Transformers As Approximations of Solomonoff Induction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Young, Nathan, Witbrock, Michael
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913476608262144
author Young, Nathan
Witbrock, Michael
author_facet Young, Nathan
Witbrock, Michael
contents Solomonoff Induction is an optimal-in-the-limit unbounded algorithm for sequence prediction, representing a Bayesian mixture of every computable probability distribution and performing close to optimally in predicting any computable sequence. Being an optimal form of computational sequence prediction, it seems plausible that it may be used as a model against which other methods of sequence prediction might be compared. We put forth and explore the hypothesis that Transformer models - the basis of Large Language Models - approximate Solomonoff Induction better than any other extant sequence prediction method. We explore evidence for and against this hypothesis, give alternate hypotheses that take this evidence into account, and outline next steps for modelling Transformers and other kinds of AI in this way.
format Preprint
id arxiv_https___arxiv_org_abs_2408_12065
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Transformers As Approximations of Solomonoff Induction
Young, Nathan
Witbrock, Michael
Artificial Intelligence
Solomonoff Induction is an optimal-in-the-limit unbounded algorithm for sequence prediction, representing a Bayesian mixture of every computable probability distribution and performing close to optimally in predicting any computable sequence. Being an optimal form of computational sequence prediction, it seems plausible that it may be used as a model against which other methods of sequence prediction might be compared. We put forth and explore the hypothesis that Transformer models - the basis of Large Language Models - approximate Solomonoff Induction better than any other extant sequence prediction method. We explore evidence for and against this hypothesis, give alternate hypotheses that take this evidence into account, and outline next steps for modelling Transformers and other kinds of AI in this way.
title Transformers As Approximations of Solomonoff Induction
topic Artificial Intelligence
url https://arxiv.org/abs/2408.12065