skwb

20 ms·

151.
by skwb 8y ago
These are only a few data points in a almost half decade of promising for healthcare solutions and failing to meet those expectations. The MD Anderson audit is particularly bad: https://www.utsystem.edu/sites/default&#x
152.
by skwb 8y ago
In my own experience, I've found it much easier to install tensorflow with GPU support from scratch (including cuDNN...) than it is to install Spark/Hadoop.
153.
by skwb 8y ago
Sounds more like TPS model.. "Build a culture of stopping to fix problems, to get quality right from the start. (Jidoka)" [ https://en.wikipedia.org/wiki/Toyota_Production_System ]
154.
by skwb 8y ago
Some companies are currently working in applying deep learning techniques. Arterys has cardiac, lung, and liver products...
155.
by skwb 8y ago
But due to insurance churn, you don’t want to be footing the bill for your competitor down the road. As an insurance manager, your unit of analysis is not healthcare system or society, but your quarterly bottom line.
156.
by skwb 8y ago
I would first start our by using CT since soft tissue subtraction is a very easy task. This is because soft tissue has a certain density compared to bone (houndsfield unit [1].) You would maybe need to smooth for a good 3D print, but it&#x2
157.
by skwb 8y ago
Depends on the data transfer agreements. However, there are plenty of way you can use public data if even you don't end up using the actual data in the final commercialized model. You just got to be clever. ;)
158.
by skwb 8y ago
Costs in healthcare aren't very "real". US providers use what they call a charge master to determine prices. So while an "IV" may show up on the bill, it's not the actual IV you're purchasing when you&#x27
159.
by skwb 8y ago
Google has a long long history of failing in the medical space. They tend to overengineer their products in a way that makes sense if you’re a computer engineer, but not if you understand anything about the healthcare system. Facebook has n
160.
by skwb 8y ago
I'll believe it if I see it at RSNA 2019.
161.
by skwb 8y ago
MRI manufacturers want to develop products that bring on new functionality and ability to justify their prices. A brand new 1.5T scanner easily costs several MIL to make, and for a clinic to purchase one will have to fulfill a business case
162.
by skwb 8y ago
To be honest I am highly skeptical that FB would be able to have the sort of institutional capacity to launch any sort of medical device health product. It takes a non trivial amount of effort and time to fully productize it (in terms of
163.
by skwb 8y ago
Depends on what you're trying to image. Tbh, if it takes 15 mins vs 30 mins to get an image of your knee, it's maybe not really that much of a game changer. From the perspective of a MRI purchaser (say clinic or hospital director)
164.
by skwb 8y ago
I find it somewhat interesting that purely analytical techniques can be replicated using CNNs. Sorta goes to show that there's a sorta divergence in a purely EECS method vs some of the newer approaches that are able to integrate advanc
165.
by skwb 8y ago
Or bioinformatics. Or bioengineering.
166.
by skwb 8y ago
You don't unit test a NN. You can unit test certain functions, but fundamentally this is an integration test. This is why serious scientific training is needed to understand these complex systems when health and safety are on the line.
167.
by skwb 8y ago
No, it's still the issue. If there's a bug in the software code for a programmable pacemaker, it doesn't matter if the fundamental design is flawed. For governmental regulators, it's the actual implimentation and real wo
168.
by skwb 8y ago
I feel her fundamental problem was that she was mistook "we always do it this way because it historically has been such" with "we always do it this way because there is a scientific reason".
169.
by skwb 8y ago
Considering this is the same Tim Draper that wanted to separate California into 6 states that would effectively concentrate silicon valley wealth while leaving counties to the eastern states some of the poorest in the nation just to weaken
170.
by skwb 8y ago
The problem is that physiologically the blood you get for chem-10 (stuff like Sodium/Potassium) are completely invalid using finger prick (you rupture too many cells and they leak out their intracellular fluids, messing with your resul
171.
by skwb 8y ago
Does this support modern health informatics interoperability standards currently? I'm particularly interested if there is good interface with this platform and FHIR/HL7.
172.
by skwb 8y ago
It's fairly common at my university. You publish 3 papers surrounding a singular theme or topic and you use that for your PhD. You still need to write an intro and have some boilerplate stuff, but it's not uncommon at all.
173.
by skwb 8y ago
You forgot to mention the permutation of companies "wanting" 2), but using the search strategy optimized for 1) approach due to: - poor HR communication - poor ability of hiring managers to know what the job entails - hiring manag
174.
by skwb 8y ago
The major problem from the federated model is that this assumes a blind trust of the labels for the learning you're trying to do. I find this approach frankly lazy and trying to make up for not having good data (and importantly good in
175.
by skwb 8y ago
Holy crap $60?! Talk about a unitasker. There's plenty of great resources to make clear cocktail ice without needing to spend any extra money. https://www.youtube.com/watch?v=sNohIFrbkXM
176.
by skwb 8y ago
Care to share a citation on the study?
177.
by skwb 8y ago
You can actually get decent milk foam these days with automatic frothers. James Hoffman reviewed an automatic espresso machine, and said that he finds little excuse anymore for subpar latte foam when you can achieve such high quality in a d
178.
by skwb 8y ago
| Tourists and social media influencers do not positively contribute, IMO. Also see: Fyre
179.
by skwb 8y ago
One thing I am a bit sad to not take off more is enchanted devices. I sorta feel it was a CES 2015-2017 news cycle coverage that was ultimately unable to have traction. The concept of integrating the design of devices into manufacturing and
180.
by skwb 8y ago
What is the plotting experience like though? As I previously mentioned, plotting is one of the main reasons our group uses python. Another reason now that I think about it, is the number of scientific libraries that I can just "pip ins

More ›