Comments
Deep Learning: Make Sense of Humongous Datasets
View Comments: Newest First | Oldest First | Threaded View
Page 1 / 2   >   >>
Deep learning technique
  • 1/18/2016 12:20:07 PM
NO RATINGS

Kishore, can you tell me which deep learning technique you have been talking about: dropout, no-drop, dropconnect or gradient descent? Which technique would you consider the most effective in reducing health care costs?

Re: Better medicine
  • 11/29/2015 4:41:15 PM
NO RATINGS

I'm not confident that deep learning tech in healthcare will reduce costs. We have imaging technologies that make diagnosis more accurate, targeted nanoscale treatments with new drugs, doctors handing off routine care to nurse practitioners and physicians assistants, robotic surgery, and limits on hospital stays. Yet, costs continue to soar when each of those steps should have cut costs. What doctor or hospital is going to reduce their fees?

Re: Better medicine
  • 11/29/2015 3:50:42 PM
NO RATINGS

Predictions about human health will be accurate because these are chemical and biological processes.

I'm less confident about predicting the financial markets since these depend on human behavior. Am I the only one that finds it difficult to predict how people will act and react?

Re: Better medicine
  • 11/29/2015 12:46:42 PM
NO RATINGS

Deep learning should eventually prove to be real lifesavers for the health industry. And if the costs of doctors reading charts and numbers can be reduced, theoretically healthcare cost should be able to be reducted significantly. I wonder if deep learning of financial data may lead to some important understandings, or just demonstrate that perhaps financial markets are more random that we think?

Re: A Rethorical Look at the Potential of the Human Brain
  • 11/23/2015 2:46:21 PM
NO RATINGS

"...So I see humans and machines playing complementary roles."

 

@Kishore    No disagreement here.   I have just been noticing more and more that no one wants to think or learn anymore.   Just Google it, get an answer and not have enough intellectual background to even question the results.   It is as if we have all given up on our potential because of the almighty computer.

Sad.

Re: A Rethorical Look at the Potential of the Human Brain
  • 11/18/2015 12:37:12 AM
NO RATINGS

@Louis Watson: There are pros and cons of the human brain and that of deep learning by machines. Humans are certainly more creative, analytical and intuitive while machines process much larger volumes of data at a higher speed. So I see humans and machines playing complementary roles.

Re: Let's automate this
  • 11/18/2015 12:33:48 AM
NO RATINGS

@Terry Sweeney: It is more like rushing towards automated analytics. Humans play a role later when you try to make sense of the patterns identified. For real-time analytics, the entire process is automated. The technology is quite advanced and is especially useful in medical diagnostics, speech recognition, etc

A Rethorical Look at the Potential of the Human Brain
  • 11/17/2015 11:13:44 PM
NO RATINGS

Interesting arguments for the wonders of computing.  This argument that the human brain ( even though we do not even understand it completely) is not capable of coming to the same conclusions as say, Watson is something I am not prepared to throw the towel in on. 

Sure only a small percentage ( Einstein and the rest) have even come close to harnessing the power of one's own intellect or brain for the sake of argument.

But do we know for a fact that the brain cannot process as effectively Watson ?   Because few could ever operate it (their brain) to it's fullest potential ?

There is no doubt I would lose to Watson rather quickly in chess for instance but this result does not inherently demenish the ability of the brain to function almost as efficiently as Watson itself.

Just because we have never seen this result (often) does not mean this possibity cannot exist. 

Let's automate this
  • 11/17/2015 10:42:40 PM
NO RATINGS

Are we tiptoeing up to automated analytics with deep un/supervised learning? This makes a huge amount of sense for big data volumes where human power can't touch its subtleties, nuances or important insights.

Predicitng ear infections
  • 11/17/2015 4:02:46 AM
NO RATINGS

The idea of predicting ear infection with a mobile device connected to a supercomputer such as Watson would be a great idea.  It would help prevent hearing problems by predicting early signs of ear infections when a child is very young.

Page 1 / 2   >   >>


INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +