Cohen
Children face an enormously difficult task in learning their native language. It is widely believed that they do not receive or make little use of negative evidence (Marcus, 1993), and yet it has...
AFS and HAC: Domain-General Agent Simulation and Control (1998)
Marc Atkin, David L. Westbrook, Paul R. Cohen, Gregory D. Jorstad
We present two systems for simulating and designing agents, respectively. The first, the Abstract Force Simulator (AFS), is a domain-general simulator of agents applying forces; many domains can be...
Physical Planning and Dynamics (1998)
The Capture the Flag domain is uncertain, adversarial, and continuous. It poses several hard planning problems. We have developed a planner that attempts to exploit Capture the Flag's inherent...
Applicability of Reinforcement Learning (1998)
We describe our experiences in trying to implement a hierarchical reinforcement learning system, and follow with conclusions that we have drawn from the difficulties that we encountered. We present...
Automatically Acquiring Rules for Event Correlation From Event Logs Tim (1997)
Tim Oates, David Jensen, Paul R. Cohen
A single fault in a complex network can generate a cascade of events, potentially overloading a manager's console with information. One way to reduce the number of events is event correlation, a...
Neo: Learning conceptual knowledge by sensorimotor interaction with an environment (1997)
Paul R. Cohen, Marc S. Atkin, Tim Oates, Carole R. Beal
Recent developments in philosophy, linguistics, developmental psychology and arti cial intelligence make it possible to envision a developmental path for an arti cial agent, grounded in...
Building an EDA assistant: A progress report (1997)
Robert St. Amant, Paul R. Cohen
Since 1993 we have beenworking on a system to help people with exploratory data analysis (EDA). Aide, an Assistant for Intelligent Data Exploration, is a knowledge-based planning system that...
A family of algorithms for finding temporal structure in data (1997)
Tim Oates, Matthew D. Schmill, David Jensen, Paul R. Cohen
Finding patterns in temporally structured data is an important and di cult problem. Examples of temporally structured data include time series of economic indicators, distributed network status...
Action representation, prediction and concepts (1997)
Michael T. Rosenstein, Paul R. Cohen, Matthew D. Schmill, Marc S. Atkin
A conceptual framework is a valuable resource for planning by situated agents. In this paper, we discuss the acquisition of such a framework. We take the position that concepts are abstractions of...
Neo: Learning conceptual knowledge by sensorimotor interaction with an environment (1997)
Paul R. Cohen, Marc S. Atkin, Tim Oates, Carole R. Beal
Recent developments in philosophy, linguistics, developmental psychology and artificial intelligence make it possible to envision a developmental path for an artificial agent, grounded in...
Dawn E. Gregory, Paul R. Cohen
this document describes the important features and contributions of our work with sea. At this stage, sea supplies a framework for modelling, experimentation, and analysis, and supports a subset of...
Michael T. Rosenstein, Paul R. Cohen, Matthew D. Schmill, and Marc S. Atkin (1997)
Michael T. Rosenstein, Paul R. Cohen, Matthew D. Schmill, Marc S. Atkin
A conceptual framework is a valuable resource for planning by situated agents. In this paper, we discuss the acquisition of such a framework. We take the position that concepts are abstractions of...
Building an EDA Assistant: A Progress Report (1997)
Robert St. Amant, Paul R. Cohen
this paper should take. This research is supported by ARPA/Rome Laboratory under contracts F30602-91C -0076 and F30602-93-C-0010. The U.S. Government is authorized to reproduce and distribute...
A Family of Algorithms for Finding Temporal Structure in Data (1997)
Tim Oates, Matthew D. Schmill, David Jensen, Paul R. Cohen
Finding patterns in temporally structured data is an important and difficult problem. Examples of temporally structured data include time series of economic indicators, distributed network status...
Neo: Learning Conceptual Knowledge by Sensorimotor Interaction with an Environment (1997)
Paul R. Cohen, Marc S. Atkin, Tim Oates, Carole R. Beal
Recent developments in philosophy, linguistics, developmental psychology and artificial intelligence make it possible to envision a developmental path for an artificial agent, grounded in...
