Conducting a Data Science Contest in Your Organization w/ Ashutosh Sanzgiri

ashutosh-sanzgiriAshutosh Sanzgiri (@sanzgiri) is a Data Scientist at AppNexus, the world’s largest independent Online Advertising (Ad Tech) company. I develop algorithms for machine learning products and services that help digital publishers optimize the monetization of their inventory on the AppNexus platform.

Ashutosh has a diverse educational and career background. He’s attained a Bachelor’s degree in Engineering Physics from the Indian Institute of Technology, Mumbai, a Ph.D. in Particle Physics from Texas A&M University and he’s conducted Post-Doctoral research in Nuclear Physics at Yale University. In addition to these achievements Ashutosh also has a certificate in Computational Finance and an MBA from the Oregon Health & Sciences University.

Prior to joining AppNexus, Ashutosh has held positions in Embedded Software Development, Agile Project Management, Program Management and Technical Leadership at Tektronix, Xerox, Grass Valley and Nike.

Scaling Machine Learning (ML) at your organization means increasing ML knowledge beyond the Data Science (DS) department. Several companies have Data / ML literacy strategies in place, usually through an internal data science university or a formal training program. At AppNexus, we’ve been experimenting with different ways to expand the use of ML in our products and services and share responsibility for its evaluation. An internal contest adds a competitive element, and makes the learning process more fun. It can engage people to work on a problem that’s important to the company instead of working on generic examples (e.g. “cat vs dog” classification), and gives contestants familiarity with the tools used by the DS team.

In this talk, Ashutosh will present the experience of conducting a “Kaggle-style” internal DS contest at AppNexus. he’ll discuss our motivations for doing it and how we went about it. Then he’ll share the tools we developed to host the contest. The hope being you too will find inspiration to try something in your organization!

From Zero to Machine Learning for Everyone w/ Poul Peterson

poul-petersenBigML was founded in January 2011 in Corvallis, Oregon with the mission of making Machine Learning beautifully simple for everyone. We pioneered Machine Learning as a Service (MLaaS), creating our platform that effectively lowers the barriers of entry to help organizations of all industries and sizes to adopt Machine Learning.

As a local company with a mission in complete alignment with that of the conference, BigML would be delighted to partake in this first edition of ML4ALL.

“So you’ve heard of Machine Learning and are eager to make data driven decisions, but don’t know where to start? The first step is not to read all the latest and greatest research in artificial intelligence, but rather to focus on the data you have and the decisions you want to make. Fortunately, this is easier than ever because platforms like BigML have commoditized Machine Learning, providing a consistent abstraction making it simple to solve use cases across industries, no Ph.D.s required.

As a practical, jump-start into Machine Learning, Poul Petersen, CIO of BigML, will demonstrate how to build a housing recommender system. In just 30 minutes, he will cover a blend of foundational Machine Learning techniques like classification, anomaly detection, clustering, association discovery and topic modeling to create an end-to-end predictive application. More importantly, using the availability of an API will make it easy to put this model into production, on stage, complete with a voice interface and a GUI. Learn Machine Learning and find a great home – all without paying expensive experts!”

Poul Petersen (@pejpgrep) is the Chief Infrastructure Officer at BigML. An Oregon native, he has an MS degree in Mathematics as well as BS degrees in Mathematics, Physics and Engineering Physics from Oregon State University. With 20 plus years of experience building scalable and fault-tolerant systems in data centers, Poul currently enjoys the benefits of programmatic infrastructure, hacking in python to run BigML with only a laptop and a cloud.


Your First NLP Machine Learning Project: Perks and Pitfalls of Unstructured Data w/ Anna Widiger

anna-widigerAnna Widiger (@widiger_anna) has a B.A. degree in Computational Linguistics from University of Tübingen. She’s been doing NLP since her very first programming assignment, specializing in Russian morphology, German syntax, cross-lingual named entity recognition, topic modeling, and grammatical error detection.

Anna describes “Your First NLP Machine Learning Project: Perks and Pitfalls of Unstructured Data” to us. Faced with words instead of numbers, many data scientists prefer to feed words straight from csv files into lists without filtering or transformation, but there is a better way! Text normalization improves the quality of your data for future analysis and increases the accuracy of your machine learning model.

Which text preprocessing steps are necessary and which ones are “nice-to-have” depends on the source of your data and the information you want to extract from it. It’s important to know what goes into the bag of words and what metrics are useful to compare word frequencies in documents. In this hands-on talk, I will show some do’s and don’ts for processing tweets, Yelp reviews, and multilingual news articles using spaCy.

Jump or Not to Jump: Solving Flappy Bird with Deep Reinforcement Learning w/ Kaleo Ha’o

kaleo-haoThe holy grail of machine learning has always been a “Skynet” type of artificial intelligence that could generally learn anything (artificial general intelligence), and the biggest advancement toward achieving such A.I. is deep reinforcement learning (DRL). The reason for this is DRL’s ability to solve a vast array of problems; its application ranges from learning any video game, to driving autonomous cars, to landing Space-X Falcon 9 rockets. As such, DRL is one of the most exciting fields in machine learning, and will continue to be so for the foreseeable future.

