PASTRIE: A Corpus of Prepositions Annotated with Supersense Tags in Reddit International English

Related tags

Deep Learningpastrie
Overview

PASTRIE

CC BY-SA 4.0

Official release of the corpus described in the paper:

Michael Kranzlein, Emma Manning, Siyao Peng, Shira Wein, Aryaman Arora, and Nathan Schneider (2020). PASTRIE: A Corpus of Prepositions Annotated with Supersense Tags in Reddit International English [link]. Proceedings of the 14th Linguistic Annotation Workshop.


Overview

PASTRIE is a corpus of English data from Reddit annotated with preposition supersenses from the SNACS inventory.

While the data in PASTRIE is in English, it was produced by presumed speakers of four L1s:

  • English
  • French
  • German
  • Spanish

For details on how L1s were identified, see section 3.1 of Rabinovich et al. (2018).

Annotation Example

Below is an example sentence from the corpus, where annotation targets are bolded and preposition supersenses are annotated with the notation SceneRole↝Function. Together, a scene role and function are known as a construal.


Data Formats

PASTRIE is released in the following formats. We expect that most projects will be best served by one of the JSON formats.

  • .conllulex: the 19-column CoNLL-U-Lex format originally used for STREUSLE.
  • .json: a JSON representation of the CoNLL-U-Lex that does not require a CoNLL-U-Lex parser.
  • .govobj.json: an extended version of the JSON representation that contains information about each preposition's syntactic parent and object.

PASTRIE mostly follows STREUSLE with respect to the data format and SNACS annotation practice. Primary differences in the annotations are:

  • Lemmas, part-of-speech tags, and syntactic dependencies aim to follow the UD standard in both cases. They are gold in STREUSLE, versus automatic with some manual corrections in PASTRIE.
    • PASTRIE does not group together base+clitic combinations, whereas STREUSLE does (multiword tokens—where a single orthographic word contains multiple syntactic words).
    • PASTRIE does not regularly specify SpaceAfter=No to indicate alignment between the tokens and the raw text.
    • In PASTRIE, the raw text string accompanying the sentence may contain two or more consecutive spaces.
    • PASTRIE lacks enhanced dependencies.
  • Multiword expression annotations in PASTRIE are limited to expressions containing a preposition. Depending on the syntactic head, the expression may or may not have a SNACS supersense.
    • Verbal multiword expressions in PASTRIE are not subtyped in the lexcat; they all have a lexcat of V.
  • Noun and verb expressions in PASTRIE do not have supersense labels.
Comments
  • Misc. annotation errors and/or conversion script bugs

    Misc. annotation errors and/or conversion script bugs

    There are some annotations which I'm fairly sure are incorrect and are choking up the JSON conversion script. (These errors occur using the unmodified versions of all scripts taken straight from STRUESLE.) One or two might also be indicative of a bug in the conllulex2json.py file.

    1. vs mistagged as a noun--should be prep

    AssertionError: ('french-fad32caf-e595-e3cb-07bf-aaea891e53cb-02', {'lexlemma': 'versus', 'lexcat': 'CCONJ', 'ss': 'c', 'ss2': 'c', 'toknums': [3]}, {'#': 3, 'word': 'vs', 'lemma': 'versus', 'upos': 'NOUN', 'xpos': 'NN', 'feats': None, 'head': 8, 'deprel': 'nsubj', 'edeps': None, 'misc': None, 'smwe': None, 'wmwe': None, 'lextag': 'O-CCONJ-`c'})

    1. ditto

    AssertionError: ('french-fad32caf-e595-e3cb-07bf-aaea891e53cb-02', {'lexlemma': 'versus', 'lexcat': 'CCONJ', 'ss': 'c', 'ss2': 'c', 'toknums': [3]}, {'#': 3, 'word': 'vs', 'lemma': 'versus', 'upos': 'NOUN', 'xpos': 'NN', 'feats': None, 'head': 8, 'deprel': 'nsubj', 'edeps': None, 'misc': None, 'smwe': None, 'wmwe': None, 'lextag': 'O-CCONJ-`c'})

