deep learning – The Health Care Blog

Get Ready for Deepfakes

Oct 13, 2021

By KIM BELLARD

The Tom Cruise TikTok deepfakes last spring didn’t spur me into writing about deepfakes, not even when Justin Bieber fell so hard for them that he challenged the deepfake to a fight. When 60 Minutes covered the topic last night though, I figured I’d best get to it before I missed this particular wave.

We’re already living in an era of unprecedented misinformation/disinformation, as we’ve seen repeatedly with COVID-19 (e.g., hydroxychloroquine, ivermectin, anti-vaxxers), but deepfakes should alert us that we haven’t seen anything yet.

ICYMI, here’s the 60 Minutes story:

The trick behind deepfakes is a type of deep learning called “generative adversarial network” (GAN), which basically means neural networks compete on which can generate the most realistic media (e.g., audio or video). They can be trying to replicate a real person, or creating entirely fictitious people. The more they iterate, the most realistic the output gets.

The Rise and Rise of Quantitative Cassandras

Oct 7, 2019

By SAURABH JHA, MD

Despite an area under the ROC curve of 1, Cassandra’s prophesies were never believed. She neither hedged nor relied on retrospective data – her predictions, such as the Trojan war, were prospectively validated. In medicine, a new type of Cassandra has emerged – one who speaks in probabilistic tongue, forked unevenly between the probability of being right and the possibility of being wrong. One who, by conceding that she may be categorically wrong, is technically never wrong. We call these new Minervas “predictions.” The Owl of Minerva flies above its denominator.

Deep learning (DL) promises to transform the prediction industry from a stepping stone for academic promotion and tenure to something vaguely useful for clinicians at the patient’s bedside. Economists studying AI believe that AI is revolutionary, revolutionary like the steam engine and the internet, because it better predicts.

Recently published in Nature, a sophisticated DL algorithm was able to predict acute kidney injury (AKI), continuously, in hospitalized patients by extracting data from their electronic health records (EHRs). The algorithm interrogated nearly million EHRS of patients in Veteran Affairs hospitals. As intriguing as their methodology is, it’s less interesting than their results. For every correct prediction of AKI, there were two false positives. The false alarms would have made Cassandra blush, but they’re not bad for prognostic medicine. The DL- generated ROC curve stands head and shoulders above the diagonal representing randomness.

The researchers used a technique called “ablation analysis.” I have no idea how that works but it sounds clever. Let me make a humble prophesy of my own – if unleashed at the bedside the AKI-specific, DL-augmented Cassandra could unleash havoc of a scale one struggles to comprehend.

Leaving aside that the accuracy of algorithms trained retrospectively falls in the real world – as doctors know, there’s a difference between book knowledge and practical knowledge – the major problem is the effect availability of information has on decision making. Prediction is fundamentally information. Information changes us.