Details on Microsoft Singularity for AI Workflow

Worth a look, I will.  Note indication of scheduling AI workloads.

Microsoft goes public with details on its 'Singularity' AI infrastructure service

Microsoft's Azure and Research teams are working together on the 'Singularity' AI infrastructure service.

By Mary Jo Foley

Posted in All About Microsoft on February 21, 2022 | Topic: AI & Robotics

Microsoft's Azure and Research teams are working together to build a new AI infrastructure service, codenamed "Singularity." The Singularity team is working to build what Microsoft describes in some of its job postings for the group as "a new AI platform service ground-up from scratch that will become a major driver for AI, both inside Microsoft and outside."

A group of those working on the project have published a paper entitled "Singularity: Planet-Scale, Preemptible and Elastic Scheduling of AI Workloads,"  which provides technical details about the Singularity effort. The Singularity service is about providing data scientists and AI practitioners with a way to build, scale, experiment and iterate on their models on a Microsoft-provided distributed infrastructure service built specifically for AI.

Authors listed on the newly published paper include Azure Chief Technical Officer Mark Russinovich; Partner Architect Rimma Nehme, who worked on Azure Cosmos DB until moving to Azure to work on AI and deep learning in 2019; and Technical Fellow Dharma Shukla. From that paper:

"At the heart of Singularity is a novel, workload-aware scheduler that can transparently preempt and elastically scale deep learning workloads to drive high utilization without impacting their correctness or performance, across a global fleet of accelerators (e.g., GPUs, FPGAs)."  (Details at her ZDnet site linked to above) .... ' 

Walgreens Automates Vaccine Scheduling with AI Bot

Interesting example here of integrating basic interactions and more complex data based decisions.   

Walgreens automates COVID-19 vaccine scheduling     By Dan Berthiaume - 03/10/2021 in Chainstorage

Walgreens is extending access to COVID-19 vaccine information and scheduling with an artificial intelligence (AI)-based bot.

The drugstore giant, which is receiving a weekly federal allocation of more than 480,000 vaccine doses for distribution across 26 U.S. states and jurisdictions, is now deploying an intelligent conversational voicebot from Nuance Communications. The voicebot will help customers schedule their COVID-19 vaccine appointments 24 hours a day, seven days a week, beyond Walgreens’ existing web-based vaccine portal.

Walgreens customers can now call 1-800-Walgreens or a Walgreens store and speak in English or Spanish to the Nuance voicebot to get answers to COVID-19 questions, confirm eligibility according to applicable guidelines, and schedule their vaccine appointments wherever vaccines are available. The Nuance solution also sends the customer an SMS text to confirm appointments after the call. ... '

Chatbots for Setting up and Maintaining Meetings

For some time have been using assistants on Siri, Alexa and Google to manage meetings.   And several times its been suggested that this kind of very common activity should be done by a meeting chatbot.  I agreed, but quickly found it was far less than trivial.  Sounds simple, but the complexity assistant logic among multiple human and machine agents is hard.    Which further suggests that further extending other multi agent intelligent conversations are not easy either.

Read this article for a good intro:

However trivial it may sound, creating an AI program to successfully schedule meetings is a monstrously difficult challenge. But the employees of X.ai are some of the most dedicated nerds you’ll ever meet. 

Author:  John H. Richardson in Wired. 
AI Chatbots try to schedule Meetings without Enraging us.

Clara Schedules Meetings for You

Another example of AI assistance.  Via Jim Spohrer.  The improvement of Email and communications is a huge opportunity.

If You CC This Robot, It Will Schedule Meetings For You  .... Clara may be the first AI built around you and your email.  

Cortana. Siri. Google Now. Facebook M. This quartet of artificial intelligence platforms, made by Microsoft, Apple, Google, and Facebook, respectively, can find you a place to eat dinner, look up the score of the football game, or even crack a decent joke.

But none promise the same specific functionality of Clara. Just CC Clara onto any email, and it’ll take over the conversation just like a real personal assistant would. It can juggle lunch meetings and your kids soccer practice, to go through the tedious "does 3pm on thursday work for you?" conversation to schedule an appointment. ....  " 

Baseball Scheduling

Scheduling complex sports events, with many restrictions and preferences, is a known difficult problem to automate.  New work at Johns Hopkins addresses the problem.  " ... A team of Johns Hopkins University researchers has developed a system that uses thousands of lines of computer code to satisfy all of a league's scheduling rules, and as many of the teams' requests and preferences as possible.

