Pareto Probing: Trading Off Accuracy for Complexity
Open accessThe DOI matches its Crossref record; 14 of the 14 references it cites were found in a registry or an index.
Abstract
The question of how to probe contextual word representations for linguistic structure in a way that is both principled and useful has seen significant attention recently in the NLP literature.In our contribution to this discussion, we argue for a probe metric that reflects the fundamental trade-off between probe complexity and performance: the Pareto hypervolume.To measure complexity, we present a number of parametric and non-parametric metrics.Our experiments using Pareto hypervolume as an evaluation metric show that probes often do not conform to our expectations-e.g., why should the non-contextual fastText representations encode more morpho-syntactic information than the contextual BERT representations?These results suggest that common, simplistic probing tasks, such as part-of-speech labeling and dependency arc labeling, are inadequate to evaluate the linguistic structure encoded in contextual word representations.This leads us to propose full dependency parsing as a probing task.In support of our suggestion that harder probing tasks are necessary, our experiments with dependency parsing reveal a wide gap in syntactic knowledge between contextual and non-contextual representations.Our code can be found at https://github. com/rycolab/pareto-probing.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325
Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
Wu, Y., Schuster, M., Chen, Z., Le, Q. V., Norouzi, M., Macherey, W., Krikun, M., Cao, Y., Gao, Q., Macherey, K., et al. (2016) · arXiv · doi:10.48550/arxiv.1609.08144 · Cited by 5,319 · cited 1×
“For each of these models, we feed it a sentence and average the output word piece (Wu et al., 2016) representations for each word, as tokenized in the treebank.”
MethodALBERT: A Lite BERT for Self-supervised Learning of Language Representations
Lan, Z., Chen, M., Goodman, S., Gimpel, K., Sharma, P., & Soricut, R. (2019) · arXiv · doi:10.48550/arxiv.1909.11942 · Cited by 4,308 · cited 3×
“…86% 67% English 86% 68% Finnish 87% 63% Marathi 72% 62% Turkish 83% 58% to human-level scores on large, non-trivial tasks, e.g. natural language inference (Liu et al., 2019) and question answering (Lan et al., 2020)—with different performance patterns being observed on the toyish probing tasks.”
BackgroundMethodUnderstanding deep learning requires rethinking generalization
Zhang, C., Bengio, S., Hardt, M., Recht, B., & Vinyals, O. (2016) · arXiv · doi:10.48550/arxiv.1611.03530 · Cited by 2,838 · cited 2×
“…measures are rarely explicitly discussed as complexity metrics—although they are intuitive for that purpose—and have become common recently: Zhang et al. (2017) originated this trend by shuf-fling outputs of image data so the images were no longer predictive of the labels, using this result…”
BackgroundOn the Uniform Convergence of Relative Frequencies of Events to Their Probabilities
Vapnik, V. N., & Chervonenkis, A. Y. (1971) · Theory of Probability & Its Applications · doi:10.1137/1116025 · Cited by 2,448
BERT Rediscovers the Classical NLP Pipeline
Tenney, I., Das, D., & Pavlick, E. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1452 · Cited by 1,307 · cited 1×
“For instance, two of the most common probing tasks are part-of-speech labeling (POSL; Hewitt and Liang, 2019; Belinkov et al., 2017) and dependency arc labeling (DAL; Tenney et al., 2019a,b; Voita and Titov, 2020).”
BackgroundA Closer Look at Memorization in Deep Networks
Arpit, D., Jastrzębski, S., Ballas, N., Krueger, D., Bengio, E., Kanwal, M. S., Maharaj, T., Fischer, A., Courville, A., Bengio, Y., et al. (2017) · arXiv · doi:10.48550/arxiv.1706.05394 · Cited by 1,307 · cited 1×
“These structured inputs may be easier to represent internally regardless of the outputs, given current theories that early stages of training are committed to memorizing inputs (Arpit et al., 2017).”
Background(2019) · doi:10.18653/v1/p19-1356 · Cited by 1,291
Keeping the neural networks simple by minimizing the description length of the weights
Hinton, G. E., & van Camp, D. (1993) · Proceedings of the sixth annual conference on Computational learning theory - COLT '93 · doi:10.1145/168304.168306 · Cited by 994 · cited 1×
“4 While low-rank regularization is assumed to produce models that generalize better (Hinton and Van Camp, 1993; Langenberg et al., 2019), contrary to the classic bias–variance tradeoff, Gold-blum et al. (2020) found that biasing towards small nuclear norms instead hurts generalization.”
BackgroundA hybrid neural network‐first principles approach to process modeling
Psichogios, D. C., & Ungar, L. H. (1992) · AIChE Journal · doi:10.1002/aic.690381003 · Cited by 982
Cited by
Implicit Representations of Meaning in Neural Language Models
Li, B. Z., Nye, M., & Andreas, J. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.143 · Cited by 84
“A phenomenon is judged to be encoded by a model if the probe’s accuracy cannot be explained by its accuracy when trained on control tasks (Hewitt and Liang, 2019) or baseline models (Pimentel et al., 2020).”
Can Language Models Encode Perceptual Structure Without Grounding? A Case Study in Color
Abdou, M., Kulmizev, A., Hershcovich, D., Frank, S., Pavlick, E., & Søgaard, A. (2021) · arXiv · doi:10.48550/arxiv.2109.06129 · Cited by 70
“We vary α across a wide range of settings to examine the effect of probe complexity, which we measure using the nuclear norm of the linear projection matrix W ∈ Rφ×ι; ||W ||∗ = ∑min(φ,ι) i=1 σi(W ), where σi(W ) is the ith singular value of W (Pimentel et al., 2020).”
MethodCan Language Models Encode Perceptual Structure Without Grounding? A Case Study in Color
Abdou, M., Kulmizev, A., Hershcovich, D., Frank, S., Pavlick, E., & Søgaard, A. (2021) · Proceedings of the 25th Conference on Computational Natural Language Learning · doi:10.18653/v1/2021.conll-1.9 · Cited by 64
Emergent Linear Representations in World Models of Self-Supervised Sequence Models
Nanda, N., Lee, A., & Wattenberg, M. (2023) · Proceedings of the 6th BlackboxNLP Workshop: Analyzing and Interpreting Neural Networks for NLP · doi:10.18653/v1/2023.blackboxnlp-1.2 · Cited by 55
Metadata
- Authors
- Pimentel, T., Saphra, N., Williams, A., & Cotterell, R.
- Venue
- Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
- Published
- 2020
- Pages
- 3138-3153
- OpenAlex
- W3090805418
- Cited by
- 36
- References
- 14
- FWCI
- 3.43
- Type
- conference paper
- Language
- en
- Open access
- hybrid
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Pareto Probing: Trading Off Accuracy for Complexity. https://app.cochat.ai/papers/10.18653/v1/2020.emnlp-main.254
- The paper
- Pimentel, T., Saphra, N., Williams, A., & Cotterell, R. (2020). Pareto Probing: Trading Off Accuracy for Complexity. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). https://doi.org/10.18653/v1/2020.emnlp-main.254
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.