Intelligent Assistant for Computational Scientists: Integrated Modelling, Experimentation and Analysis

Dawn E. Gregory, Paul R. Cohen
Proceedings of the Sixth International Workshop on Artificial Intelligence and Statistics, PMLR R1:231-238, 1997.

Abstract

Computing technology has changed the way scientists work. Among the contributions of this new paradigm are the computational sciences, which involve the study of computer simulations rather than physical systems. This transition to a simulated world carries with it an important scientific advantage: the opportunity to run experiments that are expensive, dangerous, or impossible in the real world. Unfortunately, such experiments are often too easy, and the scientist is overwhelmed with empirical data. The fields of Artificial Intelligence (AI) and Statistics are concerned with modelling and analyzing such large bodies of data. AI employs heuristic reasoning and knowledge to select potential models, and statistical analysis verifies a proposed model. The combinatio  of knowledge-based, heuristic, and statistical techniques is quite successful at modelling experiment data (e.g. [11]). Our goal is to provide an intelligent, integrated environment for scientific modelling, experimentation , and analysis, called the Scientist ’s Empirical Assistant (SEA). SEA is an assistant to human scientists: it automates model generation and verification, experiment design and data collection, but also relies on a human user for guidance, domain knowledge, and decision-making. SEA designs and runs prospective experiments with a simulator, allowing it to draw stronger conclusions than with post-hoc data analysis alone. SEA employs a variety of techniques from both AI and Statistics. It uses heuristic and knowledge-based reasoning to propose models, design experiments, and select analyses. It applies statistical techniques to verify models against experiment data. It develops plans to direct its course of action, and learns which plans are most successful based on past experience. Finally, it models the knowledge of the user to ensure its suggestions and decisions are appropriate.

Cite this Paper

BibTeX
@InProceedings{pmlr-vR1-gregory97a, title = {Intelligent Assistant for Computational Scientists: Integrated Modelling, Experimentation and Analysis}, author = {Gregory, Dawn E. and Cohen, Paul R.}, booktitle = {Proceedings of the Sixth International Workshop on Artificial Intelligence and Statistics}, pages = {231--238}, year = {1997}, editor = {Madigan, David and Smyth, Padhraic}, volume = {R1}, series = {Proceedings of Machine Learning Research}, month = {04--07 Jan}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/r1/gregory97a/gregory97a.pdf}, url = {https://proceedings.mlr.press/r1/gregory97a.html}, abstract = {Computing technology has changed the way scientists work. Among the contributions of this new paradigm are the computational sciences, which involve the study of computer simulations rather than physical systems. This transition to a simulated world carries with it an important scientific advantage: the opportunity to run experiments that are expensive, dangerous, or impossible in the real world. Unfortunately, such experiments are often too easy, and the scientist is overwhelmed with empirical data. The fields of Artificial Intelligence (AI) and Statistics are concerned with modelling and analyzing such large bodies of data. AI employs heuristic reasoning and knowledge to select potential models, and statistical analysis verifies a proposed model. The combinatio  of knowledge-based, heuristic, and statistical techniques is quite successful at modelling experiment data (e.g. [11]). Our goal is to provide an intelligent, integrated environment for scientific modelling, experimentation , and analysis, called the Scientist ’s Empirical Assistant (SEA). SEA is an assistant to human scientists: it automates model generation and verification, experiment design and data collection, but also relies on a human user for guidance, domain knowledge, and decision-making. SEA designs and runs prospective experiments with a simulator, allowing it to draw stronger conclusions than with post-hoc data analysis alone. SEA employs a variety of techniques from both AI and Statistics. It uses heuristic and knowledge-based reasoning to propose models, design experiments, and select analyses. It applies statistical techniques to verify models against experiment data. It develops plans to direct its course of action, and learns which plans are most successful based on past experience. Finally, it models the knowledge of the user to ensure its suggestions and decisions are appropriate.}, note = {Reissued by PMLR on 30 March 2021.} }
Endnote
%0 Conference Paper %T Intelligent Assistant for Computational Scientists: Integrated Modelling, Experimentation and Analysis %A Dawn E. Gregory %A Paul R. Cohen %B Proceedings of the Sixth International Workshop on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 1997 %E David Madigan %E Padhraic Smyth %F pmlr-vR1-gregory97a %I PMLR %P 231--238 %U https://proceedings.mlr.press/r1/gregory97a.html %V R1 %X Computing technology has changed the way scientists work. Among the contributions of this new paradigm are the computational sciences, which involve the study of computer simulations rather than physical systems. This transition to a simulated world carries with it an important scientific advantage: the opportunity to run experiments that are expensive, dangerous, or impossible in the real world. Unfortunately, such experiments are often too easy, and the scientist is overwhelmed with empirical data. The fields of Artificial Intelligence (AI) and Statistics are concerned with modelling and analyzing such large bodies of data. AI employs heuristic reasoning and knowledge to select potential models, and statistical analysis verifies a proposed model. The combinatio  of knowledge-based, heuristic, and statistical techniques is quite successful at modelling experiment data (e.g. [11]). Our goal is to provide an intelligent, integrated environment for scientific modelling, experimentation , and analysis, called the Scientist ’s Empirical Assistant (SEA). SEA is an assistant to human scientists: it automates model generation and verification, experiment design and data collection, but also relies on a human user for guidance, domain knowledge, and decision-making. SEA designs and runs prospective experiments with a simulator, allowing it to draw stronger conclusions than with post-hoc data analysis alone. SEA employs a variety of techniques from both AI and Statistics. It uses heuristic and knowledge-based reasoning to propose models, design experiments, and select analyses. It applies statistical techniques to verify models against experiment data. It develops plans to direct its course of action, and learns which plans are most successful based on past experience. Finally, it models the knowledge of the user to ensure its suggestions and decisions are appropriate. %Z Reissued by PMLR on 30 March 2021.
APA
Gregory, D.E. & Cohen, P.R.. (1997). Intelligent Assistant for Computational Scientists: Integrated Modelling, Experimentation and Analysis. Proceedings of the Sixth International Workshop on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research R1:231-238 Available from https://proceedings.mlr.press/r1/gregory97a.html. Reissued by PMLR on 30 March 2021.

Related Material