Language model

From Wikipedia, the free encyclopedia

A language model is a probabilistic model of a natural language[1] that can generate probabilities of a series of words, based on text corpora in one or multiple languages it was trained on. In 1980, the first significant statistical language model was proposed, and during the decade IBM performed ‘Shannon-style’ experiments, in which potential sources for language modeling improvement were identified by observing and analyzing the performance of human subjects in predicting or correcting text.[2]

Language models are useful for a variety of tasks, including speech recognition[3] (helping prevent predictions of low-probability (e.g. nonsense) sequences), machine translation,[4] natural language generation (generating more human-like text), optical character recognition, handwriting recognition,[5] grammar induction,[6] information retrieval,[7][8] and other.

Large language models, currently their most advanced form, are a combination of larger datasets (frequently using scraped words from the public internet), feedforward neural networks, and transformers. They have superseded recurrent neural network-based models, which had previously superseded the pure statistical models, such as word n-gram language model.

Pure statistical models[edit]

Models based on word n-grams[edit]

A word n-gram language model is a purely statistical model of language. It has been superseded by recurrent neural network-based models, which has been superseded by large language models. [9] It is based on an assumption that the probability of the next word in a sequence depends only on a fixed size window of previous words. If only one previous word was considered, it was called a bigram model; if two words, a trigram model; if n-1 words, an n-gram model.[10] Special tokens were introduced to denote the start and end of a sentence and .

To prevent a zero probability being assigned to unseen words, each word's probability is slightly lower than its frequency count in a corpus. To calculate it, various methods were used, from simple "add-one" smoothing (assign a count of 1 to unseen n-grams, as an uninformative prior) to more sophisticated models, such as Good–Turing discounting or back-off models.


Maximum entropy language models encode the relationship between a word and the n-gram history using feature functions. The equation is

where is the partition function, is the parameter vector, and is the feature function. In the simplest case, the feature function is just an indicator of the presence of a certain n-gram. It is helpful to use a prior on or some form of regularization.

The log-bilinear model is another example of an exponential language model.

Skip-gram model[edit]

Skip-gram language model is an attempt at overcoming the data sparsity problem that preceding (i.e. word n-gram language model) faced. Words represented in an embedding vector were not necessarily consecutive anymore, but could leave gaps that are skipped over.[11]

Formally, a k-skip-n-gram is a length-n subsequence where the components occur at distance at most k from each other.

For example, in the input text:

the rain in Spain falls mainly on the plain

the set of 1-skip-2-grams includes all the bigrams (2-grams), and in addition the subsequences

the in, rain Spain, in falls, Spain mainly, falls on, mainly the, and on plain.

In skip-gram model, semantic relations between words are represented by linear combinations, capturing a form of compositionality. For example, in some such models, if v is the function that maps a word w to its n-d vector representation, then

where ≈ is made precise by stipulating that its right-hand side must be the nearest neighbor of the value of the left-hand side.[12][13]

Neural models[edit]

Recurrent neural network[edit]

Continuous representations or embeddings of words are produced in recurrent neural network-based language models (known also as continuous space language models).[14] Such continuous space embeddings help to alleviate the curse of dimensionality, which is the consequence of the number of possible sequences of words increasing exponentially with the size of the vocabulary, furtherly causing a data sparsity problem. Neural networks avoid this problem by representing words as non-linear combinations of weights in a neural net.[15]

Large language models[edit]

A large language model (LLM) is a language model characterized by its large size. Its size is enabled by AI accelerators, which are able to process vast amounts of text data, mostly scraped from the Internet.[16] LLMs are artificial neural networks which can contain a billion to a trillion weights, and are (pre-)trained using self-supervised learning and semi-supervised learning. Transformer architecture contributed to faster training.[17] Alternative architectures include the mixture of experts (MoE), which has been proposed by Google in 2017.[18][19][20]

As language models, they work by taking an input text and repeatedly predicting the next token or word.[21] Up to 2020, fine tuning was the only way a model could be adapted to be able to accomplish specific tasks. Larger sized models, such as GPT-3, however, can be prompt-engineered to achieve similar results.[22] They are thought to acquire embodied knowledge about syntax, semantics and "ontology" inherent in human language corpora, but also inaccuracies and biases present in the corpora.[23]

Notable examples include OpenAI's GPT models (e.g., GPT-3.5 and GPT-4, used in ChatGPT), Google's PaLM (used in Bard), and Meta's LLaMa, as well as BLOOM, Ernie 3.0 Titan, and Anthropic's Claude 2.

