Thinking ahead: spontaneous prediction in context as a keystone of language in humans and machines
Open accessThe DOI matches its Crossref record; 42 of the 42 references it cites were found in a registry or an index.
Abstract
Departing from traditional linguistic models, advances in deep learning have resulted in a new type of predictive (autoregressive) deep language models (DLMs). Using a self-supervised next-word prediction task, these models are trained to generate appropriate linguistic responses in a given context. We provide empirical evidence that the human brain and autoregressive DLMs share three fundamental computational principles as they process natural language: 1) both are engaged in continuous next-word prediction before word-onset; 2) both match their pre-onset predictions to the incoming word to calculate post-onset surprise (i.e., prediction error signals); 3) both represent words as a function of the previous context. In support of these three principles, our findings indicate that: a) the neural activity before word-onset contains context-dependent predictive information about forthcoming words, even hundreds of milliseconds before the words are perceived; b) the neural activity after word-onset reflects the surprise level and prediction error; and c) autoregressive DLM contextual embeddings capture the neural representation of context-specific word meaning better than arbitrary or static semantic embeddings. Together, our findings suggest that autoregressive DLMs provide a novel and biologically feasible computational framework for studying the neural basis of language.
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
Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing
Benjamini, Y., & Hochberg, Y. (1995) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1995.tb02031.x · Cited by 109,482
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
Lewis, M., Liu, Y., Goyal, N., Ghazvininejad, M., Mohamed, A., Levy, O., Stoyanov, V., & Zettlemoyer, L. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.703 · Cited by 8,821
Self-citationChomsky, N. (1957) · doi:10.1515/9783112316009 · Cited by 7,885
Whatever next? Predictive brains, situated agents, and the future of cognitive science
Clark, A. (2013) · Behavioral and Brain Sciences · doi:10.1017/s0140525x12000477 · Cited by 5,924
Receiver Operating Characteristic Curve in Diagnostic Test Assessment
Mandrekar, J. N. (2010) · Journal of Thoracic Oncology · doi:10.1097/jto.0b013e3181ec173d · Cited by 4,506 · cited 1×
“Appendix II - Model Details Layer (type) Output Shape Param # input_1 (InputLayer) [(10, 164)] 0 conv1d (Conv1D) (8, 160) 78720 activation (Activation) (8, 160) 0 batch_normalization (BatchNo) (8, 160) 640 dropout (Dropout) (8, 160) 0 max_pooling1d (MaxPooling1D) (4, 160) 0 conv1d_1 (Conv1D) (3, 160) 51200 activation_…”
Kutas, M., & Federmeier, K. D. (2011) · Annual Review of Psychology · doi:10.1146/annurev.psych.093008.131123 · Cited by 4,377
Statistical Modeling: The Two Cultures (with comments and a rejoinder by the author)
Breiman, L. (2001) · Statistical Science · doi:10.1214/ss/1009213726 · Cited by 4,340 · cited 1×
“Such models are performance-oriented and are focused on learning how to generate formed linguistic outputs as a function of context 66 .”
BackgroundALBERT: 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
Prediction and Entropy of Printed English
Shannon, C. E. (1951) · Bell System Technical Journal · doi:10.1002/j.1538-7305.1951.tb01366.x · Cited by 2,686
“Cloze Procedure”: A New Tool for Measuring Readability
Taylor, W. L. (1953) · Journalism Quarterly · doi:10.1177/107769905303000401 · Cited by 2,332
Performance-optimized hierarchical models predict neural responses in higher visual cortex
Yamins, D. L. K., Hong, H., Cadieu, C. F., Solomon, E. A., Seibert, D., & DiCarlo, J. J. (2014) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1403112111 · Cited by 1,924
Early vocabulary growth: relation to language input and gender
Huttenlocher, J., Haight, W., Bryk, A., Seltzer, M., & Lyons, T. (1991) · Developmental Psychology · doi:10.1037/0012-1649.27.2.236 · Cited by 1,783
Natural speech reveals the semantic maps that tile human cerebral cortex
Huth, A. G., de Heer, W. A., Griffiths, T. L., Theunissen, F. E., & Gallant, J. L. (2016) · Nature · doi:10.1038/nature17637 · Cited by 1,504
Talking to Children Matters: Early Language Experience Strengthens Processing and Builds Vocabulary
Weisleder, A., & Fernald, A. (2013) · Psychological Science · doi:10.1177/0956797613488145 · Cited by 1,273
What do we mean by prediction in language comprehension?