Action Representation, Prediction and Concepts (1997)
Michael T. Rosenstein, Paul R. Cohen, Matthew D. Schmill, Marc S. Atkin
A conceptual framework is a valuable resource for planning by situated agents. In this paper, we discuss the acquisition of such a framework. We take the position that concepts are abstractions of...
A Randomized ANOVA Procedure for Comparing Performance Curves (1997)
Justus Piater, Justus H. Piater, Paul R. Cohen, Paul R. Cohen
Three factors enter into analyses of performance curves such as learning curves: the amount of training, the learning algorithm, and performance. Often we want to know whether the algorithm affects...
Evaluation of a Semi-Autonomous Assistant for Exploratory Data Analysis (1997)
Robert St, Robert St. Amant, Paul R. Cohen
Aide is a knowledge-based planning assistant for intelligent data exploration that draws on research in mixed-initiative planning and collaborative systems. Aide incrementally explores a dataset,...
Automatically Acquiring Rules for Event Correlation from Event Logs (1997)
Tim Oates, David Jensen, Paul R. Cohen
A single fault in a complex network can generate a cascade of events, potentially overloading a manager's console with information. One way to reduce the number of events is event correlation, a...
Paul R. Cohen, Tim Oates, Marc S. Atkin
Recent developments in philosophy, linguistics, developmental psychology and arti cial intelligence make it possible to envision a developmental path for an arti cial agent, grounded in...
Getting what you deserve from data (1996)
The focus of this article will appear at rst to be a narrow, prescriptive little corner of the methodological landscape. Data analysis is often dismissed as no more complicated than calculating some...
Parallel and distributed search for structure in multivariate time series (1996)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
E cient data mining algorithms are crucial for e ective knowledge discovery. We present the Multi-Stream Dependency Detection (msdd) data mining algorithm that performs a systematic search for...
Monitoring strategies for embedded agents: Experiments and analysis (1996)
Monitoring is an important activity for any embedded agent. To operate e ectively, agents must gather information about their environment. The policy by which theydo this is called a monitoring...
An empirical study of dynamic scheduling on rings of processors (1996)
Dawn E. Gregory, Lixin Gao, Arnold L. Rosenberg, Paul R. Cohen
We empirically analyze and compare two distributed, low-overhead policies for scheduling dynamic treestructured computations on rings of identical PEs. Our experiments show that both policies give...
On-line planning simulation (1996)
Scott D. Anderson, Paul R. Cohen
Mess is a substrate for building simulation environments suitable for testing plans and on-line or real-time planners. The article describes the design of Mess, howsimu-lations are built and how...
Control representation in an EDA assistant (1996)
Robert St. Amant, Paul R. Cohen
To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be...
Parallel and distributed search for structure in multivariate time series (1996)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
Abstract. E cient data mining algorithms are crucial for e ective knowledge discovery. Wepresent the Multi-Stream Dependency Detection (msdd) data mining algorithm that performs a systematic search...
Learning planning operators with conditional and probabilistic effects (1996)
Providing a complete and accurate domain model for an agent situated in a complex environment can be an extremely di cult task. Actions may have di erent e ects depending on the context in which they...
Monitoring strategies for embedded agents: Experiments and analysis (1996)
Monitoring is an important activity for any embedded agent. To operate effectively, agents must gather information about their environment. The policy by which they do this is called a monitoring...
Design-To-Time Real-Time Scheduling (1996)
Alan J. Garvey, Paul R. Cohen, John A. Stankovic, Shlomo Zilberstein Member, David W. Stemple
DESIGN-TO-TIME REAL-TIME SCHEDULING FEBRUARY 1996 ALAN J. GARVEY B.S., PACIFIC LUTHERAN UNIVERSITY M.S., STANFORD UNIVERSITY Ph.D., UNIVERSITY OF MASSACHUSETTS AMHERST Directed by: Professor Victor...
An Empirical Study of Dynamic Scheduling on Rings of Processors (Extended Abstract) (1996)
Dawn E. Gregory, Lixin Gao, Arnold L. Rosenberg, Paul R. Cohen
Dawn E. Gregory Lixin Gao Arnold L. Rosenberg Paul R. Cohen Department of Computer Science, University of Massachusetts Amherst, MA 01003, USA Abstract We empirically analyze and compare two...
Control Representation in an EDA Assistant (1996)
Robert St Amant, Paul R. Cohen
To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be...
