Panel Discussion on Artificial Intelligence
This is the AI panel discussion content post. The panel was sponsored by RSM.
George Casey - Panel Moderator, Principal at RSM
Larry Fulton - Professor of Data Science at Texas State University
Dale Sayers - Cloud Solution Architect at Microsoft
The discussion centered around the current use and best practices of ML/AI in industry. Here are some of the key takeaways and my thoughts from the panel discussion. This is also based on follow-up conversations with various members of the audience.
The short answer is Yes. But, this is no different than other industrial and technological revolutions.
Long Answer: AI/ML inputs will be a complementary addition to human expertise. In many aspects, this will bring SME and contextual knowledge to the fore i.e. who can ask the right insightful questions, who can implement the additional insights provided by this technology, etc.
There will be a range of tasks and processes where humans and machines (AI/ML) will collaborate in new ways. I also highly recommend this book aptly titled 'Human + Machine' where this topic is covered in great detail.
Picture for the post: I had a recent work trip to Gdansk, Poland. Here is a picture of the waterfront in Gdansk. World War II started in Gdansk - an interesting fact that I learned.
George Casey - Panel Moderator, Principal at RSM
Larry Fulton - Professor of Data Science at Texas State University
Dale Sayers - Cloud Solution Architect at Microsoft
The discussion centered around the current use and best practices of ML/AI in industry. Here are some of the key takeaways and my thoughts from the panel discussion. This is also based on follow-up conversations with various members of the audience.
1. AI/ML application, implementation is still low and unclear to most organizations
Based on my interaction with people at this conference, the companies present were closer to the beginning of the AI/ML journey. To be honest, there were not a lot of pointed questions from the audience. The discussions were more around specific software tools - ML/AI is not a piece of software or a toolset. The application of AI/ML has to align with the strategies of each organization and will have to turn into a competitive advantage.
Whatever the competitive advantage of an organization is, AI/ML will put it on steroids (make it better and faster!)
Whatever the competitive advantage of an organization is, AI/ML will put it on steroids (make it better and faster!)
2. Invest in Data Engineering
"How to get started on the ML/AI journey?" - this was a FAQ.
My answer to this is to invest in your data storage and engineering. Data Science is built on data. The Science and the Art/Application follows after a critical mass of 'clean' data is collected.
Data can be both expensive and cheap to obtain. That's a contradictory statement - I will explain that in more nuance here.
The expensive argument: It will be expensive to collect targeted data about a particular customer base, location, industry, etc. This data set needs to be bought or plans put in place for collection with employees or contractual workers. This can also include investment in 'on-premise' data systems or using on-demand data cloud-based data storage (AWS, Google Cloud, Microsoft Azure etc.). This is data specific to an organization's operations like transactional data, orders, customer information, supplier info, etc.
Also, look at Data Collection Outsourcing in an earlier blog post.
Also, look at Data Collection Outsourcing in an earlier blog post.
The cheap argument: Outside of the organization's internal data as described above, there will be many other complementary data that will enable the AI/ML journey. Traffic (for example, Google Maps API) and weather (eg. weather.com data from IBM) data can be very useful for a supply chain/logistics/trucking company. Currency exchange rates and financial market data (eg. Yahoo Finance) can be essential for an investment trading company. This type of generic complementary data is readily available online close to real-time via API's to enable other organizations. This complementary data is relatively cheap compared to the value that can be created for specific analyses and decisions that will be enabled by this.
3. AI/ML Embedded Processes
AI/ML at its very core is to come up with a prediction in most cases. This prediction, in turn, will fill in some missing information or correct existing information. The availability of more complete and correct information will enable better decisions leading to more value.
For the above to happen, AI needs to be part of a task/process/job enabling the organization strategy.
I cannot stress this enough.
Prediction Machines is a great book to get more information on how to think about the economics of AI/ML and the process thinking aspect is discussed in a lot of good detail. The authors have also posted a few YouTube video links to presentations and discussions on the book content.
For the above to happen, AI needs to be part of a task/process/job enabling the organization strategy.
I cannot stress this enough.
Prediction Machines is a great book to get more information on how to think about the economics of AI/ML and the process thinking aspect is discussed in a lot of good detail. The authors have also posted a few YouTube video links to presentations and discussions on the book content.
4. Focus on Quick Wins, Value Creation
AI/ML projects do NOT have to be large and all-consuming. Focus on quick wins and value creation. This becomes significantly more important in a larger organization where such positive outcomes can influence the share of budget resources.
In my experience, value creation will be a little trickle to start with. When embedded in processes (#3), the scale of the process and closed-loop feedback generated will scale grow that value creation.
The FAQ here seems to be the following, 'Will AI/ML take over some human jobs?'In my experience, value creation will be a little trickle to start with. When embedded in processes (#3), the scale of the process and closed-loop feedback generated will scale grow that value creation.
5. Humans and Machines
The short answer is Yes. But, this is no different than other industrial and technological revolutions.
Long Answer: AI/ML inputs will be a complementary addition to human expertise. In many aspects, this will bring SME and contextual knowledge to the fore i.e. who can ask the right insightful questions, who can implement the additional insights provided by this technology, etc.
There will be a range of tasks and processes where humans and machines (AI/ML) will collaborate in new ways. I also highly recommend this book aptly titled 'Human + Machine' where this topic is covered in great detail.
Picture for the post: I had a recent work trip to Gdansk, Poland. Here is a picture of the waterfront in Gdansk. World War II started in Gdansk - an interesting fact that I learned.
This increases the home edge, however generally reduces net income for the casino as gamers expertise less "play time" on the machine, busting out of their buy-in at an earlier level. Casinos that cut back paytables generally have to extend promotions to compensate 카지노 사이트 and attract prospects. Progressive jackpots are one of the better methods to advantage play video poker in markets the place there are not any full pay machines. Casinos in distant areas with little or no competition typically supply progressive jackpots on a royal flush to entice gamers to give motion to machines would possibly be} otherwise give poor returns. These don't all the time hit when they need to} and might run up properly into positive expectation, especially when mixed with different promotions.
ReplyDelete