# What Has Been Enhanced in my Knowledge-Enhanced Language Model?

arXiv: https://arxiv.org/abs/2202.00964 · HTML: https://mlsanity.com/paper/2202.00964/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 40 of its 57 references that are in the map.
- First posted on arXiv 2 Feb 2022.

## Builds on (most cited first, 30 of 40)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering](https://mlsanity.com/paper/1606.09375/) (2016)
- [Language Models as Knowledge Bases?](https://mlsanity.com/paper/1909.01066/) (2019)
- [Modeling Relational Data with Graph Convolutional Networks](https://mlsanity.com/paper/1703.06103/) (2017)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [Spectral Networks and Locally Connected Networks on Graphs](https://mlsanity.com/paper/1312.6203/) (2013)
- [MINE: Mutual Information Neural Estimation](https://mlsanity.com/paper/1801.04062/) (2018)
- [Designing and Interpreting Probes with Control Tasks](https://mlsanity.com/paper/1909.03368/) (2019)
- [ERNIE: Enhanced Language Representation with Informative Entities](https://mlsanity.com/paper/1905.07129/) (2019)
- [How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answering](https://mlsanity.com/paper/2012.00955/) (2020)
- [Linguistic Knowledge and Transferability of Contextual Representations](https://mlsanity.com/paper/1903.08855/) (2019)
- [KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representation](https://mlsanity.com/paper/1911.06136/) (2019)
- [KG-BERT: BERT for Knowledge Graph Completion](https://mlsanity.com/paper/1909.03193/) (2019)
- [K-Adapter: Infusing Knowledge into Pre-Trained Models with Adapters](https://mlsanity.com/paper/2002.01808/) (2020)
- [K-BERT: Enabling Language Representation with Knowledge Graph](https://mlsanity.com/paper/1909.07606/) (2019)
- [Knowledge Enhanced Contextual Word Representations](https://mlsanity.com/paper/1909.04164/) (2019)
- [Time-Aware Language Models as Temporal Knowledge Bases](https://mlsanity.com/paper/2106.15110/) (2021)
- [Factual Probing Is [MASK]: Learning vs. Learning to Recall](https://mlsanity.com/paper/2104.05240/) (2021)
- [Information-Theoretic Probing for Linguistic Structure](https://mlsanity.com/paper/2004.03061/) (2020)
- [Universal Invariant and Equivariant Graph Neural Networks](https://mlsanity.com/paper/1905.04943/) (2019)
- [On the equivalence between graph isomorphism testing and function approximation with GNNs](https://mlsanity.com/paper/1905.12560/) (2019)
- [Do NLP Models Know Numbers? Probing Numeracy in Embeddings](https://mlsanity.com/paper/1909.07940/) (2019)
- [Interpreting Graph Neural Networks for NLP With Differentiable Edge Masking](https://mlsanity.com/paper/2010.00577/) (2020)
- [Knowledgeable or Educated Guess? Revisiting Language Models as Knowledge Bases](https://mlsanity.com/paper/2106.09231/) (2021)
- [Attention-based Graph Neural Network for Semi-supervised Learning](https://mlsanity.com/paper/1803.03735/) (2018)

## Built on by (most cited first, 2 of 2)

- [Combining Knowledge Graphs and Large Language Models](https://mlsanity.com/paper/2407.06564/) (2024)
- [Revisiting the Knowledge Injection Frameworks](https://mlsanity.com/paper/2311.01150/) (2023)

## Abstract (arXiv)

Pretrained language models (LMs) do not capture factual knowledge very well. This has led to the development of a number of knowledge integration (KI) methods which aim to incorporate external knowledge into pretrained LMs. Even though KI methods show some performance gains over vanilla LMs, the inner-workings of these methods are not well-understood. For instance, it is unclear how and what kind of knowledge is effectively integrated into these models and if such integration may lead to catastrophic forgetting of already learned knowledge. This paper revisits the KI process in these models with an information-theoretic view and shows that KI can be interpreted using a graph convolution operation. We propose a probe model called textitGraph Convolution Simulator (GCS) for interpreting knowledge-enhanced LMs and exposing what kind of knowledge is integrated into these models. We conduct experiments to verify that our GCS can indeed be used to correctly interpret the KI process, and we use it to analyze two well-known knowledge-enhanced LMs: ERNIE and K-Adapter, and find that only a small amount of factual knowledge is integrated in them. We stratify knowledge in terms of various relation types and find that ERNIE and K-Adapter integrate different kinds of knowledge to different extent. Our analysis also shows that simply increasing the size of the KI corpus may not lead to better KI; fundamental advances may be needed.