Parallel and Distributed Search for Structure in Multivariate Time Series (1996)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
. Efficient data mining algorithms are crucial for effective knowledge discovery. We present the Multi-Stream Dependency Detection (msdd) data mining algorithm that performs a systematic search for...
Control Representation in an EDA Assistant (1996)
Robert St, Robert St. Amant, Paul R. Cohen
To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be...
Searching for Planning Operators with Context-Dependent and Probabilistic Effects (1996)
Providing a complete and accurate domain model for an agent situated in a complex environment can be an extremely difficult task. Actions may have different effects depending on the context in which...
Getting What You Deserve from Data (1996)
this article will appear at first to be a narrow, prescriptive little corner of the methodological landscape. Data analysis is often dismissed as no more complicated than calculating some means and...
Searching for Structure in Multiple Streams of Data (1996)
Finding structure in multiple streams of data is an important problem. Consider the streams of data flowing from a robot's sensors, the monitors in an intensive care unit, or periodic...
Learning Planning Operators with Conditional and Probabilistic Effects (1996)
Providing a complete and accurate domain model for an agent situated in a complex environment can be an extremely difficult task. Actions may have different effects depending on the context in which...
On-line Planning Simulation (1996)
Scott D. Anderson, Paul R. Cohen
Mess is a substrate for building simulation environments suitable for testing plans and on-line or real-time planners. The article describes the design of Mess, how simulations are built and how...
Paul R. Cohen, Tim Oates, Marc S. Atkin
Recent developments in philosophy, linguistics, developmental psychology and artificial intelligence make it possible to envision a developmental path for an artificial agent, grounded in...
Paul R. Cohen, David Jensen, Producing C
Overfitting arises when model components are evaluated against the wrong reference distribution. Most modeling algorithms iteratively find the best of several components and then test whether this...
Parallel and Distributed Search for Structure in Multivariate Time Series (1996)
Tim Oates, Matthew D. Schmill, Paul R. Cohen
Efficient data mining algorithms are crucial for effective knowledge discovery. We present the Multi-Stream Dependency Detection (msdd) data mining algorithm that performs a systematic search for...
Preliminary system design for an EDA assistant (1995)
Robert St. Amant, Paul R. Cohen
Data analysis plays a central role in our attempts to understand the behavior of complex systems. While research in both statistics and arti cial intelligence has addressed issues in the automation...
Monitoring Progress with Dynamic Programming Envelopes (1995)
Robert St. Amant, Yoshitaka Kuwata, Paul R. Cohen
Envelopes are a form of decision rule for monitoring plan execution. We describe one type, the DP envelope, that draws its decisions from a look-up table computed o-line by dynamic programming. Based...
Detecting complex dependencies in categorical data (1995)
Tim Oates, Dawn Gregory, Paul R. Cohen
Locating and evaluating relationships among values in multiple streams of data is a di cult and important task. Consider the data owing from monitors in an intensive care unit. Readings from various...
Two algorithms for inducing structural equation models from data (1995)
Paul R. Cohen, Dawn E. Gregory, Lisa A. Ballesteros, Robert St. Amant
We present two algorithms for inducing structural equation models from data. Assuming no latent variables, these models have a causal interpretation and their parameters may be estimated by linear...
Tools for detecting dependencies in AI systems (1995)
Tim Oates, Paul R. Cohen, Matthew D. Schmill, Matthew D. Schmill
We present a methodology for learning complex dependencies in data based on streams of categorical, time series data. The streams representation is applicable in a variety of situations: a...
Preliminary System Design for an EDA Assistant (1995)
Robert St Amant, Paul R. Cohen
This paper gives an overview of the design of aide, the Assistant for Intelligent Data Exploration, which assists humans in the early stages of data analysis. The system adopts a script-based...
Tools for Empirically Analyzing AI Programs (1995)
Scott Anderson David, David M. Hart, David L. Westbrook, Paul R. Cohen, Adam Carlson
The paper describes two separate but synergistic tools for running experiments on large Lisp systems such as Artificial Intelligence planning systems, by which we mean systems that produce plans and...
