Category: NLP

A theory of error analysis

Manual error analyses can help to identify the strengths and weaknesses of computational systems, ultimately suggesting future improvements and guiding development. However, they are often treated as an afterthought or neglected altogether. In three of my recent papers, we have been slowly developing what might be called a theory of error analysis. The systems evaluated include:

  • number normalization (Gorman & Sproat 2016); e.g., mapping 97000 onto quatre vingt dix sept mille,
  • inflection generation (Gorman et al. 2019); e.g., mapping pairs citation form and inflectional specification like (aufbauen, V;IND;PRS;2) onto inflected forms like baust auf, and
  • grapheme-to-phoneme conversion (Lee et al. 2020); e.g., mapping orthographic forms like almohadilla onto phonemic or phonetic forms like /almoaˈdiʎa/ and [almoaˈðiʎa].

While these are rather different types of problems, the systems all have one thing in common: they generate linguistic representations. I discern three major classes of error such systems might make.

  • Target errors are only apparent errors; they arise when the gold data, the data to be predicted, is linguistically incorrect. This is particularly likely to arise with crowd-sourced data though such errors are also present in professionally annotated resources.
  • Linguistic errors are caused by misapplication of independently attested linguistic behaviors to the wrong input representations.
    • In the case of number normalization, these include using the wrong agreement affixes in Russian numbers; e.g., nom.sg. *семьдесят миллион for gen.sg. семьдесят миллионов ‘nine hundred million’ (Gorman & Sproat 2016:516)
    • In inflection generation, these are what Gorman et al. 2019 call allomorphy errors; e.g., for instance, overapplying ablaut to the Dutch weak verb printen ‘to print’ to produce a preterite *pront instead of printte (Gorman et al. 2019:144).
    • In grapheme-to-phoneme conversion, these include failures to apply allophonic rules; e,g, in Korean, 익명 ‘anonymity’ is incorrectly transcribed as [ikmjʌ̹ŋ] instead of [iŋmjʌ̹ŋ], reflecting a failure to apply a rule of obstruent nasalization not indicated in the highly abstract hangul orthography (Lee et al. under review).
  • Silly errors are those errors which cannot be analyzed as either target errors or linguistic errors. These have long been noted as a feature of neural network models (e.g., Pinker & Prince 1988, Sproat 1992:216f. for discussion of *membled) and occur even with modern neural network models.

I propose that this tripartite distinction is a natural starting point when building an error taxonomy for many other language technology tasks, namely those that can be understood as generating linguistic sequences.

References

K. Gorman, A. D. McCarthy, R. Cotterell, E. Vylomova, M. Silfverberg, and M. Markowska (2019). Weird inflects but OK: making sense of morphological generation errors. In CoNLL, 140-151.
K. Gorman and R. Sproat (2016). Minimally supervised number normalization. Transactions of the Association for Computational Linguistics 4: 507-519.
J. L. Lee, L. F.E. Ashby, M. E. Garza, Y. Lee-Sikka, S. Miller, A. Wong, A. D. McCarthy, and K. Gorman (under review). Massively multilingual pronunciation mining with WikiPron.
S. Pinker and A. Prince (1988). On language and connectionism: analysis of a parallel distributed processing model of language acquisition. Cognition 28(1–2):73–193.
R. Sproat (1992). Morphology and computation. Cambridge: MIT Press.

Action, not ritual

It is achingly apparent that an overwhelming amount of research in speech and language technologies considers exactly one human language: English. This is done so unthinkingly that some researchers seem to see the use of English data (and only English) as obvious, so obvious as to require no comment. This is unfortunate in part because English is, typologically speaking, a bit of an outlier. For instance, it has uncommonly impoverished inflectional morphology, a particularly rigid word order, and rather large vowel inventory. It is not hard to imagine how lessons learned designing for—or evaluating on—English data might not generalize to the rest of the world’s languages. In an influential paper, Bender (2009) encourages researchers to be more explicit about the languages studied, and this, framed as an imperative, is has come to be called the Bender Rule.

This “rule”, and the aforementioned observations underlying it, have taken on an almost mythical interpretation. They can easily be seen as a ritual granting the authors a dispensation to continue their monolingual English research. But this is a mistake. English hegemony is not merely bad science, nor is it a mere scientific inconvenience—a threat to validity.

It is no accident of history that the scientific world is in some sense an English colony. Perhaps you live in a country that owes an enormous debt to a foreign bank, and the bankers are demanding cuts to social services or reduction of tariffs: then there’s an excellent chance the bankers’ first language is English and that your first language is something else. Or maybe, fleeing the chaos of austerity and intervention, you find yourself and your children in cages in a foreign land: chances are you in Yankee hands. And, it is no accident that the first large-scale treebank is a corpus of English rather than of Delaware or Nahuatl or Powhatan or even Spanish, nor that the entire boondoggle was paid for by the largest military apparatus the world has ever known.

