--- title: "Noah A. Smith — papers, page 3 — Elamur Research" description: "Noah A. Smith, Allen Institute: 97 papers on Elamur Research (2003–2026), h-index 93 — affiliations, topics, co-authors and who cites them…" canonical: "https://research.elamur.ai/free/en/authors/noah-a-smith/page/3/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/noah-a-smith/page/3/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / [Noah A. Smith](https://md.research.elamur.ai/free/en/authors/noah-a-smith/) / page 3 # Noah A. Smith · page 3 97 papers in our library · 2003–2026 · page 3 of 4 [The profile, affiliations and co-authors →](https://md.research.elamur.ai/free/en/authors/noah-a-smith/) ## Papers The most cited first (OpenAlex and our library). [← prev](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/2/) [1](https://md.research.elamur.ai/free/en/authors/noah-a-smith/) [2](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/2/) **3** [4](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/4/) [next →](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/4/) 51. **[Do All Languages Cost the Same? Tokenization in the Era of Commercial Language Models](https://md.research.elamur.ai/free/en/papers/ahia-2023-do-all-languages-cost-same/)** *(2023 · 174 citations)* Commercial LM APIs bill per token, but how many tokens a text needs depends on the tokenizer and its training data. 52. **[Binding Language Models in Symbolic Languages](https://md.research.elamur.ai/free/en/papers/cheng-2022-binding-language-models-symbolic-languages/)** *(2022 · 170 citations)* End-to-end neural systems perform well on NLP tasks but are hard to interpret and fragile. Symbolic programs are transparent but limited to what their grammar can express. Earlier neural-symbolic hybrids needed custom languages, task-specific modules and large training sets. The authors introduce BINDER, a framework… 53. **[Frame-Semantic Parsing with Softmax-Margin Segmental RNNs and a Syntactic Scaffold](https://md.research.elamur.ai/free/en/papers/swayamdipta-2017-frame-semantic-parsing-softmax-margin/)** *(2017 · 160 citations)* The paper addresses frame-semantic argument identification: finding and labeling the text spans that fill the roles of FrameNet frames. 54. **[Discriminative Lexical Semantic Segmentation with Gaps: Running the MWE Gamut](https://md.research.elamur.ai/free/en/papers/schneider-2014-discriminative-lexical-semantic-segmentation-gaps/)** *(2014 · 159 citations)* The paper addresses the automatic identification of multiword expressions (MWEs) in running English text. It covers the full range of MWE types, from proper names and compounds to phrasal verbs, support-verb constructions and fixed phrases. Earlier work typically targeted one MWE subclass at a time and handled only… 55. **[Is GPT-3 Text Indistinguishable from Human Text? Scarecrow: A Framework for Scrutinizing Machine Text](https://md.research.elamur.ai/free/en/papers/dou-2022-is-gpt-3-text-indistinguishable/)** *(2022 · 152 citations)* Recent work suggested that ordinary crowd workers can no longer reliably tell GPT-3 output from human writing. 56. **[Generation from Abstract Meaning Representation using Tree Transducers](https://md.research.elamur.ai/free/en/papers/flanigan-2016-generation-abstract-meaning-representation-tree/)** *(2016 · 152 citations)* The paper generates English from AMR graphs in two steps. A breadth-first spanning tree is taken from the graph, and a weighted tree-to-string transducer turns that tree into a sentence. The transducer rules are extracted from AMR-aligned sentences: each aligned subgraph becomes a rule with argument slots for its… 57. **[In-Context Learning for Few-Shot Dialogue State Tracking](https://md.research.elamur.ai/free/en/papers/hu-2022-context-learning-few-shot-dialogue/)** *(2022 · 144 citations)* Annotating task-oriented dialogues at the turn level is slow and expensive, and fine-tuned dialogue state tracking (DST) models must be retrained whenever a schema gains new slots or domains. 58. **[Dynamic Entity Representations in Neural Language Models](https://md.research.elamur.ai/free/en/papers/ji-2017-dynamic-entity-representations-neural-language/)** *(2017 · 144 citations)* Most neural language models do not represent entities beyond the words used to mention them, so they cannot follow who or what a long text is about. 59. **[NeuroLogic A*esque Decoding: Constrained Text Generation with Lookahead Heuristics](https://md.research.elamur.ai/free/en/papers/lu-2022-neurologic-esque-decoding-constrained-text/)** *(2022 · 140 citations)* Left-to-right decoders such as beam search and top-k sampling pick each token based only on the prefix generated so far. 60. **[Inoculation by Fine-Tuning: A Method for Analyzing Challenge Datasets](https://md.research.elamur.ai/free/en/papers/liu-2019-inoculation-fine-tuning-method-analyzing/)** *(2019 · 138 citations)* Challenge (adversarial) datasets often cause sharp drops in model accuracy. A drop alone does not show whether the model architecture is limited or whether the original training data simply lacked certain patterns. Liu, Schwartz and Smith address this with "inoculation by fine-tuning." A model trained on the standard… 61. **[Deep Encoder, Shallow Decoder: Reevaluating Non-autoregressive Machine Translation](https://md.research.elamur.ai/free/en/papers/kasai-2020-deep-encoder-shallow-decoder-reevaluating/)** *(2020 · 134 citations)* Non-autoregressive translation (NAR) is often presented as a faster alternative to standard autoregressive (AR) Transformers, with only a small loss in quality. 