Time-and-space-efficient weighted deduction

J Eisner - Transactions of the Association for Computational …, 2023 - direct.mit.edu
Many NLP algorithms have been described in terms of deduction systems. Unweighted
deduction allows a generic forward-chaining execution strategy. For weighted deduction …

Transparency at the source: Evaluating and interpreting language models with access to the true distribution

J Jumelet, W Zuidema - arXiv preprint arXiv:2310.14840, 2023 - arxiv.org
We present a setup for training, evaluating and interpreting neural language models, that
uses artificial, language-like data. The data is generated using a massive probabilistic …

A fast algorithm for computing prefix probabilities

F Nowak, R Cotterell - Proceedings of the 61st Annual Meeting of …, 2023 - aclanthology.org
Multiple algorithms are known for efficiently calculating the prefix probability of a string under
a probabilistic context-free grammar (PCFG). Good algorithms for the problem have a …

An Exploration of Left-Corner Transformations

A Opedal, E Tsipidi, T Pimentel… - Proceedings of the …, 2023 - research-collection.ethz.ch
The left-corner transformation (Rosenkrantz and Lewis, 1970) is used to remove left
recursion from context-free grammars, which is an important step towards making the …

The whole truth and nothing but the truth: Faithful and controllable dialogue response generation with dataflow transduction and constrained decoding

H Fang, A Balakrishnan, H Jhamtani, J Bufe… - arXiv preprint arXiv …, 2022 - arxiv.org
In a real-world dialogue system, generated text must be truthful and informative while
remaining fluent and adhering to a prescribed style. Satisfying these constraints …

Finding structure in language models

J Jumelet - arXiv preprint arXiv:2411.16433, 2024 - arxiv.org
When we speak, write or listen, we continuously make predictions based on our knowledge
of a language's grammar. Remarkably, children acquire this grammatical knowledge within …