Two Algorithms for Inducing Structural Equation Models from Data (1995)
Paul Cohen Dawn, Paul R. Cohen, Dawn E. Gregory, Lisa A. Ballesteros, Robert St. Amant
We present two algorithms for inducing structural equation models from data. Assuming no latent variables, these models have a causal interpretation and their parameters may be estimated by linear...
Learning Predictive Generalizations for Multiple Streams: An Incremental Algorithm (1995)
Matthew Schmill And, Matthew D. Schmill, Paul R. Cohen
We present an approach to learning complex dependencies among multiple streams of timeseries data incrementally. Given a set of input streams that contain categorical values that change over time, we...
Segregating Planners and Their Environments (1995)
Scott Anderson And, Paul R. Cohen
By implementing agents and environments using a domain-independent, extensible simulation substrate, described in this paper, agents will have clean interfaces to their environments. This makes it...
Finding Structure in Streams (1995)
Paul Cohen And, Paul R. Cohen, Tim Oates
Finding structure in streams (series of categorical data) is an important task. Consider a patient in an intensive care unit, where monitors record different aspects of the patient's condition....
Understanding Planner Behavior (1995)
Adele Howe Computer, Adele E. Howe, Paul R. Cohen
As planners and their environments become increasingly complex, planner behavior becomes increasingly difficult to understand. We often do not understand what causes them to fail, so that we can...
Issues in Automating Exploratory Data Analysis (1995)
Robert St. Amant, Paul R. Cohen
Exploratory data analysis often plays a central role in the early stages of scientific inquiry. Models of complex phenomena are built incrementally, based on suggestive patterns in data and iterative...
A Case Study of Planning for Exploratory Data Analysis (1995)
Robert St Amant, Paul R. Cohen
To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be...
Tools for Detecting Dependencies in AI Systems (1995)
Tim Oates, Paul R. Cohen, Matthew Schmill, Matthew D. Schmill
We present a methodology for learning complex dependencies in data based on streams of categorical, time series data. The streams representation is applicable in a variety of situations: a...
Common Lisp Analytical Statistics Package: User Manual (1995)
Scott D. Anderson, David L. Westbrook, Matthew Schmill, Adam Carlson, David M. Hart, Paul R. Cohen, ...
This document is the user manual for the Common Lisp Analytical Statistics Package (Clasp), a tool for visualizing and statistically analyzing data, and also for the Common Lisp Instrumentation...
A Toolbox for Analyzing Programs (1995)
Scott D. Anderson, David M. Hart, David L. Westbrook, Paul R. Cohen
The paper describes two separate but synergistic tools for running experiments on large Lisp programs. The first tool, called Clip (Common Lisp Instrumentation Package), allows the researcher to...
Preliminary System Design for an EDA Assistant (1995)
Eda Assistant, Robert St. Amant, Paul R. Cohen
This paper gives an overview of the design of aide, the Assistant for Intelligent Data Exploration, which assists humans in the early stages of data analysis. The system adopts a script-based...
Monitoring Progress with Dynamic Programming Envelopes (1995)
Robert St. Amant, Yoshitaka Kuwata, Paul R. Cohen
Envelopes are a form of decision rule for monitoring plan execution. We describe one type, the DP envelope, that draws its decisions from a look-up table computed off-line by dynamic programming....
Two Algorithms for Inducing Structural Equation Models from Data (1995)
Paul R. Cohen, Dawn E. Gregory, Lisa A. Ballesteros, Robert St. Amant
We present two algorithms for inducing structural equation models from data. Assuming no latent variables, these models have a causal interpretation and their parameters may be estimated by linear...
A Case Study of Planning for Exploratory Data Analysis (1995)
Robert St, Robert St. Amant, Paul R. Cohen
To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be...
Understanding Planner Behavior (1995)
As planners and their environments become increasingly complex, planner behavior becomes increasingly difficult to understand. We often do not understand what causes them to fail, so that we can...
Control Representation in an EDA Assistant (1995)
Robert St. Amant, Paul R. Cohen
To develop an initial understanding of complex data, one often begins with exploration. Exploratory data analysis (EDA) provides a set of statistical tools through which patterns in data may be...
Preliminary System Design for an EDA Assistant (1995)
Robert St, Robert St. Amant, Paul R. Cohen
Data Analysis We can best illustrate the EDA approach with a simple example. Much of our research deals with the behavior of AI planners in demanding simulation environments. One such system is...
