RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
RT @SteveJoffe: Reidentification of participants in "deidentified" data sets: not just a challenge for genomics any more. https://t.co/msQg…
RT @jeremyprovance: As machine learning and industrial data warehousing (not only for health) technologies continue to acclerate, a thoughtful and enforceable code of data ethics is critical to prevent harm. See @dpatil for further reading. https://t.co/VK
RT @LibSkrat: Repeating for the people, especially librarians, in the back: WE NEED TO ADMIT DEIDENTIFICATION IS JUST NOT POSSIBLE. ALA’s…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @SteveJoffe: Reidentification of participants in "deidentified" data sets: not just a challenge for genomics any more. https://t.co/msQg…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @LibSkrat: Repeating for the people, especially librarians, in the back: WE NEED TO ADMIT DEIDENTIFICATION IS JUST NOT POSSIBLE. ALA’s…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @LibSkrat: Repeating for the people, especially librarians, in the back: WE NEED TO ADMIT DEIDENTIFICATION IS JUST NOT POSSIBLE. ALA’s…
RT @DavidCMohr: Accelerometery data can be used to identify individuals. We need better ways of protecting privacy while allowing data shar…
I’ve been involved in reifentification projects. We could reidemtofybover 85%, and that was with just one additional database (insurance/Medicare/Medicaid). Give me one or two more datasets and I would be over 95%.
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @jeremyprovance: As machine learning and industrial data warehousing (not only for health) technologies continue to acclerate, a thought…
Seems what we're learning from #MachineLearning is that personal data is not secure (and the need for ever-increasing #research safeguards may be inevitable) https://t.co/YJpZkLuIAW via @JAMA_current With #ArtificialIntelligence we must not have Real I
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @jeremyprovance: As machine learning and industrial data warehousing (not only for health) technologies continue to acclerate, a thought…
As machine learning and industrial data warehousing (not only for health) technologies continue to acclerate, a thoughtful and enforceable code of data ethics is critical to prevent harm. See @dpatil for further reading.
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
RT @hmkyale: This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity d…
This is big deal…regarding national health surveys…"this study suggest current practices for deidentifying physical activity data insufficient for privacy” We need to admit deidentification is just not possible. More about risk than complete protection. ht
Can you Reidentify Individuals in Physical Activity Data From Which Protected Health Information Has Been Removed using #MachineLearning algorithms? #nhanes #epitwitter #dataprivacy #dataethics #DataScience https://t.co/adh1AiShIK
Accelerometery data can be used to identify individuals. We need better ways of protecting privacy while allowing data sharing to promote open science. https://t.co/HJ2z5109Sa
RT @SeanKhozin: Study using #MachineLearning methods successfully reidentified protected health information in accelerometer collected phys…
RT @SeanKhozin: Study using #MachineLearning methods successfully reidentified protected health information in accelerometer collected phys…
Study using #MachineLearning methods successfully reidentified protected health information in accelerometer collected physical activity data #datasharing #digitalhealth #AI #dataprivacy https://t.co/5XgQG2VOkb
RT @IAmSusanPersky: Over and over again, we see examples of deidentified data being reidentified in different domains - big implications fo…
RT @tamir_i: Study demonstrates that data anonymization techniques (frequently used to avoid the need to comply with privacy and confidenti…
RT @IAmSusanPersky: Over and over again, we see examples of deidentified data being reidentified in different domains - big implications fo…
well then… https://t.co/TzJHHRuw35
RT @DrSherryPagoto Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raising new concerns about confidentiality. https://t.co/eMxuDF1awt https://t.co/SZRywxiRuD
RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
This study has huge privacy implications for anyone collecting #physicalactivity accelerometer data. If you're sharing deidentified data, you should reconsider your aggregation level @DavidRaichlen @HermanPontzer
This is a really critical point to consider, esp in terms of sharing data.
RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
Final thought after reading this fascinating paper (https://t.co/iiKYt8caO1) and the NYT story on Facebook (As Facebook Raised a Privacy Wall, It Carved an Opening for Tech Giants https://t.co/KnTLuYi1RE): privacy may be no more than an illusion with dig
But closer inspection shows that it’s the combination of activity AND demographic data that made reidentification possible https://t.co/iiKYt8caO1 https://t.co/cbDjpspZb9
Woah - physical activity data in 20 min epochs dramatically increased accuracy of data reidentification vs demographics alone, especially for adults https://t.co/iiKYt8caO1
RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
Over and over again, we see examples of deidentified data being reidentified in different domains - big implications for data sharing and data privacy
RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
Anonymous #patientdata may not be as private as previously thought https://t.co/R7wOq1Vndo | Linda Carroll @Reuters_Health | study by Liangyuan Na et al https://t.co/0G3DJOblSE | #anonymisation #reidentification #dataprotection
RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets From Which Protected Health Information Has Been Removed With Use of Machine Learning https://t.co/RiedVLbcPq https://t.co/2dWYpTXOxr
Yikes.
In many cases, like this one, it’s really the volume and type of information that provides the risk. In a time of increasing commitment to open science and data sharing, we need to find ways to minimize these risks. Using summary scores in data share? Othe
RT @DrSherryPagoto: Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raisin…
Can machine learning be used to reidentify deidentified data? This @JAMA_current study shows that indeed it can, raising new concerns about confidentiality. https://t.co/VH9BEmMZIi
RT @tamir_i: Study demonstrates that data anonymization techniques (frequently used to avoid the need to comply with privacy and confidenti…
Feasibility of Reidentifying Individuals in Large National Physical Activity Data Sets, From Which Protected Health Information Has Been Removed, With Use of Machine Learning https://t.co/pXvMl4xtcs
Feasibility of Reidentifying Individuals by Their Protected Health Information https://t.co/iuV08snxHx @JAMANetworkOpen @JAMANetwork
RT @tamir_i: Study demonstrates that data anonymization techniques (frequently used to avoid the need to comply with privacy and confidenti…
RT @tamir_i: Study demonstrates that data anonymization techniques (frequently used to avoid the need to comply with privacy and confidenti…
Study demonstrates that data anonymization techniques (frequently used to avoid the need to comply with privacy and confidentiality obligations) can be rendered broadly ineffective through AI-driven re-identification techniques: https://t.co/rSlk19MPNl
RT @CLTCBerkeley: In a new paper, #CLTCgrantee @anil_aswani finds that despite data aggregation and removal of protected health information…
In a new paper, #CLTCgrantee @anil_aswani finds that despite data aggregation and removal of protected health information, there is concern that deidentified physical activity data collected from wearable devices can be reidentified. Read the report: https