The Traveling Observer Model: Multi-task Learning Through Spatial Variable Embeddings

TL;DR

This paper introduces the Traveling Observer Model, a multi-task learning framework that embeds variables in a shared space, enabling the model to predict across diverse tasks by capturing underlying spatial and temporal regularities.

Contribution

It presents a novel framework that learns variable embeddings jointly with model parameters, allowing multi-task learning across unrelated datasets and tasks.

Findings

Successfully recovers intuitive variable locations in space and time.

Outperforms task-specific and other multi-task models.

Exploits regularities across diverse, unrelated datasets.

Abstract

This paper frames a general prediction system as an observer traveling around a continuous space, measuring values at some locations, and predicting them at others. The observer is completely agnostic about any particular task being solved; it cares only about measurement locations and their values. This perspective leads to a machine learning framework in which seemingly unrelated tasks can be solved by a single model, by embedding their input and output variables into a shared space. An implementation of the framework is developed in which these variable embeddings are learned jointly with internal model parameters. In experiments, the approach is shown to (1) recover intuitive locations of variables in space and time, (2) exploit regularities across related datasets with completely disjoint input and output spaces, and (3) exploit regularities across seemingly unrelated tasks, outperforming task-specific single-task models and multi-task learning alternatives. The results suggest that even seemingly unrelated tasks may originate from similar underlying processes, a fact that the traveling observer model can use to make better predictions.