Tools for Empirically Analyzing AI Programs (1995)
Scott D. Anderson, David M. Hart, David L. Westbrook, Paul R. Cohen, Adam Carlson
The paper describes two separate but synergistic tools for running experiments on large Lisp systems such as Artificial Intelligence planning systems, by which we mean systems that produce plans and...
Toward a plan steering agent: Experiments with schedule maintenance (1994)
When a plan involves hundreds or thousands of events over time it can be di cult or impossible to tell whether those events are unfolding \according to plan " and to assess the impact of...
Tools for experiments in planning (1994)
Scott D. Anderson, Adam Carlson, David M. Hart, Paul R. Cohen
The paper describes two separate but synergistic tools for running experiments on large Lisp systems such as Arti cial Intelligence planning systems, by whichwe mean systems that produce plans and...
Amant, “Regression can build predictive causal models”, Computer Science (1994)
Paul R. Cohen, Lisa A. Ballesteros, Dawn E. Gregory, Robert St. Amant
Covariance information can help an algorithm search for predictive causal models and estimate the strengths of causal relationships. This information should not be discarded after conditional...
Common Lisp Analytical Statistics Package: User Manual (1994)
Scott D. Anderson, Adam Carlson, David L. Westbrook, David M. Hart, Paul R. Cohen, Scott D. Anderson, ...
This document is the user manual for the Common Lisp Analytical Statistics Package (Clasp), a tool for visualizing and statistically analyzing data, and also for the Common Lisp Instrumentation...
A planning representation for automated exploratory data analysis (1994)
Robert St. Amant, Paul R. Cohen
Igor is a knowledge-based system for exploratory statistical analysis of complex systems and environments. Igor has two related goals: to help automate the search for interesting patterns in data...
Toward the integration of exploration and modeling in a planning framework (1994)
Robert St. Amant, Paul R. Cohen
Statistical operations are often facilitated by other operations. We can facilitate modeling operations by testing their input for irregularities and removing problems wherever possible. A planning...
Learning Monitoring Strategies: A Di cult Genetic Programming Application (1994)
Abstract | Finding optimal or at least good monitoring strategies is an important consideration when designing an agent. We have applied genetic programming to this task, with mixed results. Since...
The Interval Reduction Strategy for Monitoring Cupcake problems (1994)
Paul R. Cohen, Marc S. Atkin, Eric A. Hansen
Monitoring is the process by whichagents assess their environments. Most AI applications rely on periodic monitoring, but for a large class of problems this is ine cient. The interval reduction...
Common lisp instrumentation package: User manual (1994)
David L. Westbrook, David L. Westbrook, Scott D. Anderson, Scott D. Anderson, David M. Hart, David M. Hart, ...
This document is the user manual for the Common Lisp Instrumentation Package (Clip), a tool for automating experimentation and data collection. Clip was designed to be used in combination with the...
Automated analysis of complex data (1994)
Robert St. Amant, Paul R. Cohen
Igor is a knowledge-based system for exploratory statistical analysis of complex data. Igor has two related goals: to help automate the search for interesting patterns in data sets, and to help...
Mixed-Initiative Schedule Maintenance: a First Step Toward Plan Steering. To appear (1994)
When a plan involves hundreds or thousands of events over time it can be di cult or impossible to tell whether those events are unfolding \according to plan " and to assess the impact of...
Humans plus agents maintain schedules better than either alone (1994)
Tracking and evaluating the progress of large, complex plans or schedules as they unfold in real time is extremely di cult for humans. In this paper we present a mixed-initiative system for the task...
A planning representation for automated exploratory data analysis (1994)
Robert St. Amant, Paul R. Cohen
Igor is a knowledge-based system for exploratory statistical analysis of complex systems and environments. Igor has two related goals: to help automate the search for interesting patterns in data...
Automated analysis of complex data (1994)
Robert St. Amant, Paul R. Cohen
Igor is a knowledge-based system for exploratory statistical analysis of complex data. Igor has two related goals: to help automate the search for interesting patterns in data sets, and to help...
Mixed-Initiative Schedule Maintenance: A First Step Toward Plan Steering (1994)
Tim Oates And, Tim Oates, Paul R. Cohen
When a plan involves hundreds or thousands of events over time it can be difficult or impossible to tell whether those events are unfolding "according to plan" and to assess the impact of...
