MakeAIReal

Recommendation engines are something I experience daily and has improved my overall user experience

Stuart Miniman12

Would you prefer to run AI in the public cloud and/or on-premises? How are you deploying (or thinking about it)? https://www.crowdchat.net/s/260ox

Yes. My answer is Yes.

Everywhere!

We've seen customers using AI across the datacenter, public cloud and now it's reaching into edge computing as well as IOT. #iamintel #MakeAIReal

Esther Baldwin

@CTOAdvisor I agree Keith the new norm is many data sources especially for edge to enterprise or cloud uses. More than one of the solutions Jeremy mentioned include that capability.

@CTOAdvisor Yes! Customer need solutions to run AI anywhere

@CTOAdvisor Is this a have your cake and eat it too response? #iamintel #MakeAIReal

Stuart Miniman

train in the cloud, process at the edge - see David Floyer's https://wikibon.com/matrix-workloads-power-real-ti...

Matrix Workloads Power Real-time Compute - Wikibon Research
End-to-end data architectures & Heterogeneous Compute Architectures will power Matrix workloads to extract all the value from huge amounts of parallel data.

Preethi Venkatesh

Additionally we have seen container/Kubernetes usage around development and even deployment, whether data center/cloud

Esther Baldwin

@Lucas_A.Wilson - and what I like about the transformation for medical imaging is that it supports the doctors making the diagnosis, helps them see things they might not have seen before AI

This is a great transformation for good that stems from the massive amount of data we have

(edited)

Esther Baldwin

to answer your on cloud or on prem question.Speaking personally for all the usages already mentioned for health I would prefer my data was on prem. #iamintel #MakeAIReal

Lucas A. Wilson

.@EstherBaldwin AI is an ideal instant second opinion! And by providing the physician with object boundary detection and confidence levels, it can help make sure subtle but important features are not missed.

Paul Calleja8

By working with Dell and Intel we were able to create a world leading I/O solution from open standards hardware and open source software, This removed I/O bottlenecks from our workflow

7 Votes Vote

Truly and impressive solution!

wonderful work

While we are hearing some customer stories, would welcome commentary from Dell and Intel on the biggest hurdle faced by customers looking to adopt AI.

A big challenge we are seeing is the lack of AI skills in house, there is also a large gap between POC and full scale production

Are others seeing this challenge too?

Challenges for Deep learning include security, especially for cases like the video shows where they involve massive proprietary datasets. Peoples personal information and it is highly regulated in the USA and Europe and elsewhere. #iamintel #makeaireal

(edited)

Other challenges include the tendency for IT managers to "Silo" their compute for deep learning which doesn't allow them to take advantage of all of their compute resources and can cause a bottleneck getting the data in and out of that silo. #iamintel #MakeAIreal

Lucas A. Wilson

Not just AI skills. Getting AI to full production means having the AI/data science skills, software engineering and deployment skills, and IT skills to make sure that those different groups have the hardware and software they need to #makeAIreal

3 Votes Vote

Chad Martin

Other key challenges are data pre-processing, model development, and scaling from development to production. Each of these can derail a project. #MakeAIReal #IamIntel

Janet Morss6

We can't do it alone! Thanks Paul Calleja and team for partnering with us on the Data Accelerator!

4 Votes Vote

Stuart Miniman

Paul shared some great data in yesterday's chat https://www.crowdchat.net/chat/c3BvdF9vYmpfMzEwNg=... - he also won "CrowdCaptain" 😁

[LIVE CHAT] Enabling Real AI with Dell
Learn how Dell Technologies is making AI Real for customers through partnerships with VMware

This is such a great story of performance and really being able to #MakeAIReal

https://www.dellemc.com/resources/en-us/asset/customer-profiles-case-studies/products/ready-solutions/dell-mcgill-case-study.pdf

Janet Morss6

This team of researchers is so inspiring! So cool! https://www.dellemc.com/resources/en-us/asset/cust...

4 Votes Vote

Lucas A. Wilson

such a fun project! So glad the Dell Technologies HPC & AI Innovation Lab was able to be a part of this research!

Stuart Miniman5

This next question builds on what @CTOAdvisor said about having a dashboard for #CIOs. How would you like to be able to use AI ideally? https://www.crowdchat.net/s/260ol

This combines a couple of topics. First, there's the ongoing development and operations of AI. Enterprises don't have the talent. Solutions need to look similar in maturity to modern content management platforms. They need to be extensible and focused on displaying outcomes.

3 Votes Vote

@CTOAdvisor End users need to be able to serve multiple business units with the same underlying data. As infrastructure professionals, that's the goal of AI platforms. A platform to create applications and insights across multiple domains.

With Dell Ready Made Solutions for AI powered by Intel, you can bridge the gap in skills and accelerate AI deployments with automated functions and feature engineering using opensource tools and libraries. #iamintel #MakeAIReal

Paul Calleja5

The Data Accelerator is able to completely remove I/O Slowdown from an application, meaning engineers and scientist can now imagen completely new data intensive workflows in both AI and simulation

5 Votes Vote

Stuart Miniman5

@shawnamr mentioned that #AI is impacting every industry. What applications are you using with AI or thinking of using? I've seen many customers making impactful solutions in healthcare, food production, and community services leveraging AI. https://www.crowdchat.net/s/860nr

In my experience Pharma organizations use AI/ML to match patients to nurses. Nurses are a limited resource and ideally they'd like every patient taking a drug to talk to a nurse. San the availability or nurses, they use AI to target the must appropriate patients.

Another healthcare application I've run into is on the payer side. AI/ML is used to determine payment eligibility. There is a huge need for maturity in this space.

Going back to my space the obvious use for AI/ML is in Ops or AIOPs. Using ML to reduce the number of false negatives for security and operational alerts helps to increase application availability.

@CTOAdvisor Interesting application

@CTOAdvisor With built-in AI acceleration with Intel Deep Learning Boost, Dell Technologies is positioned to support healthcare services for such items at natural language processing, critical in patient care. #iamintel #MakeAIReal

Where I'd like to see improvement is the ability to tie operational alerts to business outcomes. A dash board CIO's could use to talk to their peers.

@CTOAdvisor How's your datacenter build going? Will you be using AI to create an autonmous datacenter? #iamintel #MakeAIReal

https://www.dellemc.com/resources/en-us/asset/white-papers/products/ready-solutions/pisa-vmware-white-paper.pdf

Janet Morss4

Thank you to Maurizio Davini and the team at the University of Pisa for their work with Dell Technologies and Intel to prove out new code, new technologies and solutions! https://www.dellemc.com/resources/en-us/asset/whit...

4 Votes Vote