# Learning to Abstract with Nonparametric Variational Information Bottleneck

arXiv: https://arxiv.org/abs/2310.17284 · HTML: https://mlsanity.com/paper/2310.17284/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 39 references that are in the map.
- First posted on arXiv 26 Oct 2023.

## Builds on (most cited first, 25 of 25)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Neural Machine Translation of Rare Words with Subword Units](https://mlsanity.com/paper/1508.07909/) (2015)
- [Deep contextualized word representations](https://mlsanity.com/paper/1802.05365/) (2018)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [Deep Variational Information Bottleneck](https://mlsanity.com/paper/1612.00410/) (2016)
- [Object-Centric Learning with Slot Attention](https://mlsanity.com/paper/2006.15055/) (2020)
- [On the Variance of the Adaptive Learning Rate and Beyond](https://mlsanity.com/paper/1908.03265/) (2019)
- [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://mlsanity.com/paper/2105.13626/) (2021)
- [What you can cram into a single vector: Probing sentence embeddings for linguistic properties](https://mlsanity.com/paper/1805.01070/) (2018)
- [Synthetic and Natural Noise Both Break Neural Machine Translation](https://mlsanity.com/paper/1711.02173/) (2017)
- [SentEval: An Evaluation Toolkit for Universal Sentence Representations](https://mlsanity.com/paper/1803.05449/) (2018)
- [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://mlsanity.com/paper/2103.06874/) (2021)
- [Charformer: Fast Character Transformers via Gradient-based Subword Tokenization](https://mlsanity.com/paper/2106.12672/) (2021)
- [Character-Level Language Modeling with Deeper Self-Attention](https://mlsanity.com/paper/1808.04444/) (2018)
- [Hierarchical Multiscale Recurrent Neural Networks](https://mlsanity.com/paper/1609.01704/) (2016)
- [Implicit Reparameterization Gradients](https://mlsanity.com/paper/1805.08498/) (2018)
- [Conditional probing: measuring usable information beyond a baseline](https://mlsanity.com/paper/2109.09234/) (2021)
- [Bridging the Gap for Tokenizer-Free Language Models](https://mlsanity.com/paper/1908.10322/) (2019)
- [Learning to Create and Reuse Words in Open-Vocabulary Neural Language Modeling](https://mlsanity.com/paper/1704.06986/) (2017)
- [A Latent Morphology Model for Open-Vocabulary Neural Machine Translation](https://mlsanity.com/paper/1910.13890/) (2019)
- [The Architectural Bottleneck Principle](https://mlsanity.com/paper/2211.06420/) (2022)
- [Inducing Meaningful Units from Character Sequences with Dynamic Capacity Slot Attention](https://mlsanity.com/paper/2102.01223/) (2021)

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

- [Learning Distribution-Wise Control in Representation Space for Language Models](https://mlsanity.com/paper/2506.06686/) (2025)
- [Nonparametric Variational Regularisation of Pretrained Transformers](https://mlsanity.com/paper/2312.00662/) (2023)

## Abstract (arXiv)

Learned representations at the level of characters, sub-words, words and sentences, have each contributed to advances in understanding different NLP tasks and linguistic phenomena. However, learning textual embeddings is costly as they are tokenization specific and require different models to be trained for each level of abstraction. We introduce a novel language representation model which can learn to compress to different levels of abstraction at different layers of the same model. We apply Nonparametric Variational Information Bottleneck (NVIB) to stacked Transformer self-attention layers in the encoder, which encourages an information-theoretic compression of the representations through the model. We find that the layers within the model correspond to increasing levels of abstraction and that their representations are more linguistically informed. Finally, we show that NVIB compression results in a model which is more robust to adversarial perturbations.