Learning Monitoring Strategies: A Difficult Genetic Programming Application (1994)
Marc Atkin And, Marc S. Atkin, Paul R. Cohen
Finding optimal or at least good monitoring strategies is an important consideration when designing an agent. We have applied genetic programming to this task, with mixed results. Since the agent...
The Interval Reduction Strategy for Monitoring Cupcake Problems (1994)
Paul Cohen Marc, Paul R. Cohen, Marc S. Atkin, Eric A. Hansen
Monitoring is the process by which agents assess their environments. Most AI applications rely on periodic monitoring, but for a large class of problems this is inefficient. The interval reduction...
Detecting Complex Dependencies in Categorical Data (1994)
Tim Oates Dawn, Dawn E. Gregory, Paul R. Cohen
Locating and evaluating relationships among values in multiple streams of data is a difficult and important task. Consider the data flowing from monitors in an intensive care unit. Readings from...
Toward the Integration of Exploration and Modeling in a Planning Framework (1994)
Robert St Amant, Paul R. Cohen
Statistical operations are often facilitated by other operations. We can facilitate modeling operations by testing their input for irregularities and removing problems wherever possible. A planning...
Tools for Experiments in Planning (1994)
Scott Anderson Adam, Scott D. Anderson, Adam Carlson, David L. Westbrook, David M. Hart, Paul R. Cohen
The paper describes two separate but synergistic tools for running experiments on large Lisp systems such as planning systems, by which we mean systems that produce plans and execute them in some...
Toward a Plan Steering Agent: Experiments with Schedule Maintenance (1994)
Tim Oates And, Tim Oates, Paul R. Cohen
When a plan involves hundreds or thousands of events over time it can be difficult or impossible to tell whether those events are unfolding "according to plan" and to assess the impact of...
Tools for Experiments in Planning (1994)
Scott D. Anderson, Adam Carlson, David L. Westbrook, David M. Hart, Paul R. Cohen
The paper describes two separate but synergistic tools for running experiments on large Lisp systems such as planning systems, by which we mean systems that produce plans and execute them in some...
Evaluation of a Mixed-Initiative Approach to Schedule Maintenance (1994)
Introduction Plans and schedules formulated to run in the real world will often fail due to the complexity and unpredictability of the environment. An unexpected, high priority order may arrive at...
Detecting Complex Dependencies in Categorical Data (1994)
Tim Oates, Matthew D. Schmill, Dawn E. Gregory, Paul R. Cohen
Locating and evaluating relationships among values in multiple streams of data is a difficult and important task. Consider the data flowing from monitors in an intensive care unit. Readings from...
Detecting Complex Dependencies in Categorical Data (1994)
Tim Oates, Dawn Gregory, Paul R. Cohen
Locating and evaluating relationships among values in multiple streams of data is a difficult and important task. Consider the data flowing from monitors in an intensive care unit. Readings from...
Toward the Integration of Exploration and Modeling in a Planning Framework (1994)
Robert St. Amant, Paul R. Cohen
Statistical operations are often facilitated by other operations. We can facilitate modeling operations by testing their input for irregularities and removing problems wherever possible. A planning...
Two Algorithms for Inducing Causal Models from Data (1994)
Many methods have been developed for inducing cause from statistical data. Those employing linear regression have historically been discounted, due to their inability to distinguish true from...
Detecting Complex Dependencies in Categorical Data (1994)
Tim Oates, Dawn E. Gregory, Paul R. Cohen
Locating and evaluating relationships among values in multiple streams of data is a difficult and important task. Consider the data flowing from monitors in an intensive care unit. Readings from...
Common Lisp Analytical Statistics Package CLASP - Common Lisp Instrumentation Package CLIP (1994)
Scott D. Anderson, Adam Carlson, David L. Westbrook, David M. Hart, Paul R. Cohen, David A. Fisher
This document is the user manual for the Common Lisp Analytical Statistics Package (Clasp), a tool for visualizing and statistically analyzing data, and also for the Common Lisp Instrumentation...