Kuperberg, G. R., & Jaeger, T. F. (2016) · Language, Cognition and Neuroscience · doi:10.1080/23273798.2015.1102299 · Cited by 1,087
A deep learning framework for neuroscience
Richards, B. A., Lillicrap, T. P., Beaudoin, P., Bengio, Y., Bogacz, R., Christensen, A., Clopath, C., Costa, R. P., de Berker, A., Ganguli, S., et al. (2019) · Nature Neuroscience · doi:10.1038/s41593-019-0520-2 · Cited by 1,027 · cited 2×
“To the extent that the brain relies on a next-word prediction objective to learn how to use language in context, it likely uses a different implementation 57 .”
BackgroundManning, J. R., Jacobs, J., Fried, I., & Kahana, M. J. (2009) · The Journal of Neuroscience · doi:10.1523/jneurosci.2041-09.2009 · Cited by 973
A hierarchy of temporal receptive windows in human cortex
Hasson, U., Yang, E., Vallines, I., Heeger, D. J., & Rubin, N. (2008) · The Journal of Neuroscience · doi:10.1523/jneurosci.5487-07.2008 · Cited by 874
Self-citationEthayarajh, K. (2019) · Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) · doi:10.18653/v1/d19-1006 · Cited by 841
Hierarchical process memory: memory as an integral component of information processing
Hasson, U., Chen, J., & Honey, C. J. (2015) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2015.04.006 · Cited by 749
Self-citationThe effect of word predictability on reading time is logarithmic
Smith, N. J., & Levy, R. (2013) · Cognition · doi:10.1016/j.cognition.2013.02.013 · Cited by 725
Slow cortical dynamics and the accumulation of information over long timescales
Honey, C. J., Thesen, T., Donner, T. H., Silbert, L. J., Carlson, C. E., Devinsky, O., Doyle, W. K., Rubin, N., Heeger, D. J., & Hasson, U. (2012) · Neuron · doi:10.1016/j.neuron.2012.08.011 · Cited by 532
Self-citationTwo Guidelines for Bootstrap Hypothesis Testing
Hall, P., & Wilson, S. R. (1991) · Biometrics · doi:10.2307/2532163 · Cited by 471
Huang, Y., & Rao, R. P. N. (2011) · WIREs Cognitive Science · doi:10.1002/wcs.142 · Cited by 470
If deep learning is the answer, what is the question?
Saxe, A., Nelli, S., & Summerfield, C. (2021) · Nature Reviews Neuroscience · doi:10.1038/s41583-020-00395-8 · Cited by 397 · cited 1×
“Human brains and DLMs share computational principles 57,58 , but they are likely to implement them using radically different neural architectures 56,59 .”
BackgroundDirect Fit to Nature: An Evolutionary Perspective on Biological and Artificial Neural Networks
Hasson, U., Nastase, S. A., & Goldstein, A. (2020) · Neuron · doi:10.1016/j.neuron.2019.12.002 · Cited by 387
Self-citationA hierarchy of linguistic predictions during natural language comprehension
Heilbron, M., Armeni, K., Schoffelen, J.-M., Hagoort, P., & de Lange, F. P. (2022) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2201968119 · Cited by 343
Toward a universal decoder of linguistic meaning from brain activation
Pereira, F., Lou, B., Pritchett, B., Ritter, S., Gershman, S. J., Kanwisher, N., Botvinick, M., & Fedorenko, E. (2018) · Nature Communications · doi:10.1038/s41467-018-03068-4 · Cited by 324 · cited 2×
“Past research has leveraged language models and machine learning to extract information about semantic representation structure in the brain 13–19 .”
BackgroundResultMachine translation of cortical activity to text with an encoder–decoder framework
Makin, J. G., Moses, D. A., & Chang, E. F. (2020) · Nature Neuroscience · doi:10.1038/s41593-020-0608-8 · Cited by 321
Same Story, Different Story: The Neural Representation of Interpretive Frameworks
Yeshurun, Y., Swanson, S., Simony, E., Chen, J., Lazaridi, C., Honey, C. J., & Hasson, U. (2017) · Psychological Science · doi:10.1177/0956797616682029 · Cited by 300 · cited 1×
“Even a single change to one word or one sentence at the beginning of a story can alter the neural responses to all subsequent sentences 46,54 .”