Although sometimes matching human performance, it is not clear they are plausible cognitive models. At least for recurrent neural networks it has been shown that they sometimes learn patterns which humans do not learn, but fail to learn patterns that humans typically do learn.[24]

Evaluation and benchmarks[edit]

Evaluation of the quality of language models is mostly done by comparison to human created sample benchmarks created from typical language-oriented tasks. Other, less established, quality tests examine the intrinsic character of a language model or compare two such models. Since language models are typically intended to be dynamic and to learn from data it sees, some proposed models investigate the rate of learning, e.g. through inspection of learning curves. [25]

Various data sets have been developed to use to evaluate language processing systems.[26] These include:

  • Corpus of Linguistic Acceptability[27]
  • GLUE benchmark[28]
  • Microsoft Research Paraphrase Corpus[29]
  • Multi-Genre Natural Language Inference
  • Question Natural Language Inference
  • Quora Question Pairs[30]
  • Recognizing Textual Entailment[31]
  • Semantic Textual Similarity Benchmark
  • SQuAD question answering Test[32]
  • Stanford Sentiment Treebank[33]
  • Winograd NLI
  • BoolQ, PIQA, SIQA, HellaSwag, WinoGrande, ARC, OpenBookQA, NaturalQuestions, TriviaQA, RACE, MMLU (Massive Multitask Language Understanding), BIG-bench hard, GSM8k, RealToxicityPrompts, WinoGender, CrowS-Pairs.[34] (LLaMa Benchmark)

See also[edit]


