Carl H. Lindner College of BusinessCarl H. Lindner College of BusinessUniversity of Cincinnati

Carl H. Lindner College of Business
Machine Learning Day Graphic

Machine Learning Day

February 14, 2018  7:30 AM - 4:15 PM

Sharonville Convention Center   (directions)

Three keynote speakers will present in the morning session, and the afternoon will feature machine learning technical presentations and case studies by data scientists and analytics companies.  The event fee ($225) includes all sessions and meals.  Parking is free at the Convention Center.

The event will begin with a networking breakfast at 7:30 AM with the keynote sessions beginning at 8:00AM.

Speaker presentations (ones we are allowed to share) will be available after the event.

Register for Machine Learning Day

 

Thank You Event Sponsors

Machine Learning Day Sponsors

MORNING KEYNOTES

  1. Title TBD: Mohammad Taghi Saffar, Machine Learning Engineer, Google

  2. "Machine Learning at Kroger" : Doug Meiser, General Manager of R&D Operations Research, Kroger

  3. "The Predictability Predicament: Your Model Overlooks the Real Target": Claudia Perlich, Chief Scientist, Dstillery
    In the context of building predictive models, predictability is usually considered a blessing. After all - that is the goal... to build the model that has the highest predictive performance. The rise of 'big data' has, in fact, vastly improved our ability to predict human behavior, thanks to the introduction of much more informative features. However, in practice, the target variable is often more differentiated than accounted for in the data. For example, some customers churn (from a telecom provider) because they are moving, others because they got a better offer in the mail, and the third because their home is in a location with terrible reception. These are all positives for a model that learns to predict churn, but the predicted outcome has occurred for very different reasons. In many applications, such mixed scenarios mean the model will automatically gravitate to the one that is easiest to predict at the expense of the others. This even holds if the predictable scenario is by far less common or relevant. In the worst case, predictive models can introduce biases NOT even present in the training data. In this talk, we will cover a number of applications where this takes place: clicks on ads being performed 'intentionally' vs. 'accidentally', consumers visiting store locations vs. their phones pretending to be there, and finally customers filling out online forms vs. bots defrauding the advertising industry. In conclusion, the combination of different and highly informative features can have a significantly negative impact on the usefulness and ethics of predictive modeling.

_________________________________________________________________

AFTERNOON SESSIONS

There will be two sessions with 3 speakers in each session.  You can attend one presentation in each session.

Session 1  1:30 - 2:45 PM
  • "Random Forests and Gradient Boosting Machines in R": Brandon Greenwell, Senior Data Scientist, Ascend Innovations
    Good modeling tools should be universally applicable in classification and regression, have state-of-the-art accuracy, scale well to large data sets, and handle missing values effectively. Additionally, it would be nice for these tools to be able to automatically discover which variables are important, how they interact, and whether there are any novel cases or outliers. In this presentation, we discuss two such modeling tools: random forests and gradient boosting machines. The talk will cover a brief background of both methodologies (including decision trees) as well as various implementations of each in the R software environment for statistical computing. The pros and cons of each implementation will also be covered.
  • "Using Deep Learning in Tensorflow to Evaluate Consumers' Credit Worthiness in The Financial Industry": Kaveh Bastani, Hamed Namavari, Data Scientists, Recovery Decision Science
    Conventionally, in the financial services industry, analysts have been cross-referencing consumers’ personally identifiable data against other person-level and aggregate datasets such as Credit Bureau Reports and Census Data to analyze consumer behavior. However, the recent advances in ubiquitous digital imagery and machine vision techniques have given an extreme capability to analysts to dig deeper and extract new consumer insights that were impossible by the conventional methods. In this work, we present a method using deep learning-based computer vision to gain novel consumer insights by analyzing images of consumer's residential address.

  • “Machine Learning at Cloud Speed”: Clark MacDonald, Principal; Melissa Lacefield, Azure Cloud Solution Architect; Andrej Kyselica, Azure Cloud Solution Architect, Microsoft Azure Platform Solutions
    Learn how to leverage the latest Machine Learning technologies and improve the speed of innovation. Build and deploy ML applications in the cloud, on-premises, and at the edge. Get started by wrangling your data into shape easily and efficiently, then experiment quickly with Azure ML Studio.  Envision enterprise possibilities taking advantage of popular tools like Cognitive Toolkit, Jupyter, and Tensorflow in the Azure ML Workbench to build advanced ML models and train them locally or at large scale in the cloud.  Discuss how  real world examples of Machine Learning use cases can help your organization leverage machine learning at cloud speed.
Session 2  3:00 - 4:15 PM
  • “Price and Promotion Optimization in a Python-based Analytics Environment”: Mark Wells, VP Business Development, Opalytics
    This demonstration illustrates a use case (currently in production), that involves ML to predict sales of beverages (given price and other factors), predicting market share of competitors, and optimizing price and promotion recommendations.  A predictive analytics app and a custom optimization app are seamlessly connected with workflow within a Python advanced analytics environment in the cloud (Opalytics).  The predictions leverage the scikit-learn library.  The custom optimization app, instantly deployed as an interactive business app, uses the open source ticdat data library and Gurobi optimization.   The demo will also feature embedded Jupyter-style notebooks for data wrangling, custom visualization, and other solution tailoring.
  • Title TBD:  DOMO
  • TBD

