256lie

5 ms·

1.
by 256lie 4y ago
conformal prediction
2.
FDA Cleared AI Algorithms(models.acrdsi.org)
1 points by 256lie 5y ago|0 comments
3.
by 256lie 5y ago
There are efforts like the UK Biobank but healthcare institutions are very sensitive about their patient data.
4.
by 256lie 5y ago
For breast screening, this task is high volume and low prevalence and AI can help with radiologist burnout from increased caseload.
5.
by 256lie 5y ago
For a breast screening application, it will always be confirmed with manual review before biopsy.
6.
by 256lie 5y ago
For screening, it depends on the false positives rate. A radiologist with have to check every positive prediction. Although, I believe in Europe, they have approved AI to be used as a second reader.
7.
by 256lie 5y ago
Even the average radiologist is high variable, not to mention inter-reader variability.
8.
by 256lie 5y ago
The article is simplified (a retrospective metastudy) and might not be indicative of what real-life performance. Even reader studies (which would be more rigorous) skip so much that would be crucial to actual deployment (integration into th
9.
by 256lie 5y ago
Clinical AI (which is currently regulated as a CAD medical device by the FDA) won't replace radiologists but treated as an additional clinical vendor application integrated into existing software. Similar to speech recognition diction
10.
by 256lie 5y ago
It's not an issue of resolution but of generalizability. Populations and scanners shift over time and the biggest issue in clinical AI is the changing data distribution, such as data acquired at different times at different institution
11.
by 256lie 5y ago
Also publications are not what determines if AI get deployed in clinical practice. That's the job of the FDA and million of dollars spent on validation like clinical trials and quality management systems.
12.
by 256lie 5y ago
It's known as automation bias and a problem in pilots as well as doctors. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7651899/
13.
by 256lie 5y ago
Meta analysis is more common in medical journals than computer science conferences. The evaluation of AI medical devices is determined by regulatory agencies like the FDA.
14.
by 256lie 5y ago
Additionally, breast cancer screening is a high-volume and low-prevalence task and CAD applications has been developed for decades (although not with the performance of latest CNN algorithms).
15.
by 256lie 5y ago
Would be interesting to see the time advantage. Mammography is high-volume and low-prevalence task with standards such as BI-RADS. While AI will not replace radiologists, breast cancer screening is a prime application to assist with radiolo
16.
by 256lie 5y ago
This article is shallow and generic. Replace "software" with any other business activity and the strength of argument remains the same.
17.
by 256lie 5y ago
They mention a GAN which is a generative model and currently no good measure of evaluation (beside metrics like Fréchet inception distance). The article only mentions a qualitative comparison but no incorporation of causal / physic bas
18.
by 256lie 5y ago
Watson missed the DL train and IBM should have partnered with a company that had experience in getting medical devices through the FDA (like MSFT are doing with Nuance).
19.
by 256lie 5y ago
There are healthcare startups around fraud detection, reducing no-shows, telemedicine, drug discovery, and patient triage. Just radiology alone is prime for ML due to existing digital infrastructure and clinical use cases. The trend in FDA
20.
by 256lie 6y ago
Is DNA intelligent? What about virus? Ants? Dolphins? A corporation?
21.
by 256lie 6y ago
How much more of a human are you than a radio? A computer?
22.
by 256lie 6y ago
That presumes technical people believe in AGI. I would think more ML researchers don't so just avoid the term "AI".
23.
by 256lie 6y ago
What would a principled reason for association look like beyond mere convention? Language is used by different groups to mean different things. Machine learning, logic, control, robotics, linguistics, and cognitive science were publishing i
24.
by 256lie 6y ago
Is that a bad thing? Tesla's consumer base is not large or diverse enough to be representative of the national economy or population.
25.
by 256lie 6y ago
Humans can do a lot of other things such as hear car horns, reason about driver behavior, interpret road signs, and understand casuality of driving off a cliff.
26.
by 256lie 6y ago
Go can be simulated and self-trained. Self driving cars are robots interacting in the real world with multiple, stochastic agents. Real time hardware systems that are reliable and commercially viable will also be extremely difficult.
27.
by 256lie 6y ago
Not processing power but a better world model of physics and reasoning. Humans can understand other drivers intentions and when laws can be bent based on the context to avoid dangerous situations.
28.
by 256lie 6y ago
It is unlikely to me that any company will "win" the self-driving race in this decade.
29.
by 256lie 6y ago
Humans can do a lot of things computers cannot do. Computers can do a lot of things humans cannot do.
30.
by 256lie 6y ago
Why do you imagine that? When will they fix them all?

More ›