What would machine learning look like if you mixed in DevOps? Wonder no more, we lift the lid on MLOps

Achieving production-level governance with machine-learning projects currently presents unique challenges. A new space of tools and practices is emerging under the name MLOps. The space is analogous to DevOps but tailored to the practices and workflows of machine learning. Why MLOps is Needed Machine learning models make …

[amanfromMars 1]

C42 Quantum Communication Control Systems .... AI@ITsWork

Machine learning models make predictions for new data based on the data they have been trained on.

Oh? That is bound to be retarded.

Surely the likes of SMARTR* Machine Learning Models make predictions with new data based on the metadata/intelligence and information they are discovering and training on/uncovering and deploying ....... should making predictions be one of their things rather than just a human trait and confection?

* .... Systems Monitoring Analysis Reports Tracking Resources.

As the field evolves we’ll likely see greater standardisation,...

And, of course, ever greater Alien Specialisation too ..... to ensure guaranteed overwhelming advantage in established fields venturing forth deeper and darker into a whole host of novel environments ..... is always a current solution to abiding problems which can also be a current problem to abiding solutions too and therefore something else altogether quite different to deal with and macro micromanage.

Such is quite normal though and just the nature of the beast unleashed. And there be some surely, who may or may not be more than just a few, who can and do recognise and enjoy and employ it, IT and AI in Operations which Attack Human Achilles Heels.

[VibhorTyagi]

I Engineer AI, ML, DL...

There is a reason that Machine Learning become so frustrating, I will tell you as an engineer. AI are generally formed on the basis of Deep learning, as such they rely on a huge amount of trial and error. DevOps, MLOps, etc become important because they create an aura of synergy between the other AI engineers. Without this, there is little to no coordination, management and planning between the ideal of work, and the active process wherein we engineer ai on a daily basis.

[macjules]

Re: I Engineer AI, ML, DL...

Getting sick to death of what I am supposed to call everyone. My DevOps team now calls themselves DevSecOps and this morning proudly announced they were adding NoOps and GitOps and I suppose they will add MLOps.

Soon it might be RedOps .. short for Redundant Operators.

[Mike 137]

All very up beat but ...

Considering the widely evident erratic behaviour of DevOps managed fixed logic systems in current use, there's a very high probability of complete chaos ensuing. If you don't know in advance what the answers are likely to be and don't know how they are arrived at (typical legitimate characteristics of black box investigative AI) the one thing you have left to rely on is stability of the underlying mechanisms of decision making. Make these a variable at your peril. And quite apart from this, the quality of rapid turn round code production is generally so poor (witness bug fixes for bug fixes for bug fixes from many major vendors) that the more you tinker the more hazardous it gets.

[amanfromMars 1]

Re: All very up beat but ...

If you don't know in advance what the answers are likely to be and don't know how they are arrived at (typical legitimate characteristics of black box investigative AI) the one thing you have left to rely on is stability of the underlying mechanisms of decision making. ..... Mike 137

But, Mike 137, is there always in human centric circles an inherent systemic petrification and powerful endemic command and control problem whenever one knows in advance of how things are most likely to be [because one has been so informed and advised] but one isn't privy to how such things are actually to be arrived at ...... because that particular information is of a peculiar intelligence which is proprietary intellectual property, extremely valuable and easily misused and abused for almighty destruction and/or disruption?

Do you think AI would be bothered by that or even think to bother itself with that?

[AMBxx]

I predict (without any AI)

That we're overdue a return to traditional Software Engineering with Waterfall project mangement.

Yes, projects will be late and/or over budget but at least the stuff will work as per requirements spec.