This talk will be a deep dive into the math behind an algorithm which uses DRL to solve the video game Flappy Bird. Additionally the goal of this talk will be such that even the most math-phobic beginner-to-ML will walk away excited about DRL and able to verbalize the central equation to DRL. This will be done in two simple steps. First we will derive the Bellman Equation (classic reinforcement learning), which provides a framework for the algorithm’s decision making process. Second we will use that Bellman Equation to derive a custom loss function, which will drive the training of the algorithm’s deep neural network (deep learning).

To keep things simple during presenting, Kaleo will skip the code snippets but links will be provided where everybody can download the Flappy Bird algorithm’s code, a 20 page paper (written by Kaleo) detailing the algorithm, and a list of resources for further study.

The following is a link to the algorithm code and accompanying paper (PDF), where the math details can be found in section Algorithms and Techniques:

Kaleo Ha’o (@06kahao) is a graduate from Udacity’s Machine Learning Engineering program, Kaleo Ha’o is a freelance ML engineer who loves to contribute to Open A.I. research on reinforcement learning.

Eliminating Machine Bias w/ Mary Ann Brennan & Getting from Machine Learning Outputs to Decisions… timely things!

Meet Mary Ann Brennan (@mabmierau) presenting…

Eliminating Machine Bias


Mary Ann is a jack-of-all-trades software engineer with background in iOS and full-stack web at a bunch of startups you’ve probably never heard of. She’s currently using her extended maternity leave to refresh her studies in Artificial Intelligence, the concentration for her master’s in Computer Science at Stanford. She recently completed Coursera’s Deep Learning specialization.

Mary Ann will give a quick overview of the machine bias problem. She’ll provide a definition (and contrast with statistical bias) and show examples like criminal justice system algorithms with racial bias and machine translation algorithms with gender bias.

Mary Ann will also pull an example from the courses that shows how to remove gender bias from word representations. She’s been surprised how simple and manageable it was to do and hopes this will help inspire people to take the extra step to consider how they might eliminate bias in their work. Time permitting, she’ll also aim to round the talk out with other techniques and more info on what to look for in your own data.

In Brennan’s talk, there may even be a code up of web-based interactive before and after view of word representations that you can check out, maybe. Be sure to ask about it, as it might be in under construction!

Meet Robert Dodier (@robert-dodier) presenting…

Getting from Machine Learning Outputs to Decisions


Robert is interested in Bayesian inference and decision analysis, among many other topic. He’s got a background in math, computer science, and engineering. He also wrote his PhD dissertation on Bayesian network models applied to engineering problems. Since then he’s worked on statistical analysis, including machine learning, in different domains and also worked for some years on a agent-based system for control of distributed energy resources.

You’ve set up a machine learning pipeline and it’s delivering predictions and estimates in real time just as you hoped it would! Now, what can you do with those numbers? Intuitively, we think that if we compute a prediction that an event A is more likely than another event B, we should probably take an A-related action instead of a B-related action. But not if the A-action is really expensive compared to the B-action; then we’d hesitate unless we’re very sure about A. We could set up some rules to post-process the predictions to figure out which action to take. That sounds kind of hackish.

But wait! The good news is that we can formalize the decision problem in terms of statements of beliefs (i.e., probabilities) and values or preferences (i.e., utilities). Typical machine learning algorithms can produce probabilities, which we’ll take as the inputs for the decision-making step. Given fairly broad assumptions, the right way to make a decision is to take the action which maximizes expected (i.e., weighted average) utility. This gives a straightforward recipe for deriving actions from beliefs and preferences. Typically the expected utility calculation is much simpler than the belief calculation, so we can easily a devise a decison-making back end for our machine learning pipeline.

In this talk, Robert will pose some real-world decision problems and then talk about the features they have in common. He’ll describe a framework for finding best actions for such problems in general, and show how the framework applies to the problems he posed at the beginning of the talk. He’ll emphasize the idea that a formal decision making analysis gives an obvious way to inject business logic into the pipeline in a clear and easily-modified way.

Now, Kirsten Westeinde & Rich Jolly

Processed with VSCO with c1 presetUsing Your Data to Give Your Customers Superpowers

Kirsten Westeinde (@kmkwesteinde) is a technology enthusiast and a lifelong learner. Currently a senior software engineer at Shopify, where she solves challenging web development problems on the daily. She’s toiled and built apps and services with Ruby on Rails for greater than 5 years and is constantly adding new tools to her tool belt. In the important realm of building us humans, she is passionate about building diverse teams, mentoring & teaching, and gaining new perspectives through travel & cross cultural communication.
Continue reading “Now, Kirsten Westeinde & Rich Jolly”

Say Hello to Jon and Amy @ML4ALL!

Welcome the next two speakers I’m introducing: Jon Oropeza and Amy Cheng!

jon-oropezaML Spends A Year In Burgundy

Jon (@joropeza) is coming to Portland from Portland to speak to us about Cote d’Or in Burgundy! He’s a hacker, grape lover, and Portlander that loves a good smooth wine made out of those grapes he loves!

Jon being a both a weather nerd and a wine nerd, he was curious if machine learning could be applied to vintages in an area where a) quality of wine varies greatly by year b) most of that variance has to do with weather patterns and different aspects of temperature and precipitation c) there are known, reasonably-objective classifications or ‘scores’ of each vintage, such that we could say that such-and-such year with such-and-such weather produced wines of good/bad/mediocre quality.

Continue reading “Say Hello to Jon and Amy @ML4ALL!”