SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1804.07656 · 2018

Acquisition of Phrase Correspondences using Natural Deduction Proofs

arXiv · PDF · Open in the Atlas

Code that ran

12 of 14 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 12 have no contract verdict; 0 failures of our instrument; 2 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1804.07656.json · badge: /api/ran/1804.07656.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
mynlp/ccg2lambda canonical (classifier label)
classifier basis: census
12 ran · 2 no recorded run
FunctionStatusWhere it lives
GetPremisesFromNode_ Ran
ran on a synthesized input
mynlp/ccg2lambda/extract_jsem_problems.py
code served (permissive licence) · get_code("570b63a817883f3e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import xml.etree.ElementTree as ET

# Create a mock XML node structure
# The function iterates over node (so node must be iterable)
# It accesses child.tag, child.text, child.get(), and child.attrib
# We need to create Element objects with 'p' children

root = ET.Element('root')

# Create first child element with tag 'p'
child1 = ET.SubElement(root, 'p')
child1.text = '  First premise text  '
child1.set('idx', '2')

# Create second child element with tag 'p'
child2 = ET.SubElement(root, 'p')
child2.text = '  Second premise text  '
child2.set('idx', '1')

# Create a non-'p' child that should be skipped
child3 = ET.SubElement(root, 'div')
child3.text = 'This should be ignored'

# Create third 'p' child
child4 = ET.SubElement(root, 'p')
child4.text = '  Third premise text  '
child4.set('idx', '3')

ARGS = {
    'node': root
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
GetTextFromScriptNode Ran
ran on a synthesized input
mynlp/ccg2lambda/extract_jsem_problems.py
code served (permissive licence) · get_code("b8f1e22e03a0f287")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from xml.etree import ElementTree as ET

# Create mock XML elements that match the function's usage
# The function accesses script[0].tail and script.text

# Create a child element with a tail attribute
child_elem = ET.Element("child")
child_elem.tail = "  tail text  "

# Create a parent element with text attribute and append the child
parent_elem = ET.Element("parent")
parent_elem.text = "  parent text  "
parent_elem.append(child_elem)

# Test case 1: script is a list with elements (len > 0)
script_with_elements = [child_elem]

# Test case 2: script is a list without elements (len == 0)
# In this case, the function accesses script.text, so we need script to be the parent
script_empty = ET.Element("root")
script_empty.text = "  empty script text  "

# We'll use the first test case (non-empty list) as the main argument
ARGS = {
    "script": script_with_elements
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
assign_ids_to_nodes Ran
ran on a synthesized input
mynlp/ccg2lambda/en/candc2transccg.py
code served (permissive licence) · get_code("3a51bbe51a2bb268")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

class MockCCGNode:
    def __init__(self, children=None):
        self.children = children if children is not None else []
        self.attribs = {}
    
    def set(self, key, value):
        self.attribs[key] = value
    
    def __iter__(self):
        return iter(self.children)


# Create a small tree structure
leaf1 = MockCCGNode(children=[])
leaf2 = MockCCGNode(children=[])
internal_node = MockCCGNode(children=[leaf1, leaf2])
root = MockCCGNode(children=[internal_node])

ARGS = {
    'ccg_tree': root,
    'sentence_number': 0,
    'current': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
denormalize_category Ran
ran on a synthesized input
mynlp/ccg2lambda/en/easyccg2jigg.py
code served (permissive licence) · get_code("7e55aaf47ea15ebc")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function denormalize_category takes a category string and replaces
# '-lb-' with '(' and '-rb-' with ')'
# It expects a string parameter

category = "test-lb-content-rb-end"

ARGS = {
    "category": category
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
flatten_and_rename_nodes Ran
ran on a synthesized input
mynlp/ccg2lambda/en/candc2transccg.py
code served (permissive licence) · get_code("c56f548353afc2ed")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

class TreeNode:
    def __init__(self, tag=None):
        self.tag = tag
        self.children = []
    
    def __iter__(self):
        return iter(self.children)
    
    def append(self, child):
        self.children.append(child)


