With one of our most eagerly anticipated summits around the corner, I decided to go through five of the top reasons I'm excited to attend the Machine Learning Summit May 9-10.
1) A chance to collaborate with other machine learning professionals
Few endeavors are more reliant on collaboration than machine learning. In order to create a production level model, you need experts from a myriad of fields. From data scientists to business leaders, they all need to come together and work towards a common goal.
This is why simply being in a room with fellow professionals, all of whom are looking for answers or want to share insight (most of the time, both) is so invaluable. Benefits range from increased self-awareness and troubleshooting to the feeling of support being amongst a tribe of people who understand you and the struggles you're going through.
Ahead of the event, we asked some of the speakers to share some of the challenges they have or are currently going through. Are you going through the same? Here are a few:
- How can we create learning models from extremely unbalanced binary data (beyond oversampling and undersampling)?
- How can we create learning models from extremely skewed class distributions in multiclass classification problems
- We can identify features and have the ML model rate a security based on these, but how can we take this to the next level and have the machine pinpoint the features to use in predicting price movement
- How do we let users know why the machine made the decision it did?
2) Topical and insightful presentations
First and foremost, you go to a summit like this for the presentations. One of the most efficient ways to solve a problem is to see how other professionals and brands solved a similar problem. And then ask them loads of questions about it!
Here are some of the presentation topics I am most excited about:
- Deep learning in medicine: Applications to disease diagnostics and DNA sequencing
- Growing the LinkedIn content ecosystem with machine learning
- Leveraging causal inference for demand estimation
- Deep Learning applications to online payment fraud detection
- Creating customer-facing experiences with Machine Learning
- COTA: Improving Uber customer care with NLP & Deep Learning
- Turning clicks into purchases: Revenue optimization for product search in E-Commerce
- The history (and future) of Machine Learning at Reddit!
Other themes which will be discussed over the summit:
-Building autonomous vehicles using machine learning
-Machine learning in search and recommendation systems
-Using deep learning in text-to-speech and speech-to-text
-Leveraging AI to optimize customer experience
3) Location - San Francisco
The tech Mecca for decades now, San Francisco has long been the hub of innovation. Funnily enough, San Fransisco first came to being as a railroad town. If you visit the railyard today, you can still see one of the first computers used in Silicon Valley. Tech has been an integral part of the city from its inception as both the tech and the city are relatively modern.
However, like the tech it has incubated over the years, San Francisco has exploded into one of the most influential locations in the United States. As the railroad brought in people, it also brought in internet cables which allowed supergiants to grow back when they were still baby startups contained in garages across the city.
San Fransisco is also the home of Stanford which is where companies like HP were born and where early versions of tech like semiconductors and NASA's wind tunnels were first developed. Along with thousands of engineers the university has produced over the years.
The constant flow of talent, the early pre-existent ecosystem geared towards the tech industry and liberal tax breaks have made the city the undeniable home for all things digital, and you can sense it everywhere you go there.
The year-round sun doesn't hurt either...
4) The people, the brands
Representatives from some of the biggest and most influential brands are out in full force, here are some of the people, and brands, I am most excited for:
Roelof van Zwol
Director, Product Innovation
Machine Learning Engineer
Senior Software Engineer
Head of Cloud AI & Machine Learning
Co-Founder and CTO
"Freebird shares some of their research applying Deep Learning methods to estimate the likelihood and severity of flight delays."
"Deep Learning is very accurate at making predictions, but each prediction does not have a confidence interval attached to it. Our research extends deep learning methods to have confidence intervals, which are crucial for managing risk!"
5) IBM workshop
IBM has been instrumental in the machine learning space. It was actually one of their employees, Arthur Samuels, who first coined the term.
Since then, they have consistently been at the forefront of the field. In recent weeks, they even announced their unexpected plans to team up with Apple. They will be combining IBM's flagship machine learning platform, Watson, with Apple's core ML in order to make Apple's business apps smarter.
They are yet to announce what they shall be showcasing at the summit, but from experience, they have rarely failed to impress.
There you have it, 5 reasons to attend this years Machine learning summit in San Fran this May.
Hope to see you there!