Tools for Experiments in Planning (1994)
Scott D. Anderson, Adam Carlson, David Westbrook, David M. Hart, Paul R., Paul R. Cohen
The paper describes two separate but synergistic tools for running experiments on large Lisp systems such as Artificial Intelligence planning systems, by which we mean systems that produce plans and...
Humans Plus Agents Maintain Schedules Better than Either Alone (1994)
Tracking and evaluating the progress of large, complex plans or schedules as they unfold in real time is extremely difficult for humans. In this paper we present a mixed-initiative system for the...
Learning Monitoring Strategies: A Difficult Genetic Programming Application (1994)
Finding optimal or at least good monitoring strategies is an important consideration when designing an agent. We have applied genetic programming to this task, with mixed results. Since the agent...
Mixed-Initiative Schedule Maintenance: A First Step Toward Plan Steering (1994)
When a plan involves hundreds or thousands of events over time it can be difficult or impossible to tell whether those events are unfolding "according to plan" and to assess the impact of...
Adele E. Howe, Adele E. Howe, Robert St. Amant, Robert St. Amant, Paul R. Cohen, Paul R. Cohen
Given a complex planner and or environment, it can be difficult to determine why it behaves as it does. Statistical causal modeling techniques allow us to develop models of behavior, but they tend to...
Detecting Complex Dependencies in Categorical Data (1994)
Tim Oates, Matthew D. Schmill, Dawn E. Gregory, Paul R. Cohen
Locating and evaluating relationships among values in multiple streams of data is a difficult and important task. Consider the data flowing from monitors in an intensive care unit. Readings from...
Toward a Plan Steering Agent: Experiments with Schedule Maintenance (1994)
When a plan involves hundreds or thousands of events over time it can be difficult or impossible to tell whether those events are unfolding "according to plan" and to assess the impact of...
Path Analysis Models of an Autonomous Agent in a Complex Environment (1994)
Paul R. Cohen, David M. Hart, Robert St. Amant, Lisa A. Ballesteros, Adam Carlson
We seek explanatory models of how and why AI systems work in particular environments. We are not satisfied to demonstrate performance, we want to understand it. In terms of data and models, this...
The Interval Reduction Strategy for Monitoring Cupcake Problems (1994)
Paul R. Cohen, Marc S. Atkin, Eric A. Hansen
Monitoring is the process by which agents assess their environments. Most AI applications rely on periodic monitoring, but for a large class of problems this is inefficient. The interval reduction...
Path analysis models of an autonomous agent in a complex environment (1993)
Paul R. Cohen, David M. Hart, Robert St. Amant, Lisa A. Ballesteros, Adam Carlson
We seek explanatory models of how and why AI systems work in particular environments. We are not satis ed to demonstrate performance, we want to understand it. In terms of data and models, this means...
Visualization tools for real-time search algorithms (1993)
Yoshitaka Kuwata, Paul R. Cohen
Search methods are common mechanisms in problem solving. In many AI applications, they are used with heuristic functions to prune the search space and improve performance. In last three decades, much...
Experimental trials of programs are sometimes aborted when resource bounds are exceeded. The data from these trials are called censored data. This paper discusses the inferences that can be drawn...
Automating Path Analysis for Building Causal Models from Data (1993)
Paul R. Cohen, Adam Carlson, Lisa Ballesteros, Robert St. Amant
Path analysis is a generalization of multiple linear regression that builds models with causal interpretations. It is an exploratory or discovery procedure for finding causal structure in...
Benchmarks, Testbeds, Controlled Experimentation, and the Design of Agent Architectures (1993)
Steve Hanks, Martha E. Pollack, Paul R. Cohen
The methodological underpinnings of AI are slowly changing. Benchmarks, testbeds, and controlled experimentation are becoming more common. While we are optimistic that this change can solidify the...
Steering Traffic Networks (1993)
Yoshitaka Kuwata, David M. Hart, Paul R. Cohen
When a traffic management system involves many thousands of vehicles using hundreds of streets and highways, it can be difficult or impossible to tell whether the network is flowing smoothly and to...
Benchmarks, Testbeds, Controlled Experimentation, and the Design of Agent Architectures (1993)
Martha Pollack, Paul Cohen, Steve Hanks, Steve Hanks, Martha E. Pollack, Paul R. Cohen
The methodological underpinnings of AI are slowly changing. Benchmarks, testbeds, and controlled experimentation are becoming more common. While we are optimistic that this change can solidify the...
