Event Diary - 8 Essential Tips for Implementing Machine Learning Models

Posted on December 13, 2017
By Avinash Patchava
and Rajesh Pantina, Marketing Manager, IndiaMarketing Manager, India

InMobi regularly hosts, participates in and contributes to a series of events, meet-ups, webinars, sharing best practices with partners and thought leaders, across the globe. In this edition of the InMobi Event Diary, we are looking at our recent visit to one of India’s top business schools - the Indian Institute of Management in Bangalore, where we addressed students of the Big Data and Analytics Certificate Program.



In this edition of the InMobi event diary, we cover our visit to the well-known Indian Institute of Management of Bangalore where we addressed 60 students of the Big Data Analytics Certificate Program. Speaking on Machine Learning in the Wild, the event was a great opportunity to interact with a student group comprising experienced professionals looking to move to the field of Big Data and Analytics, and address their concerns and thoughts.

InMobi was invited to address the 2017-18 batch given the on-ground experience in applying Big Data and Machine Learning technologies in the mobile advertising industry. InMobi has been applying these technologies for several years and their relevance has only grown as the industry and company have evolved. Avi Patchava, VP, Machine Learning and Artificial Intelligence, at InMobi addressed the batch. He was introduced by Dr. Shankar Venkatagiri - Associate Professor at Indian Institute of Management Bangalore.

Avi led the discussion with 8 Essential Tips to consider when designing, building and deploying a Machine Learning model towards a business problem:

Tip #1:Be clear about the business problem you seek to solve. Use a problem statement worksheet to align on the problem definition, solution intent, problem scope, hypotheses to guide the model, the data that you need, the value at stake and the team skills required.

Tip #2:Think carefully on what your model needs to achieve. Is your model descriptive, diagnostic, predictive, or prescriptive use case? Be wary of trying to do everything with one model. Ask yourself: what is the business really asking for?

Tip #3: Ensure feature availability in the production environment. Having lots of data and features is often a positive sign. But you need to ensure that all features are available in the production environment at the time of use - whether for exploration, training, evaluation or deployment.

Tip #4: Apply Occam’s razor for the choice of algorithms. Zero in on to the algorithm that will do the job - answer the problem that you set out to solve. You do not need the most glamourous algorithms, when a more modest algorithm can do the trick.

Tip #5: Be aware of the training time in production environments. If you need to train models multiple times per day, and a model takes several hours to train (because of its complexity), then your model will not be effective (regardless of its complexity!).

Tip #6: Think of data science beyond research. You need to support the drive for impact by considering what is needed to make your code ready for the production environment. For example, lighter size of files, logging and storing of key variables is something to ensure for easy evaluation of models.

Tip #7: Fortify the model with support elements. Achieving lasting business impact from a model, requires more than excellent accuracy. Many other elements are needed to ensure model impact whether UI-UX design, data infrastructure, training model users, design of the business process, or the overall change management as a model transforms a business.

Tip #8: 'De-jargonize' Data Sciences. There is a lot of jargon surrounding the field of data sciences, machine learning and especially artificial intelligence. One of your roles as a data scientist is to help your colleagues and team members cut through the jargon. Take the opportunity to clarify key terms, and even define their usage in your organization with relevant examples.