« Biologiska klockor påverkas av hjärnans temperatur? | Main | Bok: "Modeling the Internet and the Web" »
juli 17, 2003
Data mining-forskare försvarar sitt fält
I ett öppet brev "Data Mining" Is NOT Against Civil Liberties (PDF-fil) förklarar data mining-forskare i en av de mest ansedda sammanslutningarna (ACM Special Interest Group on Knowledge, Discovery and Data Mining, SIGKDD) sin inställning till data mining och individens rättigheter, speciellt med anledning av diskussionerna kring det amerikanska TIA-projektet (numera uttytt "Terrorism Information Awareness" tidigare "Total Information Awareness").
The main goal of this letter is to help differentiate the data mining technology from data collection and specific applications in specific domains. We believe that the most signif icant sources of danger to civil liberties are the unnecessary and unauthorized collection of data, and misuse of collected data, including the use of wrong data, the use of data in unauthorized ways, the wrong and unauthorized dissemination of data, and reaching wrong conclusions from data.
We are concerned that recent newsmedia reports and a recent press release from the US ACM Public Policy Committee (see http://www.acm.org/usacm/Letters/tia_final.html) may have contributed to this misimpression. We are concerned that the proposed S. 188 Data Mining Moratorium Act of 2003 (see http://feingold.senate.gov/~feingold/releases/03/01/2003116745.html ) does not reflect a sound understanding of data mining science, technology or applications. Finally, we are concerned that the public debate has not distinguished between the research and development of data mining technology and the application and use of these technologies by specific agencies on specific data for specific purposes.
...Data mining is but one of many technologies that may be used in these projects. Other technologies include database management, online analytical processing, speech recognition, image (face, iris, fingerprint, etc.) recognition, natural language understanding and translation, data warehousing, data integration, information retrieval, etc. Does it make sense to attempt to outlaw any or all of these?
I notisen Can Total Information Awareness work or how can you separate bad coins from good ones? skrivs lite om de sannolikhetsteoretiska problemen som finns i TIA. Se även The Homeland Security Act and the proposed DARPA "Total Information Awareness" (TIA) program.
Posted by hakank at juli 17, 2003 10:47 FM Posted to Machine learning/data mining