Genetic Programming to Learn an Agent's Monitoring Strategy (1993)
Many tasks require an agent to monitor its environment, but little is known about appropriate monitoring strategies to use in particular situations. Our approach is to learn good monitoring...
Visualization Tools for Real-time Search Algorithms (1993)
Yoshitaka Kuwata, Paul R. Cohen
Search methods are common mechanisms in problem solving. In many AI applications, they are used with heuristic functions to prune the search space and improve performance. In last three decades, much...
Automating Path Analysis for Building Causal Models from Data (1993)
Paul R. Cohen, Adam Carlson, Lisa Ballesteros, Robert St. Amant
Path analysis is a generalization of multiple linear regression that builds models with causal interpretations. It is an exploratory or discovery procedure for finding causal structure in...
Experimental trials of programs are sometimes aborted when resource bounds are exceeded. The data from these trials are called censored data. This paper discusses the inferences that can be drawn...
Automating Path Analysis for Building Causal Models from Data (1993)
Paul R. Cohen, Adam Carlson, Lisa Ballesteros, Robert St. Amant
Path analysis is a generalization of multiple linear regression that builds models with causal interpretations. It is an exploratory or discovery procedure for finding causal structure in...
Predicting and explaining success and task duration in the phoenix planner (1992)
Phoenix is a multi-agent planning system that fights simulated forest fires. In this paper we describe an experiment with Phoenix in which we uncover factors that affect the planner's behavior...
Learning A Decision Rule for Monitoring Tasks with Deadlines (1992)
A real-time scheduler or planner is responsible for managing tasks with deadlines. When the time required to execute a task is uncertain, it may be useful to monitor the task to predict whether it...
Debugging Plan Failures by Analyzing Execution Traces (1992)
Debugging plan failures involves isolating the actions and events that contributed to the failure and explaining how they caused the failure. In this paper, we present an approach to the first step...
AI Magazine, Vol. 11, No. 4, pp. 16-41, 1991. A survey of 150 papers from the Proceedings of the Eighth National Conference on Artificial Intelligence (AAAI-90) shows that AI research follows two...
A survey of 150 papers from the Proceedings of the Eighth National Conference on Artificial Intelligence (AAAI-90) shows that AI research follows two methodologies, each incomplete with respect to...
A survey of 150 papers from the Proceedings of the Eighth National Conference on Artificial Intelligence (AAAI-90) shows that AI research follows two methodologies, each incomplete with respect to...
Envelopes as a vehicle for improving the e ciency of plan execution (1990)
David M. Hart, Scott D. Anderson, Paul R. Cohen
Envelopes are structures which capture expectations of the progress of a plan. By comparing expected progress with actual progress, envelopes can notify the planner when the plan violates those...
Envelopes as a Vehicle for Improving the Efficiency of Plan Execution (1990)
David M. Hart, Scott D. Anderson, Paul R. Cohen
Envelopes are structures which capture expectations of the progress of a plan. By comparing expected progress with actual progress, envelopes can notify the planner when the plan violates those...
Projections as Concepts (1984)
What do the first concepts look like? I propose that the earliest concepts learned by infants are abstractions of activities. The semantics of these concepts are predictive---a good abstraction is...
A theory of heuristic reasoning about uncertainty (1983)
Paul R. Cohen, Milton R. Grinberg
This article describes a theory of reasoning about uncertainty, baaed on a representation of states of certainty called endorsements The theory of endorsements is an alternative to numerical methods...
Learning Planning Operators with Conditional and Probabilistic Effects
Providing a complete and accurate domain model for an agent situated in a complex environment can be an extremely difficult task. Actions may have different effects depending on the context in which...
Discovering Rules for Clustering and Predicting Asynchronous Events
Tim Oates, David Jensen, Paul R. Cohen
A wide variety of complex systems generate asynchronous events, including nuclear power plants, computer networks, governments, relational database systems and operating systems. We present...
Discovering Rules for Clustering and Predicting Asynchronous Events
Tim Oates, David Jensen, Paul R. Cohen
A wide variety of complex systems generate asynchronous events, including nuclear power plants, computer networks, governments, relational database systems and operating systems. We present...