goupaz.com
goupaz.com copied to clipboard
goupaz
[06272021] - Chapter04 Session07 - Data collection and processing for AI by Avi Gopal
What is your suggestion, how to collect the medical info for AI processing?
How Discreet Mathematics, Analytics, NLP, Statistics, Deep Learning, ML etc stuffs are directly linked with AI or required to learn AI ?
How software engineers can easily move to AI field? Do you have some fast track road map suggestion for it?
- Resources for good AI projects we can work on ?
- Apart from kaggle, sites we can check for getting data ?
What is complete workflow of data processing?
And resources to learn AiOps and MLOps?
What is glue catalog?
Is there any AI based database dedicated to videos and images?
Another question on Metabob . Metabob is very useful since I used. It supports only python.
But do you have any plan to get add support for other programming languages in future ?
- Resources for good AI projects we can work on ?
- Apart from kaggle, sites we can check for getting data ?
https://github.com/awesomedata/apd-core/tree/master/core
What is your suggestion, how to collect the medical info for AI processing?
My presentation on HIPAA-compliant data processing for the detection of clinical depression is at http://v.gd/ben20200116
It's from January 2020, and includes a link to a technically much more detailed paper
What is the process to be followed while implementing AI typically (in a company)? How long does this process usually take? How do you think what is the impact of AI in the near future and would it be able to replace some jobs? What recommendations can you give to the CEO or to the individual who has not used AI yet?
What is your suggestion, how to collect the medical info for AI processing?
My presentation on HIPAA-compliant data processing for the detection of clinical depression is at https://v.gd/20200116
It's from January 2020, and includes a link to a technically much more detailed paper
It says link doesn't exist.
What is your suggestion, how to collect the medical info for AI processing?
My presentation on HIPAA-compliant data processing for the detection of clinical depression is at https://v.gd/20200116 It's from January 2020, and includes a link to a technically much more detailed paper
It says link doesn't exist.
Sorry, the link should be http://v.gd/ben20200116
What is the process to be followed while implementing AI typically (in a company)?
Typically like this:
- Decide your goal, get funding from investors
- Hire Researchers & Engineers to make an architecture for training and deploying the model
- Obtain the data for training the model (in parallel with previous step)
- Train, Test, Deploy, Repeat.
There is also marketing, UX testing, legal stuff, etc.
How long does this process usually take?
It's iterative, but between initial investment and product can be 2-6 years. Most successful I've been involved with was a spinout from SRI so initial funding was 1994, first commercial product was around 1997, IPO 2000, bought by competitor in 2005. But timing is very important: if this company had started in 1997 instead of 1994, they might have failed because in 2001 investors shunned many similar companies. From October 2000 to April 2001 our stock went from 102/share to 35/share (and I think it went lower).
How do you think what is the impact of AI in the near future and would it be able to replace some jobs?
It already has replaced jobs. For example call centers, customer service, job recruiting (scanning resumes)
What recommendations can you give to the CEO or to the individual who has not used AI yet?
Be sure AI will be useful to achieve your goal. Get good advisors.
I've seen too many companies fail because CEO hired some family friend, morale fell, everyone became unproductive. But I've also seen one whose CEO hired an excellent Lawyer, then gave that Lawyer the CEO position! It worked out well.