Induced Model Matching: Restricted Models Help Train Full-Featured Models

TL;DR

This paper introduces Induced Model Matching (IMM), a novel method that leverages restricted, side-information models to improve training of full-featured models across various domains.

Contribution

IMM aligns restricted models with full models, providing a unified framework that improves training by exploiting restricted data or models, surpassing prior methods like noising and reverse knowledge distillation.

Findings

IMM improves language model training with restricted features.

IMM outperforms prior approaches in consistency and effectiveness.

Applicable across domains including NLP and reinforcement learning.

Abstract

We consider scenarios where a very accurate (often small) predictive model using restricted features is available when training a full-featured (often larger) model. This restricted model may be thought of as side-information'', and can come either from an auxiliary dataset or from the same dataset by forcing the restriction. How can the restricted model be useful to the full model? To answer this, we introduce a methodology called Induced Model Matching (IMM). IMM aligns the context-restricted, or induced, version of the large model with the restricted model. We relate IMM to approaches such as noising, which is implicit in addressing the problem, and reverse knowledge distillation from weak teachers, which is explicit but does not exploit restriction being the nature of the weakness. We show that these prior methods can be thought of as approximations to IMM and can be problematic in terms of consistency. Experimentally, we first motivate IMM using logistic regression as a toy example. We then explore it in language modeling, the application that initially inspired it, and demonstrate it on both LSTM and transformer full models, using bigrams as restricted models. We lastly give a simple RL example, which shows that POMDP policies can help learn better MDP policies. The IMM principle is thus generally applicable in common scenarios where restricted data is cheaper to collect or restricted models are easier to learn.