Such material facts respond to just one thing: concrete actions. Rituals, indulgences, or dispensations will not do. We must not confuse the act of perceiving and naming the hegemon with the far more challenging act of actually combating it. It is tempting to see the material conditions dualistically, as a sin we can never fully cleanse ourselves of. But they are the past and a more equitable world is only to be found in the future, a future of our own creation. It is imperative that we—as a community of scientists—take  steps to build the future we want.

References

Bender, Emily M. 2009. Linguistically naïve != language independent: why NLP needs linguistic typology. In EACL Workshop on the Interaction Between Linguistics and Computational Linguistics, pages 26-32.

Using a fixed training-development-test split in sklearn

The scikit-learn machine learning library has good support for various forms of model selection and hyperparameter tuning. For setting regularization hyperparameters, there are model-specific cross-validation tools, and there are also tools for both grid (e.g., exhaustive) hyperparameter tuning with the sklearn.model_selection.GridSearchCV and random hyperparameter tuning (in the sense of Bergstra & Bengio 2012) with sklearn.model_selection.RandomizedSearchCV, respectively. While you could probably could implement these yourself, the sklearn developers have enabled just about every feature you could want, including multiprocessing support.

One apparent limitation of these classes is that, as their names suggest, they are designed for use in a cross-validation setting. In the speech & language technology, however, standard practice is to use a fixed partition of the data into training, development (i.e., validation), and test (i.e., evaluation) sets, and to select hyperparameters which maximize performance on the development set. This is in part an artifact of limited computing resources of the Penn Treebank era and I’ve long suspected it has serious repercussions for model evaluation. But tuning and evaluating with a standard split is faster than cross-validation and can make exact replication much easier. And, there are also some concerns about whether cross-validation is the best way to set hyperparameters anyways. So what can we do?

The GridSearchCV and RandomSearchCV classes take an optional cv keyword argument, which can be, among other things, an object implementing the cross-validation iterator interface. At first I thought I would create an object which allowed me to use a fixed development set for hyperparameter tuning, but then I realized that I could do this with one of the existing iterator classes, namely one called sklearn.model_selection.PredefinedSplit. The constructor for this class takes a single argument test_fold, an array of integers of the same size as the data passed to the fitting method.  As the documentation explains “…when using a validation set, set the test_fold to 0 for all samples that are part of the validation set, and to -1 for all other samples.” That we can do. Suppose that we have training data x_train and y_train and development data x_dev and y_dev laid out as NumPy arrays. We then create a training-and-development set like so:

x = numpy.concatenate([x_train, x_dev])
y = numpy.concatenate([y_train, y_dev])

Then, we create the iterator object:

test_fold = numpy.concatenate([
    # The training data.
    numpy.full(-1, x_train.shape[1], dtype=numpy.int8),
    # The development data.
    numpy.zeros(x_dev.shape[1], dtype=numpy.int8)
])
cv = sklearn.model_selection.PredefinedSplit(test_fold)

Finally, we provide cv as a keyword argument to the grid or random search constructor, and then train. For instance, similar to this example we might do something like:

base = sklearn.ensemble.RandomForestClassifier()
grid = {"bootstrap": [True, False], 
        "max_features": [1, 3, 5, 7, 9, 10]}
model = sklearn.model_select.GridSearchCV(base, grid, cv=cv)
model.fit(x, y)

Now just add n_jobs=-1 to the constructor for model and to spread the work across all your logical cores.

References

Bergstra, J., and Bengio, Y. 2012. Random search for hyperparameter optimization. Journal of Machine Learning Research 13: 281-305.

arXiv vs. LingBuzz

In the natural language processing community, there has been a bit of kerfuffle about the ACL preprint policy, which essentially prevents you from submitting a manuscript to preprint aggregation websites like arXiv when the m.s. is also under review for a conference. I personally think this is a good policy: double blind review is really important for fairness. This lead me to reflect a bit on the outsized role that arXiv plays in natural language processing research. It is interesting to contrast arXiv with LingBuzz, a preprint aggregator for formal linguistics research.1 arXiv is visually ugly and cluttered, expensive (it somehow takes over $800,000 from Simons Foundations’ money to run it every year), and submissions tare subject to detailed, strict, carefully enforced editorial guidelines. In contrast, LingBuzz has a minimalistic text interface, is run and operated by a single professor (Michael Starke at the University of Tromsø), and the editorial guidelines are simple (they fit on a single page) and laxily enforced (mostly after the fact). Despite the laissez-faire attitude at LingBuzz, it has seen some rather contentious debates involving the usual trollish suspects (Postal, Everett, Behme, etc.) but it managed to keep things under control. But what I really love about LingBuzz is that unlike arXiv, no linguist is under the impression that it is any sort of substitute for peer review, or that authors need to know about (and cite) late-breaking work only available on LingBuzz. I think NLP researchers should take a hint from this and stop pretending arXiv is a reasonable alternative to peer review.