    1. Script complains about "to" in this snippet at ID=23. Not immediately clear to me what the issue is--perhaps that "to" is labeled ADP/IN? For its xpos I think it ought to be TO, not sure about its upos. Snippet:
    13      shit    shit    NOUN    NN      _       16      obl:npmod       _       _       _       _       _       _       _       _       _       _       _
    14      this    this    PRON    DT      _       16      nsubj   _       _       _       _       _       _       _       _       _       _       _
    15      can     can     AUX     MD      _       16      aux     _       _       _       _       _       _       _       _       _       _       _
    16      end     end     VERB    VB      _       4       parataxis       _       _       _       _       _       _       _       _       _       _       _
    17      right   right   ADV     RB      _       18      advmod  _       _       _       _       _       _       _       _       _       _       _
    18      now     now     ADV     RB      _       16      advmod  _       _       _       _       _       _       _       _       _       _       _
    19      if      if      SCONJ   IN      _       21      mark    _       _       _       _       _       _       _       _       _       _       _
    20      I       I       PRON    PRP     _       21      nsubj   _       _       _       _       _       _       _       _       _       _       _
    21      want    want    VERB    VBP     _       16      advcl   _       _       _       _       _       _       _       _       _       _       _
    22      it      it      PRON    PRP     _       21      obj     _       _       _       _       _       _       _       _       _       _       _
    23      to      to      ADP     IN      _       21      obl     _       _       _       _       _       `i      `i      _       _       _       _
    24      .       .       PUNCT   .       _       4       punct   _       _       _       _       _       _       _       _       _       _       _
    

    Error:

    AssertionError: ('french-a17a4340-f9c0-8fef-fa1b-1bf13879399b-02', {'lexlemma': 'to', 'lexcat': 'INF', 'ss': 'i', 'ss2': 'i', 'toknums': [23]}, {'#': 23, 'word': 'to', 'lemma': 'to', 'upos': 'ADP', 'xpos': 'IN', 'feats': None, 'head': 21, 'deprel': 'obl', 'edeps': None, 'misc': None, 'smwe': None, 'wmwe': None, 'lextag': 'O-INF-`i'})

    Relevant span of code:

                if validate_pos and upos!=lc and (upos,lc) not in {('NOUN','N'),('PROPN','N'),('VERB','V'),
                    ('ADP','P'),('ADV','P'),('SCONJ','P'),
                    ('ADP','DISC'),('ADV','DISC'),('SCONJ','DISC'),
                    ('PART','POSS')}:
                    # most often, the single-word lexcat should match its upos
                    # check a list of exceptions
                    mismatchOK = False
                    if xpos=='TO' and lc.startswith('INF'):
                        mismatchOK = True
                    elif (xpos=='TO')!=lc.startswith('INF'):
                        assert upos in ['SCONJ', "ADP"] and swe['lexlemma']=='for',(sent['sent_id'],swe,tok)
                        mismatchOK = True
    
    1. Originator as function:

    (in french-c02823ec-60bd-adce-7327-01337eb9d1c8-02) AssertionError: ('p.Originator should never be function', {'lexlemma': 'you', 'lexcat': 'PRON.POSS', 'ss': 'p.Originator', 'ss2': 'p.Originator', 'toknums': [1]})

    1. lexcat DISC with ADJ:

    AssertionError: In spanish-a25e8289-e04a-f5af-ce56-ead9faca65b1-02, single-word expression 'like' has lexcat DISC, which is incompatible with its upos ADJ

    1. "her" tagged with Possessor is incorrectly parsed as iobj and tagged as PRP instead of PRP$. Relevant snippet:
    1       My      my      PRON    PRP$    _       2       nmod:poss       _       _       _       _       _       SocialRel       Gestalt _       _       _       _
    2       grandma grandma NOUN    NN      _       3       nsubj   _       _       _       _       _       _       _       _       _       _       _
    3       had     have    VERB    VBD     _       0       root    _       _       _       _       _       _       _       _       _       _       _
    4       her     she     PRON    PRP     _       3       iobj    _       _       _       _       _       Possessor       Possessor       _       _       _       _
    5       super   super   ADV     RB      _       6       advmod  _       _       _       _       _       _       _       _       _       _       _
    6       thick   thick   ADJ     JJ      _       8       amod    _       _       _       _       _       _       _       _       _       _       _
    7       floor   floor   NOUN    NN      _       8       compound        _       _       _       _       _       _       _       _       _       _       _
    8       mats    mat     NOUN    NNS     _       3       obj     _       _       _       _       _       _       _       _       _       _       _
    9       *       *       PUNCT   NFP     _       8       punct   _       _       _       _       _       _       _       _       _       _       _
    10      over    over    ADP     IN      _       13      case    _       _       _       _       _       Locus   Locus   _       _       _       _
    11      *       *       PUNCT   NFP     _       13      punct   _       _       _       _       _       _       _       _       _       _       _
    12      the     the     DET     DT      _       13      det     _       _       _       _       _       _       _       _       _       _       _
    13      accelerator     accelerator     NOUN    NN      _       3       obl     _       _       _       _       _       _       _       _       _       _       _
    14      ,       ,       PUNCT   ,       _       3       punct   _       _       _       _       _       _       _       _       _       _       _
    

