Counterfactual decision-making across a broad set of fields

What is it about?

It uses counterfactuals (i.e., `what if' questions) to arrive at personalized decisions. This paper demonstrates the use of this technique to personalized medical decision-making. However, it can be used for decision-making in many other areas as well, e.g., policy, sports, sociology, politics, etc.

Featured Image

Why is it important?

Prior state-of-the-art was based on linear regression, which uses time-agnostic weights. However, many applications are better modeled by taking the spatiotemporal context into account. SCouT uses a transformer to perform sequence-to-sequence mapping with much better results than prior art.