Title |
Preventing undesirable behavior of intelligent machines
|
---|---|
Published in |
Science, November 2019
|
DOI | 10.1126/science.aag3311 |
Pubmed ID | |
Authors |
Philip S Thomas, Bruno Castro da Silva, Andrew G Barto, Stephen Giguere, Yuriy Brun, Emma Brunskill |
Abstract |
Intelligent machines using machine learning algorithms are ubiquitous, ranging from simple data analysis and pattern recognition tools to complex systems that achieve superhuman performance on various tasks. Ensuring that they do not exhibit undesirable behavior-that they do not, for example, cause harm to humans-is therefore a pressing problem. We propose a general and flexible framework for designing machine learning algorithms. This framework simplifies the problem of specifying and regulating undesirable behavior. To show the viability of this framework, we used it to create machine learning algorithms that precluded the dangerous behavior caused by standard machine learning algorithms in our experiments. Our framework for designing machine learning algorithms simplifies the safe and responsible application of machine learning. |
X Demographics
Geographical breakdown
Country | Count | As % |
---|---|---|
United States | 69 | 18% |
India | 13 | 3% |
United Kingdom | 11 | 3% |
Indonesia | 10 | 3% |
Bangladesh | 9 | 2% |
France | 6 | 2% |
Germany | 5 | 1% |
Russia | 5 | 1% |
Italy | 4 | 1% |
Other | 54 | 14% |
Unknown | 190 | 51% |
Demographic breakdown
Type | Count | As % |
---|---|---|
Members of the public | 313 | 83% |
Scientists | 52 | 14% |
Science communicators (journalists, bloggers, editors) | 6 | 2% |
Practitioners (doctors, other healthcare professionals) | 5 | 1% |
Mendeley readers
Geographical breakdown
Country | Count | As % |
---|---|---|
Unknown | 165 | 100% |
Demographic breakdown
Readers by professional status | Count | As % |
---|---|---|
Student > Ph. D. Student | 34 | 21% |
Researcher | 34 | 21% |
Student > Master | 17 | 10% |
Professor | 10 | 6% |
Student > Bachelor | 7 | 4% |
Other | 22 | 13% |
Unknown | 41 | 25% |
Readers by discipline | Count | As % |
---|---|---|
Computer Science | 46 | 28% |
Agricultural and Biological Sciences | 11 | 7% |
Engineering | 10 | 6% |
Social Sciences | 6 | 4% |
Mathematics | 6 | 4% |
Other | 38 | 23% |
Unknown | 48 | 29% |