 .... The system, developed by associate research scientist Anton Dahbura and professor Donniell Fishkind, allows 10,000 schedule limitations and even more factors to be fed into a supercomputer, which produces a workable schedule. The methodology employs combinatorial optimization, which is the concept that there should be at least one schedule that optimally satisfies every rule and team request, via a combination of integers. .... " 

Placebo App on Smartphones

In Mashable:   The placebo effect has been known for a long time.  Is there a way to use interaction with a smartphone to produce a similar effect?  And can that effect be more readily tailored with a smartphone?  All interesting questions.  " ... Here's how it works: You begin the experience by choosing which lifestyle aspect you'd like to change — say, quitting smoking or decreasing stress — before scheduling an alarm-like reminder to "take it" each day. Then, you can personalize it further by choosing exactly what you'll be taking (it doesn't need to be a picture of a pill). The whole idea is to create a comfortable "happy place" to achieve the proper effect. With the right mindset, the group says, the act of routinely pressing buttons and watching your smartphone's screen will be equivalent to physically swallowing a sugar pill. ... " 

Optimizing Business Process

Rexamining a company, Charlotte Software Systems.  We had worked with their CEO Kevin Kostuik, when he and other people there worked with former component BiosGroup and then NuTech.  The company was bought by Netezza, which was then folded into IBM.    Charlotte is now an independent company.  We used some of their methods to address genetic and optimization solutions to such problems as project portfolio analysis, supply chain design and product development investment.   Worth a look.  Will  report more here as I can.  Pass me any experiences you have had.

   " ... Charlotte Software Systems specializes in cutting-edge optimization software to support enhanced tactical and strategic planning. Our systems can help optimize a number of complex business processes from large scale resource scheduling to strategic asset allocation. ...The foundation for our solutions is a powerful modeling and optimization platform upon which we capture the intricate details of your specific business process. Our solutions are delivered with attractive user interfaces which allow business specialists an intuitive way to manage their process. The resulting application is precisely tuned to fit your business while leveraging a proven platform of advanced optimization technologies. ... " 

Business Analytics, Past and Future

I have been a long time practitioner of business analytics, the use of quantitative methods to improve business processes. Starting at the Defense Department, where we built military simulation models, and ending up at P&G.  My first project there, when I arrived in 1977, was to improve warehouse efficiency using IBM's mathematical programming system: MPSX.  That followed with using the same mathematical optimization approaches for scheduling, supply chain siting and executive decision making.

So I was happy to see in the July/August Issue of Analytics Magazine an article by Arnold Greenland on the history and current state of business analytics at IBM.    This allows me to reflect on the growth and impact of analytical methods at IBM and elsewhere.  Since then the original MPSX package has disappeared, and has been replaced by the acquired Ilog/Cplex, which goes far beyond the original package, adding nonlinear methods as well.

Shortly after being introduced to MPSX, in 1980, we addressed the analysis of unstructured data, typically consumer comments in unstructured text, using recently developed methods called  'Content Analysis'.  These permitted the semantic analysis of multiple human languages.  An early attempt to look at and understand unstructured 'Big Data'.

During all of this time we also used a number of statistical methods throughout the enterprise to explore and improve systems.  SPSS and SAS were in frequent use.  A package called Clementine allowed us to use advanced logical methods, like artificial neural nets, to implement what were essentially statistical methods, to store and implement specific decision rules.   These methods could then be inserted in both software and hardware processes.  Clementine also permitted the structural exploration of the decision process.  Ultimately Clementine was acquired by SPSS, and were eventually absorbed into Modeler.

During the 1990's, a heady time for artificial intelligence, we implemented expert systems using a now defunct language called M1.  Which were successful for a number of complex industrial management processes.  Some of the same capabilities can be seen in JRules.  That work has been extended into recent AI explorations like that of IBM's Watson.

What is further interesting now is that many of these methods are now available to the small and medium sized business.  We live in a time where the tools are available,  just go out and use them.

Timed Automata and Scheduling Resources

A largely technical article in the CACM that reminded me of the complexity of scheduling and resourcing activities in real time. 

" ... The problems of time-dependent behavior in general, and dynamic resource allocation in particular, pervade many aspects of modern life. Prominent examples range from reliability and efficient use of communication resources in a telecommunication network to the allocation of tracks in a continental railway network, from scheduling the usage of computational resources on a chip for durations of nano-seconds to the weekly, monthly, or longer-range reactive planning in a factory or a supply chain ..."

The Hubble and Enterprise Planning

I have been reading Eric J Chaisson's Book: The Hubble Wars, a detailed view of the launch and operation of the Hubble Space Telescope. My background is in astrophysics, so the details of goals and scheduling ring true. My later career dealt with optimizing the large scale enterprise so some of Chaisson's descriptions of the Hubbell team using AI techniques to direct performance remind me of those heady intelligence times in the late 80s and early 90s. The shuttle is over now, but the Hubble is still there and there is still much need for doing things quickly and well in the enterprise. He writes about an environment written for the Hubble: . "Science Planning Intelligent Knowledge Environment. (SPIKE) a neural net, Lisp language based algorithm that includes the latest and sundry kind of smart software, expert systems and artificial intelligence techniques..... " We could use more of that kind of advanced help today. I am now taking closer look at what Spike did.