AI & ML in Hospitality and Healthcare

Schedule:

6 PM - Registration

6:30 PM - Matthew Denesuk, SVP, Data Analytics & Artificial Intelligence at Royal Caribbean Cruises LTD.

Cruising: an Ideal Crucible for Leading Edge Enterprise AI

Enterprise AI garners little attention in the media today, but it’s potential economic impact dwarfs that of all other areas or AI combined. And it’s surprising to many that leading edge innovations in Enterprise AI will likely emerge from the cruise industry. Why? A cruise company is a unique operation. It is composed of a set of semi-independent, floating, self-contained cities, each coupled to an onshore, full-spectrum business operation. And data are being collected everywhere, to improve the all dimensions of the guest experience, and to make the operation run more effectively and more safely.

This presentation will describe how Enterprise AI is much broader, more complex, and fundamentally different from the more mature and narrower areas of AI such as NLP and computer vision. We will describe the different approaches and skills needed to be successful in Enterprise AI, and how firms are beginning to apply these. We will also talk about specific applications in the cruise industry, which cover such areas as hotel operations, industrial IoT & predictive maintenance, pricing & revenue management, social media tracking & inference, promotions optimization, supply chain & logistics, and beyond.

We will also discuss the skills and career pathways in Enterprise AI, and how to best gain entry into this emerging, world-changing discipline.

7 PM - Douglas Pestana, Manager of Data Science and Analytics at LifeExtension

Automated Forecasting and Alerts using RemixAutoML and Telegram

If your company has the following symptoms when it comes to KPI forecasting, then you maybe you need to look into automated forecasting:
• Ugly Excel spreadsheets with multiple tabs that are difficult to navigate and tedious to reverse engineer
• Too much manual and human intervention going into the forecasts, no documentation on processes or methodology, and higher-ups adding bias to the forecasts
• Lack of data science or data analyst personnel to create statistical forecasts or the data science and analytics team has low bandwidth to take on new projects
• Your current forecasts are costing you money because they’re inaccurate and the executives are furious with your team for sloppy forecast accuracy

Forecasting is a simple business problem to solve, but many companies aren’t doing it well. In fact, 20% of Amazon’s North American retail revenue can be attributed to customers buying from Amazon because its competitors lacked accurate demand forecasts. Accurate forecasts are needed in almost any vertical such as retail, eCommerce, nutriceuticals, health care, hospitality and tourism, call centers, etc. It becomes especially challenging if several have to be created all-at-once, timely and accurately.

Automated forecasting is the process of automating data wrangling and data preparation of your time series data, splitting the data into training and holdout data, training several different time series and machine learning models, testing each of those models onto a holdout data set to measure its accuracy, then choosing the most accurate model and re-fitting on the entire data set to create a forecast over a specified time horizon. This could typically take several steps and hundreds of lines of code, but RemixAutoML can do this type of automated forecasting in a single line of code.

This talk will be a concise, to-the-point business use case on how to create automated forecasts using AutoTS (Automated Time Series) and AutoCatBoostCARMA (Calendar Autoregressive Moving Average via CatBoost gradient boosting) using the open source RemixAutoML package and then send forecast alerts to your phone and desktop using Telegram, an open source encrypted messaging app.

7:30 PM – Pizza, Networking & Beer