---
title: "Sequence Extrapolation"
authors:
  - "Philip D. Laird"
  - "Ronald Saul"
doi: ""
url: "https://ijcai.org/proceedings/1993/155"
pdf: "https://ijcai.org/proceedings/1993/0155.pdf"
venue: "Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence"
year: 1993
pages: "1120-1126"
track: "Machine Learning"
keywords:
---

# Sequence Extrapolation

Philip D. Laird, Ronald Saul

*Proceedings of the Thirteenth International Joint Conference on Artificial Intelligence*, pages 1120-1126.

## Abstract

We study sequence extrapolation as a machinelearning problem. Input examples come from a stream of data elements of a single type {integers, strings, etc.). Stream values are presented in order; the problem is to construct a hypothesis that explains the given values and predicts the subsequent ones. We define a new representation language for streams called elementary stream descriptions and present a unified algorithm that learns elementary streams over a generous family of data types that includes integers, lists, and strings. The algorithm has been implemented and an application targeted.