---
title: "Mova: how it learns — papers it applies — Elamur Research"
description: "10 papers Mova applies at the stage «how it learns»: How the parsers learn: perceptrons, instance-based learning, bootstrapping and the teacher loop."
canonical: "https://research.elamur.ai/free/en/projects/mova/learn/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/projects/mova/learn/)

[home](https://md.research.elamur.ai/free/en/) / [projects](https://md.research.elamur.ai/free/en/projects/) / [Mova](https://md.research.elamur.ai/free/en/projects/mova/) / [Main list](https://md.research.elamur.ai/free/en/projects/mova/main/) / how it learns

Mova · papers by stage

# how it learns

How the parsers learn: perceptrons, instance-based learning, bootstrapping and the teacher loop. The 10 works of [Mova](https://mova.elamur.ai/free/en/)'s own list that this stage uses; measured marks a work whose effect Mova measured in its own experiments (7 here). Each title opens our card: the summary, how Mova uses the work, its references and who cites it. What the stage does is on [Mova's pipeline page](https://mova.elamur.ai/free/en/); every stage is on [the main list](https://md.research.elamur.ai/free/en/projects/mova/main/).

- [Instance-Based Learning Algorithms](https://md.research.elamur.ai/free/en/papers/aha-1991-instance-based-learning-algorithms/)measured David W. Aha, Dennis Kibler, Marc K. Albert · 1991 **What Mova took:** `prag/src/model.rs` — the k-NN / IB1 classifier, one of three models behind a shared `Classifier` trait in the pragmatics crate (sentence form, speech act, indirectness, hedging, polarity, voice).
- [IGTree: Using Trees for Compression and Classification in Lazy Learning Algorithms](https://md.research.elamur.ai/free/en/papers/daelemans-1997-igtree-trees-compression-classification-lazy/)measured Walter Daelemans, Antal van den Bosch, Ton Weijters · 1997 **What Mova took:** `prag/src/model.rs` — IGTree is one of three models behind a shared `Classifier` trait in the pragmatics crate, trained on the same features as the perceptron and k-NN.
- [Assessing the Reliability of Statistical Software: Part I](https://md.research.elamur.ai/free/en/papers/mccullough-1998-assessing-reliability-statistical-software/)measured B. D. McCullough · 1998 **What Mova took:** `mlab/tests/ext_nist.rs` — the MATLAB-compatible numeric engine `mlab` is tested on NIST StRD datasets (11 regressions, 11 ANOVA, 9 univariate; data files under `mlab/tests/data/ext/nist/`).
- [Discriminative Reranking for Natural Language Parsing](https://md.research.elamur.ai/free/en/papers/collins-2000-discriminative-reranking-natural-language-parsing/)measured Michael Collins · 2000 **What Mova took:** `math/src/steps.rs` ("top-K reranking") — the arithmetic word-problem solver's beam-search perceptron produces the top-K distinct final expression trees (`Model::topk`, candidates with tree, score, value and actions), and a…
- [Discriminative Training Methods for Hidden Markov Models: Theory and Experiments with Perceptron Algorithms](https://md.research.elamur.ai/free/en/papers/collins-2002-discriminative-training-methods-hidden-markov/)measured Michael Collins · 2002 **What Mova took:** en/src/ptag.rs: the English PTB part-of-speech tagger is an averaged perceptron in the spirit of this paper, run as a left-to-right greedy/beam tagger (following Honnibal 2013) with hashed integer features and per-tag weights.
- [Curriculum Learning](https://md.research.elamur.ai/free/en/papers/bengio-2009-curriculum-learning/)measured Yoshua Bengio et al. · 2009 **What Mova took:** math/src/bin/mathsolve.rs, flag `--curriculum` of the `steps` model (word-problem solver as a transition system trained by a structured perceptron, math/src/steps.rs): before normal training the model gets 3 epochs on problems…
- [Structured Perceptron with Inexact Search](https://md.research.elamur.ai/free/en/papers/huang-2012-structured-perceptron-inexact-search/)measured Liang Huang, Suphan Fayong, Yang Guo · 2012 **What Mova took:** `math/src/steps.rs` (`Model::run`) — the arithmetic word-problem solver builds expression trees with a beam-search perceptron trained with early update by default.
- [C4.5: Programs for Machine Learning](https://md.research.elamur.ai/free/en/papers/quinlan-1993-c45-programs-machine-learning/) J. Ross Quinlan · 1993 **What Mova took:** `prag/src/model.rs` — the function `gain_ratio` computes (H(Y) − Σ p(v) H(Y|v)) / SplitInfo for each feature slot, following this book (and the same weighting as in the TiMBL memory-based learning package).
- [Bagging Predictors](https://md.research.elamur.ai/free/en/papers/breiman-1996-bagging-predictors/) Leo Breiman · 1996 **What Mova took:** `math/src/bin/mathsolve.rs` (`--bag`) — for the arithmetic word-problem solver, two extra beam-search perceptron models are trained with reversed and with shuffled training order (instead of bootstrap resamples), and their top-K…
- [Collocational bootstrapping: A hypothesis about the learning of subject-verb agreement in humans and neural networks](https://md.research.elamur.ai/free/en/papers/hobbs-2026-collocational-bootstrapping-hypothesis-about-learning/) Claire Hobbs, R. Thomas McCoy · 2026 **What Mova took:** `en/seeds/errors/singular-subject-plural-verb.md`: cited in the seed for the warning rules `en.errors.singular-subject-vbp` and `en.errors.singular-subject-vbp-auxcop`…