SPEAKER BIOS

Keynote Speakers

Mohammad Saffar, Machine Learning Engineer, Google

Doug Meiser, General Manager of R&D Operations Research (Kroger Co.)
Doug Meiser leads the Operations Research team within the Research & Development team of Kroger Technology. Doug earned a bachelor’s degree in Mathematics and Physics and an MBA from Northern Kentucky University. He spent three years in Supply Chain Systems before transitioning to the Research & Development team to develop his current team.

Claudia Perlich, Chief Scientist, Dstillery

Claudia Perlich leads the machine learning efforts that power Dstillery's digital intelligence for marketers and media companies. With more than 50 published scientific articles, she is a widely acclaimed expert on big data and machine learning applications, and an active speaker at data science and marketing conferences around the world.
Claudia is the past winner of the Advertising Research Foundation's (ARF) Grand Innovation Award and has been selected for Crain's New York’s 40 Under 40 list, Wired Magazine's Smart List, and Fast Company's 100 Most Creative People.
Claudia holds multiple patents in machine learning. She has won many data mining competitions and awards at Knowledge Discovery and Data Mining (KDD) conferences, and served as the organization's General Chair in 2014.
Prior to joining Dstillery in 2010, Claudia worked at IBM's Watson Research Center, focusing on data analytics and machine learning. She holds a PhD in Information Systems from NYU.

Afternoon Session Speakers

Dr. Brandon Greenwell, Senior Data Scientist, Ascend Innovations
In addition to his role at Ascend Innovations, Brandon is an Adjunct Professor of Statistics at Wright State University. He is the author of several R packages on CRAN and has published several articles, including two in The R Journal, and is currently co-authoring an R book---Advanced Business Analytics with R: Description, Prediction, and Prescription---to be published by CRC Press in early 2019.

Kaveh Bastani, Data Scientist, Recovery Decision Science
Kaveh is a data scientist with Recovery Decision Science. He received his PhD in Industrial and Systems Engineering from Virginia Tech in 2016. His current research interests are predictive modeling, risk analysis, and text mining w, ith applications to financial services. His research has appeared in high-quality journals including IIE-Transactions, Decision Support Systems, IEEE Transactions on Human-Machine Systems, and IEEE Transactions on Automation Science and Engineering.

Hamed Namavari, Data Scientist, Recovery Decision Science
Hamed has been with the Analytics Team at Recovery Decision Science (RDS) as Data Scientist since 2012. He has MA in Applied Economics and MS in Business Analytics from University of Cincinnati where he is currently an Econ PhD candidate. His research interests are focused primarily on Bayesian Inference in Spatiotemporal Econometrics.

Melissa Lacefield, Azure Cloud Solution Architect, Microsoft
Melissa Lacefield has been at Microsoft since Sept 2005, and during her 12 year tenure she has held several positions including an RRE, PFE, Beta Engineer with the SQL Product Group, CSS SQL Escalation Engineer, and a Technical Account manager. She is passionate about advanced analytics, data, and cybersecurity. She has enjoyed working with large enterprise customers in situations from data center floods to sandy military engagements overseas to large Edu and SLG customers.  She has two bachelors, masters and currently working on finishing her Ph.D. in cybersecurity. 

Clark MacDonald, Principal, Azure Platform Solutions
Clark MacDonald celebrates his 4th year anniversary at Microsoft this month.  Clark is a  Principal for Data & AI technologies. He has worked with customers across multiple industries on predictive analytics projects and other advanced analytics efforts leveraging Microsoft's Cloud.  He holds a Bachelor Degree in Computer Science from The Ohio State University and has several professional certifications.

Marks Wells, VP Business Development, Opalytics
After serving as a U.S. Marine officer in transportation and logistics, Mark attended Drexel University where he received his MBA, took post-MBA coursework in operations research, and taught Production and Operations Management and Operations Research in the evening college.  While attending graduate school, he drove and loaded package delivery trucks.
    He has held analytical positions in steel distribution, direct mail, and electronic retailing.  He has also had the opportunity to consult with the management of leading global supply chains in the chemicals, medical devices, consumer goods, high tech, and retail industries.  Mark has led projects in multiple areas of process improvement, including facility location, forecasting, inventory planning, order fulfillment and warehousing, cost-to-serve analysis, and more.
    Mark speaks occasionally at industry events and has been published in several periodicals, including the Journal of Enterprise Resource Management, CSCMP Comment, and Analytics.  He is also the publisher and author of the Supply Chain Action blog.
    Prior to joining Opalytics, Mark’s recent positions included Supply Chain Account Executive for retail at SAP, Director of Advanced Analytics Sales at Aera, and member of the worldwide advanced analytics software team at IBM.

Register for Machine Learning Day