    Error:

    AssertionError: In spanish-ebba3c73-2431-c216-8f4d-d469ee8d5564-01, single-word expression 'her' has lexcat P, which is incompatible with its upos PRON

    1. "NA" is misannotated--this is NA as in North America, i.e. a PROPN/NP, but it's lemmatized as "no", and its tags are weird.

    AssertionError: ('german-35000895-1d78-c18a-01ed-f7410b9c0581-01', {'lexlemma': 'no', 'lexcat': 'ADV', 'ss': None, 'ss2': None, 'toknums': [5]}, {'#': 5, 'word': 'NA', 'lemma': 'no', 'upos': 'PART', 'xpos': 'TO', 'feats': None, 'head': 6, 'deprel': 'mark', 'edeps': None, 'misc': None, 'smwe': None, 'wmwe': None, 'lextag': 'O-ADV'})

    opened by lgessler 6
  • Prepositional supersense annotations on non-preposition targets

    Prepositional supersense annotations on non-preposition targets

    Is it OK for a verb-headed SMWE to have a prepositional supersense? The validator complains about it. Offending SMWE:

    21	give	give	VERB	VB	_	10	conj	_	_	2:1	_	give up on	p.Theme	p.Theme	_	_	_	_
    22	up	up	ADP	RP	_	21	compound:prt	_	_	2:2	_	_	_	_	_	_	_	_
    23	on	on	ADP	IN	_	24	case	_	_	2:3	_	_	_	_	_	_	_	_
    
    opened by lgessler 5
  • Prepositions unannotated for supersense

    Prepositions unannotated for supersense

    Token 6:

    # sent_id = french-f57dd6ab-5263-4c8a-e360-8ec683e6a37a-02
    # text = Once you have the hang of it it s pretty fast ( and does n't eat your clutch ) .
    1	Once	once	SCONJ	IN	_	3	mark	_	_	_	_	_	_	_	_	_	_	_
    2	you	you	PRON	PRP	_	3	nsubj	_	_	_	_	_	_	_	_	_	_	_
    3	have	have	VERB	VBP	_	11	advcl	_	_	_	_	_	_	_	_	_	_	_
    4	the	the	DET	DT	_	5	det	_	_	_	_	_	_	_	_	_	_	_
    5	hang	hang	NOUN	NN	_	3	obj	_	_	_	_	_	_	_	_	_	_	_
    6	of	of	ADP	IN	_	7	case	_	_	_	_	_	_	_	_	_	_	_
    7	it	it	PRON	PRP	_	5	nmod	_	_	_	_	_	_	_	_	_	_	_
    8	it	it	PRON	PRP	_	11	nsubj	_	_	_	_	_	_	_	_	_	_	_
    9	s	be	AUX	VBZ	_	11	cop	_	_	_	_	_	_	_	_	_	_	_
    10	pretty	pretty	ADV	RB	_	11	advmod	_	_	_	_	_	_	_	_	_	_	_
    11	fast	fast	ADJ	JJ	_	0	root	_	_	_	_	_	_	_	_	_	_	_
    12	(	(	PUNCT	-LRB-	_	16	punct	_	_	_	_	_	_	_	_	_	_	_
    13	and	and	CCONJ	CC	_	16	cc	_	_	_	_	_	_	_	_	_	_	_
    14	does	do	AUX	VBZ	_	16	aux	_	_	_	_	_	_	_	_	_	_	_
    15	n't	not	PART	RB	_	16	advmod	_	_	_	_	_	_	_	_	_	_	_
    16	eat	eat	VERB	VB	_	11	conj	_	_	_	_	_	_	_	_	_	_	_
    17	your	you	PRON	PRP$	_	18	nmod:poss	_	_	_	_	_	Possessor	Possessor	_	_	_	_
    18	clutch	clutch	NOUN	NN	_	16	obj	_	_	_	_	_	_	_	_	_	_	_
    19	)	)	PUNCT	-RRB-	_	11	punct	_	_	_	_	_	_	_	_	_	_	_
    20	.	.	PUNCT	.	_	11	punct	_	_	_	_	_	_	_	_	_	_	_
    

    I assumed that all preps were supposed to be annotated, but perhaps not?

    opened by lgessler 3
  • Apostrophes removed in preprocessing?

