GetReady4AI

infrastructure cost and lack of trained staff serve as the big roadblocks right now

.@plburris The new innovation cocktail = Data + #AI + Cloud scale so constraints are primarily data quality & accessibility (i.e. no stovepipes), AI skills & resources (talent)...access to cloud scale is pretty much table stakes by now but the other two are significant barriers

AI is a new technology, and needs to be tested first in relatively bounded applications

@dfloyer agreed. Organizations are at various stages of the AI adoption journey, but almost all start with some sort of sandboxed environment at first.

@dfloyer Starting with a really well defined use case, usually a piece of software is so important.

It's actually a set of technologies with long vintage. What's new about AI is the extent to which it's been deployed into operational business applications and infrastructure, and the extent to which it's the core focus of developers for all types of disruptive apps

@dvellante Access to automation throughout the AI pipeline is as critical as cloud-scale. It's all about speeding up the DevOps cycle for these AI-driven apps, bringing greater consistency, industrializing the process 24x7 in enterprise development/IT shops.

Government / aerospace / defense use cases

Automating data workflows

Healthcare is ripe for disruption - but really hasn't been disrupted yet - really interested to see how machine intelligence applies in the near-to-mid term there #AI

we see a lot of interest in applications that can use image or video processing to improve business outcomes. For e.g. using pictures of components as they come into a manufacturing plant to scan for defects before they are used

Also interested in: 1/ Will machines make better diagnoses than Docs? 2/ Will driving your own car become the exception? 3/ Will large retail stores mostly disappear? 4/ What's the future of warfare and what role will #AI play?

improving operational efficiency, reducing risk, and improving security have been leading the way

@MikeLeone_IT +1 on the operational efficiency. Freeing up resources to do more creative and higher value activities.

Personal favorite AI use cases are in healthcare. Already useful in object classification (e.g. tumor detection), coming soon for NLP transcription of EHRs, and eventually full NLP on patient oral responses, prediction of illness onset, better treatment plan suggestions, and more

.@JayBoisseau One of the most incredible uses in healthcare I’ve seen @TGen is delivering precision medicine based on the Human Genome just inspiring. #AI will further transform this field. #IWork4Dell
https://www.youtube....

@NickBrackney Love the @TGen folks, know them well.. and talking to them about next steps in using #AI :-)

Thanks for the great partnership

@jamonascone Machine Learning Knowledge Center featuring the latest developments in artificial intelligence, machine learning and deep learning

Thanks for the partnership

This key point seems to get overlooked WAY too many times! It was a point that we tried to stress on my last few international consulting assignments. Especially with the GDPR rules in place, the data has to be accurate.

Yes Keith... #AI doesn't solve GIGO problem #DataQualityMatters

Yes. here's my discussion of the quality conundrum surrounding this data in an AI/ML development context, from a few years ago in another life. Data quality for social media analytics: sentiment or sediment? http://www.ibmbigdat...

Data Preparation can be 2/3rds or more of the total effort/time in an AI/Deep Learning project! Need to account for that!

.@plburris we use AI daily in our data ingestion system - curating video content and identifying the best "gems"

Most of my current clients are still working on understanding what it means in their organizations. There are a lot of things being labeled AI that are not.

Many enterprises are at the start of their AI adoption. What that means is that they've adopted machine learning as an evolution of their advanced analytics practices, and are using it for focused apps: chatbots, intelligent search, security incident ad event mgt, etc

I'd put it in the category of NLP - but humans are still critical to our workflow - so I'd say we're still in the very early days

nearly 20% of organizations are depending on AI/ML to deliver significant measurable business outcomes immediately

While AI is a strategic priority for most CIOs, the trick is to offload data scientists of any mundane non-data science tasks

@jameskobielus we're seeing the same. There's also a lot of interest in image and video recognition capabilities in multiple verticals.

@jon_siegal The strategic priority should be to build data science teams of specialists for the various functions--statistical modeling, data engineering, programming, etc.---necessary to drive the entire AI DevOps workflow.

@broomio Yes. We're seeing plenty of activity in AI-driven computer vision. Check out my recent #SiliconANGLE piece re this trend: https://siliconangle...