62. **[Parsing Tweets into Universal Dependencies](https://md.research.elamur.ai/free/en/papers/liu-2018-parsing-tweets-into-universal-dependencies/)** *(2018 · 121 citations)* The authors extend the UD guidelines to tweets: emoticons and retweet markers become non-syntactic or discourse tokens, abbreviations such as "idc" stay unnormalized as single words, multiword expressions can be merged, and tweets can have several roots. 63. **[Comprehensive Annotation of Multiword Expressions in a Social Web Corpus](https://md.research.elamur.ai/free/en/papers/schneider-2014-comprehensive-annotation-multiword-expressions-social/)** *(2014 · 120 citations)* The authors annotate every multiword expression (MWE) in a web corpus of about 55,000 words, sentence by sentence, without limiting the kinds of MWE. 64. **[Branch-Train-Merge: Embarrassingly Parallel Training of Expert Language Models](https://md.research.elamur.ai/free/en/papers/li-2022-branch-train-merge-embarrassingly-parallel/)** *(2022 · 114 citations)* Large language models usually require tightly synchronized multi-node clusters, because gradients and activations must be exchanged across all GPUs. 65. **[A Corpus and Model Integrating Multiword Expressions and Supersenses](https://md.research.elamur.ai/free/en/papers/schneider-2015-corpus-model-integrating-multiword-expressions/)** *(2015 · 111 citations)* The paper defines one task: find lexical expressions in running text, single- or multi-word, and label the noun and verb ones with a supersense (41 coarse classes). 66. **[We're Afraid Language Models Aren't Modeling Ambiguity](https://md.research.elamur.ai/free/en/papers/liu-2023-we-re-afraid-language-models/)** *(2023 · 110 citations)* The paper asks whether pretrained language models can notice when a sentence has more than one plausible meaning and separate those meanings. 67. **[A Latent Variable Model for Geographic Lexical Variation](https://md.research.elamur.ai/free/en/papers/eisenstein-2010-latent-variable-model-geographic-lexical/)** *(2010 · 110 citations)* Studies of dialect and sociolinguistic variation usually start from a small set of variables that researchers choose by hand. 68. **[Parsing with Multilingual BERT, a Small Corpus, and a Small Treebank](https://md.research.elamur.ai/free/en/papers/chau-2020-parsing-multilingual-bert-small-corpus/)** *(2020 · 104 citations)* Multilingual models such as mBERT serve low-resource language varieties unevenly. Varieties that are barely present in, or absent from, the pretraining data have too little labeled and unlabeled text to train a strong monolingual model. Chau, Lin, and Smith test whether mBERT can be adapted cheaply using only a small… 69. **[Time Waits for No One! Analysis and Challenges of Temporal Misalignment](https://md.research.elamur.ai/free/en/papers/luu-2022-time-waits-no-one-analysis/)** *(2022 · 102 citations)* The paper examines temporal misalignment, the situation in which the text used to train an NLP model and the text it is evaluated or deployed on come from different periods. 70. **[Learning Latent Personas of Film Characters](https://md.research.elamur.ai/free/en/papers/bamman-2013-learning-latent-personas-film-characters/)** *(2013 · 101 citations)* The authors propose viewing narrative through its characters rather than only as a sequence of events. Their goal is to automatically discover the typical roles (personas) of film characters, that is, archetypes and stereotypes such as the villain or the protagonist's love interest. To this end, they assembled a new… 71. **[Turning on the Turbo: Fast Third-Order Non-Projective Turbo Parsers](https://md.research.elamur.ai/free/en/papers/martins-2013-turning-turbo-fast-third-order/)** *(2013 · 97 citations)* Non-projective graph parsing with third-order features (grand- and tri-siblings, head bigrams) is NP-hard, so the paper uses AD3, an accelerated dual decomposition algorithm, and extends it with specialised head automata that solve the large subproblems exactly. 72. **[Probing Across Time: What Does RoBERTa Know and When?](https://md.research.elamur.ai/free/en/papers/liu-2021-probing-across-time-what-does/)** *(2021 · 96 citations)* The paper asks when, during pretraining, a masked language model acquires different kinds of knowledge, rather than only what a finished model knows. 73. **[What's In My Big Data?](https://md.research.elamur.ai/free/en/papers/elazar-2023-s-my-big-data/)** *(2023 · 92 citations)* Language models are trained on huge text collections whose contents are rarely documented. Creators usually describe how a corpus was assembled but not what it actually contains. This paper introduces WIMBD (What's In My Big Data?), an open-source toolkit for inspecting and comparing such corpora. It rests on two… 74. **[Provable Limitations of Acquiring Meaning from Ungrounded Form: What Will Future Language Models Understand?](https://md.research.elamur.ai/free/en/papers/merrill-2021-provable-limitations-acquiring-meaning-ungrounded/)** *(2021 · 89 citations)* Can a system trained only on raw text, with no grounding, ever acquire meaning? Merrill, Goldberg, Schwartz and Smith turn the informal debate started by Bender and Koller (2020) into a precise mathematical question. 75. **[Frame-Semantic Role Labeling with Heterogeneous Annotations](https://md.research.elamur.ai/free/en/papers/kshirsagar-2015-frame-semantic-role-labeling-heterogeneous/)** *(2015 · 88 citations)* The paper addresses argument identification in frame-semantic parsing: given a predicate and the FrameNet frame it evokes, the system must find the phrases that fill the frame's roles and label them. [← prev](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/2/) [1](https://md.research.elamur.ai/free/en/authors/noah-a-smith/) [2](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/2/) **3** [4](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/4/) [next →](https://md.research.elamur.ai/free/en/authors/noah-a-smith/page/4/)