AI News, Real-world ML lessons from GAFA, top startups researchers

Real-world ML lessons from GAFA, top startups researchers

Hindi points out that this is likely to be an intractable issue for voice assistants such as Amazon Echo or Google Home, and takes the example of hosting a dinner party with many guests… Unless the data doesn’t leave the device and ML takes place on-device.

It’s interesting to see how new methodology and software are being developed to inspect and understand how black-box models work, at the same time as automated modeling techniques are being developed and made easier to use in practice — see auto-sklearn for instance, which acts as a drop-in replacement for a scikit-learn estimator that performs algorithm selection and hyper-parameter tuning automatically.

You can now develop accurate models in one click/one line of code, and then inspect them (beyond the usual aggregate pre-deployment performance metrics) to decide whether you should trust them and to ensure that they will be robust when deployed in the real world!

As modeling gets automated, it becomes also clear that automating ML involves more than the modeling component… You want to automate as much as possible of the whole workflow that goes from data collection and preparation into ML-ready data, to deployment of models and predictions at scale!

In “Machine Teaching: A New Paradigm for Building Machine Learning Systems”, researchers at Microsoft make a distinction between Machine Learning and Machine Teaching, a new discipline that extends principles of SE to the domain of predictive models and aims at enabling more individuals to build ML systems: Microsoft, Google and Amazon are driving research in ML, and at the same time they provide ML products on their public clouds that aim at boosting our productivity.

If your company is looking to build its own production-ready ML platform, PAPIs published a few articles in Proceedings of Machine Learning Research that share lessons learned by teams at Microsoft Azure ML (“Anatomy of a machine learning service”), Upwork (“Deploying high throughput predictive models with the actor framework”), and more recently Uber (“Scaling ML as a Service”), which are a great starting point.

Machine learning models + IoT data = a smarter world (Google I/O '18)

With the IoT market set to triple in size by 2020, and massive increases in computing power on small devices, the intersection of IoT and machine learning is a ...

Transform Retail with Machine Learning: Find & Recommend products (Google Cloud Next '17)

Businesses today are realizing that they can use machine learning to improve customer experience. With the most recent models, you can simplify product ...

Choosing the right ML algorithm

The Microsoft Azure Machine Learning Algorithm Cheat Sheet helps you choose the right machine learning algorithm for your predictive analytics solutions from ...

Your First ML App - Machine Learning for Hackers #1

This video will get you up and running with your first ML app in just 7 lines of Python. The app will be able to recognize Iris flowers. I created a Slack channel for ...

Introduction to Google Cloud Machine Learning (Google Cloud Next '17)

Google Cloud is at the forefront of developing cutting-edge machine learning technology. Computer vision, predictive modeling, natural language understanding ...

Adding Machine Learning to your applications

Google provides infrastructure, services, and APIs for you to create your own machine learning models. They also have pretrained machine learning APIs that ...

Machine Learning: Client-side ML vs Server-side ML

"Machine Learning (ML) and Artificial Intelligence are becoming ubiquitous in websites, chatbots and mobile apps. While JavaScript/Node.js may not be the ...

Boosting Product Categorization with Machine Learning

Product categories are the structural backbone of every online shop. To attract customers and facilitate navigation, categories need to be easily understandable, ...

Lifecycle of a machine learning model (Google Cloud Next '17)

In this video, you'll hear lessons learned from our experience with machine learning and how having a great model is only a part of the story. You'll see how ...

What is machine learning and how to learn it ?

Machine learning is just to give trained data to a program and get better result for complex problems. It is very close to data ..