Collaborating Authors

From Data to AI with the Machine Learning Canvas (Part III)


I like to think of ML tasks as questions in a certain format, for which the system we're building gives answers. The question has to be about a certain "object" of the real world (which we call the input). In the supervised learning paradigm -- which we're focusing on in this series -- we would make the system learn from example objects AND from the answers for each of them. The inputs in those questions are an email and a property. The Data Sources listed in the LEARN part of the Canvas (see Part II) should provide information about these inputs.

Video Prediction via Selective Sampling

Neural Information Processing Systems

Most adversarial learning based video prediction methods suffer from image blur, since the commonly used adversarial and regression loss pair work rather in a competitive way than collaboration, yielding compromised blur effect. In the meantime, as often relying on a single-pass architecture, the predictor is inadequate to explicitly capture the forthcoming uncertainty. Our work involves two key insights: (1) Video prediction can be approached as a stochastic process: we sample a collection of proposals conforming to possible frame distribution at following time stamp, and one can select the final prediction from it. Combining above two insights we propose a two-stage network called VPSS (\textbf{V}ideo \textbf{P}rediction via \textbf{S}elective \textbf{S}ampling). Specifically a \emph{Sampling} module produces a collection of high quality proposals, facilitated by a multiple choice adversarial learning scheme, yielding diverse frame proposal set.

Serverless predictions at scale – Yufeng G – Medium


Google's Cloud Machine Learning Engine enables let's you create a prediction service for your TensorFlow model without any ops work. Get more time to work with your data, by going from a trained model to a deployed, auto-scaling prediction service in a matter minutes. So, we've gathered our data, and finally finished training up a suitable model and validating that it performs well. We are now finally ready to move to the final phase: serving your predictions. When taking on the challenge of serving predictions, we would ideally want to deploy a model that is purpose-built for serving.

Machine Learning for Generalizable Prediction of Flood Susceptibility


Flooding is a destructive and dangerous hazard and climate change appears to be increasing the frequency of catastrophic flooding events around the world. Physics-based flood models are costly to calibrate and are rarely generalizable across different river basins, as model outputs are sensitive to site-specific parameters and human-regulated infrastructure. In contrast, statistical models implicitly account for such factors through the data on which they are trained. Such models trained primarily from remotely-sensed Earth observation data could reduce the need for extensive in-situ measurements. In this work, we develop generalizable, multi-basin models of river flooding susceptibility using geographically-distributed data from the USGS stream gauge network.

Salesforce aims to make AI predictions easier for its customers


Salesforce is launching a new feature that uses machine learning to let customers make predictions based on data stored with the tech titan's software. Einstein Prediction Builder is designed to take any of the fields that users have set up inside a piece of software -- like Salesforce's Sales Cloud CRM -- and generate predictions for future outcomes based on saved data. For example, Salesforce users could input information they have about customers and use Einstein Prediction Builder to generate a churn prediction system that would score each customer based on how likely they are to stop using a product or service. The feature is designed to make it easier for companies and users who don't have machine learning expertise to reap the benefits of the current explosion in AI technology. Custom predictions can be set up through a visual editor and don't require coding of any sort.