    Apostrophes removed in preprocessing?

    Looking through the data, there are a LOT of sentences where clitics are tokenized off but lack an apostrophe. Is that just the genre or did they get lost in preprocessing?

    opened by nschneid 2
  • Dataset requested

    Dataset requested

    Hi all,

    I would like to request the PASTRIE dataset accompanying the paper "PASTRIE: A Corpus of Prepositions Annotated with Supsersense Tags in Reddit International English".

    Thanks for reply.

    opened by fj-morales 2
  • SNACS supersense tags should start with

    SNACS supersense tags should start with "p."

    For compatibility with STREUSLE, it should be p.Locus, p.Theme, etc.

    Special labels like `i `d `c `$ ?? should not start with p.. In fact, the backtick labels from annotation are not represented as such in STREUSLE—they are reflected in the LEXCAT column of the data.

    opened by nschneid 0
  • Questionable adpositional MWEs

    Questionable adpositional MWEs

    • in_male_term — from "in male terms"; should be in_term (at most)
    • in_the_first_place
    • in_my_hand — from "in my hands"; should be in_hand (at most)
    • for_quite_some_time — just Duration for, weak MWE?
    • at_all_time — from what should have been "at all times". OK?
    • on_a_smaller_scale — omit adjective?
    • withouth — typo
    • see_as — "seeing as" (deverbal MWE acting like a preposition)
    opened by nschneid 0
  • Some undersegmentation of sentences

    Some undersegmentation of sentences

    Despite manual editing there are still places where a long sentence ought to be split up (esp. where it consists of a blockquoted sentence with > followed by a response). Looking for multiple consecutive spaces in the raw text uncovers some of these (as well as some discourse appendages like emoticons, which should probably remain in the same UD sentence).

    It would be nice to write a script to help clean these up—the tricky part is updating offsets in each parse.

    opened by nschneid 0
Releases(v2.0.1)
  • v2.0.1(Nov 21, 2021)

  • v2.0(Oct 22, 2021)

    • Switch to full .conllulex format following STREUSLE
      • add lexcats (#3), morphological features, newdoc directives
    • Scripts for validation and format conversion
    • Clean up various annotation issues, including:
      • restore apostrophes and fixing other conversion problems (#6, #9)
      • include pretokenized raw text (#12)
    Source code(tar.gz)
    Source code(zip)
  • v1.0.1(Dec 14, 2020)

    • Added .json file format
    • Switched lemmatization and pos tagging from StanfordNLP 0.2.0 to Stanza 1.1.1
    • Corrected rare encoding issue from v1.0
    Source code(tar.gz)
    Source code(zip)
Owner
NERT @ Georgetown
NERT @ Georgetown
Implementation of the master's thesis "Temporal copying and local hallucination for video inpainting".

Temporal copying and local hallucination for video inpainting This repository contains the implementation of my master's thesis "Temporal copying and

David Álvarez de la Torre 1 Dec 02, 2022
Pytorch implementation for Semantic Segmentation/Scene Parsing on MIT ADE20K dataset

Semantic Segmentation on MIT ADE20K dataset in PyTorch This is a PyTorch implementation of semantic segmentation models on MIT ADE20K scene parsing da

MIT CSAIL Computer Vision 4.5k Jan 08, 2023
Official PyTorch implementation of the paper "Deep Constrained Least Squares for Blind Image Super-Resolution", CVPR 2022.

