1/n) #automl automated machine learning is shifting a field of #ml from #knowledge centric to compute centric.
It also providing much needed abstraction layer for #businessgrowth and #product folks to directly leverage power of ml without the need of Experts.
2/n) Also from business point of view, it shifts #MachineLearning from fixed cost (Experts, systems, data) to variable cost (pay per use automl). This will result in death nail of many ml teams across Enterprise.
3/n) On top of #automl, if you can integrate #DevOps then you will get modified version of #mlops which means a single engineer with sufficient knowledge and help of cloud can replace team of engineers and researchers. #DigitalTransformation
4) Now scary and self reflection question...
What will the folks working on derivative #ML work will do?
Another up skilling 😕?
New technology?
It's time to think deeper and answer the why question for your current work and #futureofwork.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Until you appreciate what you currently have, more won’t make your life better.
2. It’s Never As Bad As You Think It Will Be
The problem with dread & fear is that it holds people back from taking on big challenges. What you will find — no matter how big or small the challenge — is that u will adapt to it.When u consciously adapt to enormous stress, u evolve.
What #AI lacks, humans can fill in.
What humans lack, who will fill in.
Dual standards of our society and human thinking. #Ethics, #fairness and other values are regulated for AI but same values for humans are not regulated.
Time to do introspection.
If humans are unable to grasp the ethics, morale and fairness values due to deep diversity of humanity, how can we ensure ethical frameworks created for #AI will be universal.
Humans have failed to uphold the values of #ethic across history and geography. Humans have used the technology to gain control and become superior. The weaponisation of #ai is inevitable. We have seen parallel cases from field of #biotech and specific case of #crispr
#antitrusthearing
Security and safety of consumer, product, partners and Algorithm is next set of questions. Again no satisfactory answers. #Algorithms#AI#security
#antitrusthearing very interesting question, how will you ensure that biases of your employees are getting in to the algorithm?
In fact research has proven that algorithms are learning not from the data but the way data was labelled and annotated. #ai#bias
#100daysoflearning#psychology
Day 18 update
Completed reading from Robert Cialdini
Started another reading from chapter 2 of Mayer’s book on Social Psychology, completed 5 pages
#100daysoflearning#psychology
Day 18 update
Key Learning
- spotlight effect is experienced when we think people are paying more attention to us then needed.
- we also suffer from illusion of transparency that our emotions are easily detectable.
#100daysoflearning#psychology
Day 19-20 update
Completed up to page 12/chapter 2 from Social psychology book by Mayers
Key Learning
- We overestimate the visibility of our social blunders and public mental slipups
- At center of our world is our sense of self
#QuantumComputing#quantum#technology#India Panel is packed with eminent and leading members from Science and Academia.
Apoorva Patel (IISc, Bangalore)
R. P. Singh (PRL, Ahmedabad)
Umakant Rapol (IISER, Pune)
Anil Prabhakar (IIT Madras)