arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2402.12513cs.LGcs.CL
收藏

诱导模型匹配:受限模型有助于训练全特征模型

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

University of Illinois Chicago(伊利诺伊大学芝加哥分校)

Usama Muneeb, Mesrob I. Ohannessian

AI总结:

本文提出诱导模型匹配(IMM)方法,通过将全特征模型的受限版本与已有的高精度受限模型对齐,从而利用受限模型辅助训练更大的全特征模型,并在逻辑回归、语言建模和强化学习任务中验证了其有效性。

摘要:

我们考虑如下场景:在训练一个全特征(通常更大)模型时,可以使用一个利用受限特征且非常准确(通常较小)的预测模型。这个受限模型可以被视为“辅助信息”,它可以来自一个辅助数据集,也可以通过强制施加限制而从同一数据集中获得。受限模型如何能够对全模型有用?为了回答这个问题,我们提出了一种称为诱导模型匹配(Induced Model Matching,IMM)的方法。IMM 将大模型的上下文受限版本(即诱导版本)与受限模型对齐。我们将 IMM 与诸如加噪(noising)之类的方法联系起来,加噪在解决该问题时是隐式的;同时我们也将 IMM 与来自弱教师的反向知识蒸馏联系起来,后者是显式的,但并未利用“限制正是弱点的本质”这一特点。我们表明,这些先前的方法可以被视为对 IMM 的近似,并且在一致性方面可能存在问题。在实验上,我们首先使用逻辑回归作为一个简单示例来阐明 IMM 的动机。随后,我们在最初启发该方法的语言建模应用中对其进行探索,并在 LSTM 和 Transformer 全模型上进行了演示,使用二元语法(bigrams)作为受限模型。最后,我们给出一个简单的强化学习示例,该示例表明 POMDP 策略可以帮助学习更好的 MDP 策略。因此,IMM 原则在受限数据收集成本更低或受限模型更容易学习的常见场景中具有普遍适用性。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