Endnotes

1. There are a few other such repositories. The Rutgers Optimality Archive (ROA) was once a popular repository for pre-prints of Optimality Theory work, but its contents are re-syndicated on LingBuzz and Optimality Theory is largely dead anyways. There is also the Semantics Archive.

Text encoding issues in Universal Dependencies

Do you know why the following comparison (in Python 3.7) fails?

>>> s1 = "ड़"
>>> s2 = "ड़"
>>> s1 == s2
False

I’ll give you a hint:

>>> len(s1)
1
>>> len(s2)
2

Despite the two strings rendering identically, they are encoded differently. The string s1 is a single-codepoint sequence, whereas s2 contains two codepoints. Thus string comparison fails, whether it’s done at the level of bytes or of Unicode codepoints.

Some NLP researchers are aware of issues arising from faulty string encoding. Eckhart de Castilho (2016), for example, describes a tool which automatically identifies misencoded pre-trained data, whereas Wu & Yarowsky (2018) report issues using an existing tool for transliteration on certain languages because of encoding issues. However, I suspect that far fewer NLP researchers are familiar with the aforementioned problem, which is specific to Unicode normalization. To put it simply, Unicode defines four normalization forms (and associated conversion algorithms) for strings, and the key distinction is between “composed” and “decomposed” forms of characters (using that term in a pretheoretic sense). The string s1 is composed into a single Unicode codepoint; s2 is decomposed into two.

Unfortunately, three columns of the Hindi Dependency Treebank (hi_hdtb, commit 54c4c0f; Bhat et al. 2017, Palmer et al. 2009) have a chaotic mix of composed and decomposed representations. It seems most if not all of these have to do with the encoding of the six nuqta (‘dot’) consonants, which are usually found in borrowings from Arabic or Persian (via Urdu, presumably). In Devangari these consonants are written by adding a dot to a phonetically similar native consonant; for instance ड [ɖə] plus the nuqta produces ड़ [ɽə]. As is usually the case in Unicode, there is more than one way to do it: you can either encode ड़ with a composed character (U+095C DEVANAGARI LETTER DDDHA) or with the native Devangari character (U+O921 DEVANAGARI LETTER DDA) plus a combining character (U+093C DEVANAGARI SIGN NUKTA). In practical terms, this means that strings containing diferent encodings of <ṛa> (as it is sometimes transliterated) will be treated as totally separate during training and evaluation, except on the off chance that all associated tools perform Unicode normalization ahead of time.

This does have negative consequences for NLP. Consider the UDPipe system (Straka & Straková 2017) at the CoNLL 2017 shared task on dependency parsing (Zeman et al. 2017), for which the primary metric is labeled attachment score (LAS). I first attempted to replicate the UDPipe results for the Hindi Dependency Treebank. Using UDPipe 1.2.0, word2vec (commit 20c129a), the hyperparameters given in the authors’ supplementary materials, and the official evaluation script, I obtain LAS = 87.09 on the “gold tokenization” subtask. However I can improve this simply by converting the training, development, and test data to a consistent normalization like so:

for FILE in *.conllu; do
    TMPFILE="$(mktemp)"
    uconv -x nfkc "${FILE}" > "${TMPFILE}"
    mv "${TMPFILE}" "${FILE}"
done

and then retraining. Here I have chosen to apply the NFKC (“compatibility composed”) normalization form. While Zeman et al. do not discuss the encoding of the labeled Universal Dependencies data, they do mention that they apply NFKC normalization to the addditional raw data. But it doesn’t really matter in this case which you choose so long as you are consistent. After retraining, I obtain LAS = 87.38, or .29 points for free. I also ran an “mismatch” experiment, where the training and testing data have different normalization forms; naturally, this causes a slight degradation to LAS = 86.98.

Straka & Straková (2017) report a separate set of experiments in which they have attempted to rebalance the training-development-test splits. Just to be sure, I repeated the above experiments using their original rebalancing script. With the baseline—mixed normalization—data, I can replicate their result exactly: LAS = 87.30. With a consistent NFKC normalization of training, development and test data, I get LAS = 87.50. And with a normalization mismatch between training and test data, I get LAS = 87.07, a slight degradation. And the improvements are more or less for free.