Self-citationPrediction during natural language comprehension
Willems, R. M., Frank, S. L., Nijhof, A. D., Hagoort, P., & van den Bosch, A. (2016) · Cerebral Cortex · doi:10.1093/cercor/bhv075 · Cited by 288
Deep reinforcement learning and its neuroscientific implications
Botvinick, M., Wang, J. X., Dabney, W., Miller, K. J., & Kurth-Nelson, Z. (2020) · Neuron · doi:10.1016/j.neuron.2020.06.014 · Cited by 269
Heeger, D. J. (2017) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1619788114 · Cited by 220
fMRI reveals language-specific predictive coding during naturalistic sentence comprehension
Shain, C., Blank, I. A., van Schijndel, M., Schuler, W., & Fedorenko, E. (2020) · Neuropsychologia · doi:10.1016/j.neuropsychologia.2019.107307 · Cited by 215
Syntactic Structure from Deep Learning
Linzen, T., & Baroni, M. (2021) · Annual Review of Linguistics · doi:10.1146/annurev-linguistics-032020-051035 · Cited by 202 · cited 3×
“Finally, the internal contextual embedding space in DLMs can capture many aspects of the latent structure of human language, including syntactic trees, voice, co-references, morphology, and long-range semantic and pragmatic dependencies 1,71–73 .”
BackgroundHasson, U., Egidi, G., Marelli, M., & Willems, R. M. (2018) · Cognition · doi:10.1016/j.cognition.2018.06.018 · Cited by 157
Self-citationIncorporating Context into Language Encoding Models for fMRI
Jain, S., & Huth, A. G. (2018) · doi:10.1101/327601 · Cited by 155 · cited 1×
“These results agree with recent studies demonstrating that contextual embeddings better model neural responses to words than static semantic embeddings 18,19,48 .”
ResultAligning context-based statistical models of language with brain activity during reading
Wehbe, L., Vaswani, A., Knight, K., & Mitchell, T. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1030 · Cited by 95 · cited 1×
“These results agree with recent studies demonstrating that contextual embeddings better model neural responses to words than static semantic embeddings 18,19,48 .”
ResultAmplification of local changes along the timescale processing hierarchy
Yeshurun, Y., Nguyen, M., & Hasson, U. (2017) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1701652114 · Cited by 94 · cited 1×
“Even a single change to one word or one sentence at the beginning of a story can alter the neural responses to all subsequent sentences 46,54 .”
BackgroundSelf-citationMcClelland, J. L., Hill, F., Rudolph, M., Baldridge, J., & Schütze, H. (2020) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1910416117 · Cited by 91 · cited 2×
“Some theoretical and empirical papers, however, have recently begun to search for shared computational principles between DLMs and the brain’s representation of language 20,21 .”
BackgroundDeep language algorithms predict semantic comprehension from brain activity
Caucheteux, C., Gramfort, A., & King, J.-R. (2022) · Scientific Reports · doi:10.1038/s41598-022-20460-9 · Cited by 91
Cited by
The neural architecture of language: Integrative modeling converges on predictive processing
Schrimpf, M., Blank, I. A., Tuckute, G., Kauf, C., Hosseini, E. A., Kanwisher, N., Tenenbaum, J. B., & Fedorenko, E. (2021) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2105646118 · Cited by 517
Over-reliance on English hinders cognitive science
Blasi, D. E., Henrich, J., Adamou, E., Kemmerer, D., & Majid, A. (2022) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2022.09.015 · Cited by 396
Brains and algorithms partially converge in natural language processing
Caucheteux, C., & King, J.-R. (2022) · Communications Biology · doi:10.1038/s42003-022-03036-1 · Cited by 318
The neural architecture of language: Integrative modeling converges on predictive processing
(2020) · doi:10.1101/2020.06.26.174482 · Cited by 129
Deep language algorithms predict semantic comprehension from brain activity
Caucheteux, C., Gramfort, A., & King, J.-R. (2022) · Scientific Reports · doi:10.1038/s41598-022-20460-9 · Cited by 91
Metadata
- Authors
- Goldstein, A., Zada, Z., Buchnik, E., Schain, M., Price, A., Aubrey, B., Nastase, S. A., Feder, A., Emanuel, D., Cohen, A., Jansen, A., Gazula, H., Choe, G., Rao, A., Kim, S. C., Casto, C., Fanda, L., Doyle, W., Friedman, D., Dugan, P., Melloni, L., Reichart, R., Devore, S., Flinker, A., Hasenfratz, L., Levy, O., Hassidim, A., Brenner, M., Matias, Y., Norman, K. A., Devinsky, O., & Hasson, U.
- Venue
- bioRxiv (Cold Spring Harbor Laboratory)
- Publisher
- openRxiv
- Published
- 2020-12-03
- OpenAlex
- W3108672584
- Cited by
- 51
- References
- 42
- Type
- posted content
- Language
- en
- Open access
- green
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
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.