Deep Constrained Least Squares for Blind Image Super-Resolution [Paper] This is the official implementation of 'Deep Constrained Least Squares for Bli

MEGVII Research 141 Dec 30, 2022
Official PyTorch implementation of "Preemptive Image Robustification for Protecting Users against Man-in-the-Middle Adversarial Attacks" (AAAI 2022)

Preemptive Image Robustification for Protecting Users against Man-in-the-Middle Adversarial Attacks This is the code for reproducing the results of th

2 Dec 27, 2021
DaReCzech is a dataset for text relevance ranking in Czech

Dataset DaReCzech is a dataset for text relevance ranking in Czech. The dataset consists of more than 1.6M annotated query-documents pairs,

Seznam.cz a.s. 8 Jul 26, 2022
Train a deep learning net with OpenStreetMap features and satellite imagery.

DeepOSM Classify roads and features in satellite imagery, by training neural networks with OpenStreetMap (OSM) data. DeepOSM can: Download a chunk of

TrailBehind, Inc. 1.3k Nov 24, 2022
Parallel Latent Tree-Induction for Faster Sequence Encoding

FastTrees This repository contains the experimental code supporting the FastTrees paper by Bill Pung. Software Requirements Python 3.6, NLTK and PyTor

Bill Pung 4 Mar 29, 2022
A modification of Daniel Russell's notebook merged with Katherine Crowson's hq-skip-net changes

Edits made to this repo by Katherine Crowson I have added several features to this repository for use in creating higher quality generative art (featu

Paul Fishwick 10 May 07, 2022
A PyTorch-based open-source framework that provides methods for improving the weakly annotated data and allows researchers to efficiently develop and compare their own methods.

Knodle (Knowledge-supervised Deep Learning Framework) - a new framework for weak supervision with neural networks. It provides a modularization for se

93 Nov 06, 2022
Use of Attention Gates in a Convolutional Neural Network / Medical Image Classification and Segmentation

Attention Gated Networks (Image Classification & Segmentation) Pytorch implementation of attention gates used in U-Net and VGG-16 models. The framewor

Ozan Oktay 1.6k Dec 30, 2022
Data Preparation, Processing, and Visualization for MoVi Data

MoVi-Toolbox Data Preparation, Processing, and Visualization for MoVi Data, https://www.biomotionlab.ca/movi/ MoVi is a large multipurpose dataset of

Saeed Ghorbani 51 Nov 27, 2022
Create images and texts with the First Order Generative Adversarial Networks

First Order Divergence for training GANs This repository contains code accompanying the paper First Order Generative Advesarial Netoworks The majority

Zalando Research 35 Dec 11, 2021
This is an official implementation of the High-Resolution Transformer for Dense Prediction.

High-Resolution Transformer for Dense Prediction Introduction This is the official implementation of High-Resolution Transformer (HRT). We present a H

HRNet 403 Dec 13, 2022
Implements Stacked-RNN in numpy and torch with manual forward and backward functions

Recurrent Neural Networks Implements simple recurrent network and a stacked recurrent network in numpy and torch respectively. Both flavours implement

Vishal R 1 Nov 16, 2021
Open source code for Paper "A Co-Interactive Transformer for Joint Slot Filling and Intent Detection"

A Co-Interactive Transformer for Joint Slot Filling and Intent Detection This repository contains the PyTorch implementation of the paper: A Co-Intera

67 Dec 05, 2022
Orbivator AI - To Determine which features of data (measurements) are most important for diagnosing breast cancer and find out if breast cancer occurs or not.

Orbivator_AI Breast Cancer Wisconsin (Diagnostic) GOAL To Determine which features of data (measurements) are most important for diagnosing breast can

anurag kumar singh 1 Jan 02, 2022
BDDM: Bilateral Denoising Diffusion Models for Fast and High-Quality Speech Synthesis

Bilateral Denoising Diffusion Models (BDDMs) This is the official PyTorch implementation of the following paper: BDDM: BILATERAL DENOISING DIFFUSION M

172 Dec 23, 2022
PyTorch code for: Learning to Generate Grounded Visual Captions without Localization Supervision

Learning to Generate Grounded Visual Captions without Localization Supervision This is the PyTorch implementation of our paper: Learning to Generate G

Chih-Yao Ma 41 Nov 17, 2022
U^2-Net - Portrait matting This repository explores possibilities of using the original u^2-net model for portrait matting.

U^2-Net - Portrait matting This repository explores possibilities of using the original u^2-net model for portrait matting.

Dennis Bappert 104 Nov 25, 2022
DCA - Official Python implementation of Delaunay Component Analysis algorithm

Delaunay Component Analysis (DCA) Official Python implementation of the Delaunay

Petra Poklukar 9 Sep 06, 2022