  1. ^ Jurafsky, Dan; Martin, James H. (2021). "N-gram Language Models". Speech and Language Processing (3rd ed.). Archived from the original on 22 May 2022. Retrieved 24 May 2022.
  2. ^ Rosenfeld, Ronald (2000). "Two decades of statistical language modeling: Where do we go from here?". Proceedings of the IEEE. 88 (8).
  3. ^ Kuhn, Roland, and Renato De Mori (1990). "A cache-based natural language model for speech recognition". IEEE transactions on pattern analysis and machine intelligence 12.6: 570–583.
  4. ^ Andreas, Jacob, Andreas Vlachos, and Stephen Clark (2013). "Semantic parsing as machine translation" Archived 15 August 2020 at the Wayback Machine. Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers).
  5. ^ Pham, Vu, et al (2014). "Dropout improves recurrent neural networks for handwriting recognition" Archived 11 November 2020 at the Wayback Machine. 14th International Conference on Frontiers in Handwriting Recognition. IEEE.
  6. ^ Htut, Phu Mon, Kyunghyun Cho, and Samuel R. Bowman (2018). "Grammar induction with neural language models: An unusual replication" Archived 14 August 2022 at the Wayback Machine. arXiv:1808.10000.
  7. ^ Ponte, Jay M.; Croft, W. Bruce (1998). A language modeling approach to information retrieval. Proceedings of the 21st ACM SIGIR Conference. Melbourne, Australia: ACM. pp. 275–281. doi:10.1145/290941.291008.
  8. ^ Hiemstra, Djoerd (1998). A linguistically motivated probabilistically model of information retrieval. Proceedings of the 2nd European conference on Research and Advanced Technology for Digital Libraries. LNCS, Springer. pp. 569–584. doi:10.1007/3-540-49653-X_34.
  9. ^ Bengio, Yoshua; Ducharme, Réjean; Vincent, Pascal; Janvin, Christian (1 March 2003). "A neural probabilistic language model". The Journal of Machine Learning Research. 3: 1137–1155 – via ACM Digital Library.
  10. ^ Jurafsky, Dan; Martin, James H. (7 January 2023). "N-gram Language Models". Speech and Language Processing (PDF) (3rd edition draft ed.). Retrieved 24 May 2022.
  11. ^ David Guthrie; et al. (2006). "A Closer Look at Skip-gram Modelling" (PDF). Archived from the original (PDF) on 17 May 2017. Retrieved 27 April 2014.
  12. ^ Cite error: The named reference mikolov was invoked but never defined (see the help page).
  13. ^ Cite error: The named reference compositionality was invoked but never defined (see the help page).
  14. ^ Karpathy, Andrej. "The Unreasonable Effectiveness of Recurrent Neural Networks". Archived from the original on 1 November 2020. Retrieved 27 January 2019.
  15. ^ Bengio, Yoshua (2008). "Neural net language models". Scholarpedia. Vol. 3. p. 3881. Bibcode:2008SchpJ...3.3881B. doi:10.4249/scholarpedia.3881. Archived from the original on 26 October 2020. Retrieved 28 August 2015.
  16. ^ "Better Language Models and Their Implications". OpenAI. 14 February 2019. Archived from the original on 19 December 2020. Retrieved 25 August 2019.
  17. ^ Merritt, Rick (25 March 2022). "What Is a Transformer Model?". NVIDIA Blog. Retrieved 25 July 2023.
  18. ^ Shazeer, Noam; Mirhoseini, Azalia; Maziarz, Krzysztof; Davis, Andy; Le, Quoc; Hinton, Geoffrey; Dean, Jeff (1 January 2017). "Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer". arXiv:1701.06538 [cs.LG].
  19. ^ Lepikhin, Dmitry; Lee, HyoukJoong; Xu, Yuanzhong; Chen, Dehao; Firat, Orhan; Huang, Yanping; Krikun, Maxim; Shazeer, Noam; Chen, Zhifeng (12 January 2021). "GShard: Scaling Giant Models with Conditional Computation and Automatic Sharding". arXiv:2006.16668 [cs.CL].
  20. ^ Dai, Andrew M; Du, Nan (9 December 2021). "More Efficient In-Context Learning with GLaM". Retrieved 9 March 2023.
  21. ^ Bowman, Samuel R. (2023). "Eight Things to Know about Large Language Models". arXiv:2304.00612 [cs.CL].
  22. ^ Brown, Tom B.; Mann, Benjamin; Ryder, Nick; Subbiah, Melanie; Kaplan, Jared; Dhariwal, Prafulla; Neelakantan, Arvind; Shyam, Pranav; Sastry, Girish; Askell, Amanda; Agarwal, Sandhini; Herbert-Voss, Ariel; Krueger, Gretchen; Henighan, Tom; Child, Rewon; Ramesh, Aditya; Ziegler, Daniel M.; Wu, Jeffrey; Winter, Clemens; Hesse, Christopher; Chen, Mark; Sigler, Eric; Litwin, Mateusz; Gray, Scott; Chess, Benjamin; Clark, Jack; Berner, Christopher; McCandlish, Sam; Radford, Alec; Sutskever, Ilya; Amodei, Dario (December 2020). Larochelle, H.; Ranzato, M.; Hadsell, R.; Balcan, M.F.; Lin, H. (eds.). "Language Models are Few-Shot Learners" (PDF). Advances in Neural Information Processing Systems. Curran Associates, Inc. 33: 1877–1901.
  23. ^ Manning, Christopher D. (2022). "Human Language Understanding & Reasoning". Daedalus. 151 (2): 127–138. doi:10.1162/daed_a_01905. S2CID 248377870.
  24. ^ Hornstein, Norbert; Lasnik, Howard; Patel-Grosz, Pritty; Yang, Charles (9 January 2018). Syntactic Structures after 60 Years: The Impact of the Chomskyan Revolution in Linguistics. Walter de Gruyter GmbH & Co KG. ISBN 978-1-5015-0692-5. Archived from the original on 16 April 2023. Retrieved 11 December 2021.
  25. ^ Karlgren, Jussi; Schutze, Hinrich (2015), "Evaluating Learning Language Representations", International Conference of the Cross-Language Evaluation Forum, Lecture Notes in Computer Science, Springer International Publishing, pp. 254–260, doi:10.1007/978-3-319-64206-2_8, ISBN 9783319642055
  26. ^ Devlin, Jacob; Chang, Ming-Wei; Lee, Kenton; Toutanova, Kristina (10 October 2018). "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding". arXiv:1810.04805 [cs.CL].
  27. ^ "The Corpus of Linguistic Acceptability (CoLA)". Archived from the original on 7 December 2020. Retrieved 25 February 2019.
  28. ^ "GLUE Benchmark". Archived from the original on 4 November 2020. Retrieved 25 February 2019.
  29. ^ "Microsoft Research Paraphrase Corpus". Microsoft Download Center. Archived from the original on 25 October 2020. Retrieved 25 February 2019.
  30. ^ Aghaebrahimian, Ahmad (2017), "Quora Question Answer Dataset", Text, Speech, and Dialogue, Lecture Notes in Computer Science, vol. 10415, Springer International Publishing, pp. 66–73, doi:10.1007/978-3-319-64206-2_8, ISBN 9783319642055
  31. ^ Sammons, V.G.Vinod Vydiswaran, Dan Roth, Mark; Vydiswaran, V.G.; Roth, Dan. "Recognizing Textual Entailment" (PDF). Archived from the original (PDF) on 9 August 2017. Retrieved 24 February 2019.{{cite web}}: CS1 maint: multiple names: authors list (link)
  32. ^ "The Stanford Question Answering Dataset". Archived from the original on 30 October 2020. Retrieved 25 February 2019.
  33. ^ "Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank". Archived from the original on 27 October 2020. Retrieved 25 February 2019.
  34. ^ Hendrycks, Dan (14 March 2023), Measuring Massive Multitask Language Understanding, archived from the original on 15 March 2023, retrieved 15 March 2023

Further reading[edit]

  • J M Ponte; W B Croft (1998). "A Language Modeling Approach to Information Retrieval". Research and Development in Information Retrieval. pp. 275–281. CiteSeerX
  • F Song; W B Croft (1999). "A General Language Model for Information Retrieval". Research and Development in Information Retrieval. pp. 279–280. CiteSeerX
  • Chen, Stanley; Joshua Goodman (1998). An Empirical Study of Smoothing Techniques for Language Modeling (Technical report). Harvard University. CiteSeerX