While I have not yet done a consistent audit, I found three other UD treebanks that have encoding issues. The ar_padt treebank has a non-canonical ordering of combining characters in the lemma column (the shaddah, which indicates geminates, should come before the fathah and not the other way around), but this is unlikely to have any major effect on model performance because it uses this non-canonical ordering consistently. The ko_kaist and ur_udtb treebanks also have minor inconsistencies.

Unfortunately my corporate overlord doesn’t permit me to file a pull request here because of the Hindi data is released under a CC BY-NC-SA license. But if you’re not so constrained, feel free to do so, and ping this thread once you have! And pay attention in the future.

References

Bhat, R. A., Bhatt, R., Farudi, A., Klassen, P., Narasimhan, B., Palmer, M., Rambow, O., Sharma, D. M., Vaidya, A., Vishnu, S. R., and Xia, F. 2017. The Hindi/Urdu Treebank Project. In Ide., N., and Pustejovsky, J. (ed.), The Handbook of Linguistic Annotation, pages 659-698. Springer.
Eckhart de Castilho, R. 2016. Automatic analysis of flaws in pre-trained NLP models. In 3rd International Workshop on Worldwide Language Service Infrastructure and 2nd Workshop on Open Infrastructures and Analysis Frameworks for Human Language Technologies, pages 19-27.
Palmer, M., Bhatt, R., Narasimhan, B., Rambow, O., Sharma, D. M., and Xia, F. 2009. Hindi syntax: Annotation dependency, lexical predicate-argument structure, and phrase structure. In ICON, pages 14-17.
Straka, M., and Straková, J. 2017. Tokenizing, POS tagging, lemmatizing and parsing UD 2.0 with UDPipe. In CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies, pages 88-99.
Wu, W. and Yarowsky, D. 2018. A comparative study of extremely low-resource transliteration of the world’s languages. In LREC, pages 938-943.
Zeman, D., Popel, M., Straka, M., Hajič, J., Nivre, J., Ginter, F., … and Li, J. 2017. CoNLL Shared Task: Multilingual parsing from raw text to Universal Dependencies. In CoNLL 2017 Shared Task: Multilingual Parsing from Raw Text to Universal Dependencies, pages 1-19.

Lessons learned from my time at Google

  • C++ 11 is a powerful, elegant language and the right choice for performant general-purpose code. Bash is an excellent lingua franca for chaining a long series of commands. Python is best for everything else.
  • Data should be passed around in schematic form, with a compact serializations over the wire and a human-readable format at rest. Protocol buffers (and the lesser-known text format) are an ideal cross-language solution.
  • Grammar development is more important than model building.
  • Model building is easier than deployment.
  • Whiteboards are useful.
  • I can only do certain sorts of work without an office (yes, that thing with a door).

A minimalist project design for NLP

Let’s say you want to build a new tagger, a new named entity recognizer, a new dependency parser, or whatever. Or perhaps you just want to see how your coreference resolution engine performs on your new database of anime reviews. So how should you structure your project? Here’s my minimalist solution.

There are two principles that guide my design. The first one is modularity. Some of these components will get run many times, some won’t. If you’re doing model comparison—and you should be doing model comparison—some components will get swapped out with someone else’s code. This sort of thing is a major lift unless you opt for modularity. The second principle is filesystem state. The filesystem is your friend. If your embedding table eats up all your RAM and you have to restart, the filesystem will be in roughly the same state as when you left. The filesystem allows you to organize things into directories and subdirectories, and give the pieces informative names; I like to record information about datasets and hyperparameter values in my file and directory names. So without further ado, here are the recommended scripts or applications to create when you’re starting off on a new project.

  1. split takes the full dataset and a random seed (which you should store for later) as input. The script reads the data in, randomly shuffles the data, and then splits it into an 80% training set, 10% development set, and a 10% test (i.e., evaluation set) which it then outptus. If you’re comparing to prior work that used a “standard split” you may want to have a separate script that generates that too, but I strongly recommend using randomly generated splits.
  2. train takes the training set as input and outputs a model file or directory. If you’re automating hyperparameter tuning you will also want to provide the development set as input; if not you will probably want to either add a bunch of flags to control the hyperparameters or allow the user to pass some kind of model configuration file (I like YAML for this).
  3. apply takes as input the model file(s) produced in (2) and the test set, and applies the model to the data, outputting a new hypothesized test data set (i.e., the model’s predictions). One open question is whether this ought to take only unlabeled data or should overwrite the existing labels: it depends.
  4. evaluate takes as input the gold test set and the hypothesized test data set generated in (3) and outputs the evaluation results (as text or in some structured data format—sometimes YAML is a good choice, other times TSV files will do). I recommend you test this with a small amount of data first.

That’s all there’s to it. When you begin doing model comparison you may find yourself swapping out (2-3) for somebody else’s code, but make sure to still stick to the same evaluation script.