# Build a small tree structure
leaf1 = TreeNode('leaf1')
leaf2 = TreeNode('leaf2')
leaf3 = TreeNode('leaf3')

internal1 = TreeNode('internal1')
internal1.append(leaf1)
internal1.append(leaf2)

root = TreeNode('root')
root.append(internal1)
root.append(leaf3)

ARGS = {
    'ccg_root': root
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_fracas_info Ran
ran on a synthesized input
mynlp/ccg2lambda/en/report_results.py
code served (permissive licence) · get_code("a5ded5178deeff0a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

def get_fracas_info(problem_id):
    """
    Given a problem ID, e.g. fracas_001_generalized_quantifiers,
    it returns the fracas number "001" and the section name
    "generalized_quantifiers".
    """
    info = re.findall('fracas_(\\d\\d\\d)_(.+)', problem_id)
    assert info, 'Correct Fracas ID information could not be retrieved for {0}'.format(problem_id)
    fracas_num, section = info[0]
    return (int(fracas_num), section)


ARGS = {
    'problem_id': 'fracas_001_generalized_quantifiers'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_nodes_by_tag Ran
ran on a synthesized input
mynlp/ccg2lambda/en/candc2transccg.py
code served (permissive licence) · get_code("0c0ad808ca6abea8")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

from xml.etree import ElementTree as ET

# Create a simple XML tree structure for testing
root = ET.Element("root")
root.set("id", "1")

child1 = ET.SubElement(root, "child")
child1.set("id", "2")

child2 = ET.SubElement(root, "target")
child2.set("id", "3")

grandchild1 = ET.SubElement(child1, "target")
grandchild1.set("id", "4")

grandchild2 = ET.SubElement(child1, "other")
grandchild2.set("id", "5")

great_grandchild = ET.SubElement(grandchild2, "target")
great_grandchild.set("id", "6")

ARGS = {
    "root": root,
    "tag": "target"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_text_from_script_Node Ran
ran on a synthesized input
mynlp/ccg2lambda/en/extract_entailment_problems.py
code served (permissive licence) · get_code("2ecf433430f8ac89")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. It ran in syntology-harvest-sandbox:v5-deps (sha256:6a292aa1aed1) at 2026-09-27T00:36:13Z. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
from xml.etree import ElementTree as ET

# Create a mock script element with children
# The function expects either:
# 1. len(script) > 0: access script[0].tail
# 2. else: access script.text

# Create an XML element to simulate the script parameter
parent_elem = ET.Element("parent")
child_elem = ET.SubElement(parent_elem, "child")
child_elem.tail = "  extracted text from tail  "
parent_elem.text = "  fallback text  "

# Test case 1: script has children (len > 0)
script_with_children = parent_elem

ARGS = {
    "script": script_with_children
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image syntology-harvest-sandbox:v5-deps (sha256:6a292aa1aed1); not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_answers Ran
ran on a synthesized input
mynlp/ccg2lambda/en/report_results.py
code served (permissive licence) · get_code("3008d57334981e1e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import os

# Create a temporary file with sample fracas answers
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_file.write("fracas_001_generalized_quantifiers 3 yes\n")
temp_file.write("fracas_002_generalized_quantifiers 2 no\n")
temp_file.write("fracas_003_implicatives 1 unknown\n")
temp_file.close()

ARGS = {
    'finput_filename': temp_file.name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
load_verbocean Ran
ran on a synthesized input
mynlp/ccg2lambda/en/verbocean_to_json.py
code served (permissive licence) · get_code("fa5a06efa76c1301")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import gzip
import os
from collections import defaultdict

# Create a temporary gzip file with sample VerbOcean data
temp_dir = tempfile.mkdtemp()
verbocean_file = os.path.join(temp_dir, "verbocean.txt.gz")

# Write sample VerbOcean data
with gzip.open(verbocean_file, 'wt', encoding='utf-8') as f:
    f.write("# VerbOcean sample data\n")
    f.write("run sprint [hypernym] 0.95\n")
    f.write("run jog [similar] 0.85\n")
    f.write("eat consume [hypernym] 0.92\n")
    f.write("eat devour [similar] 0.88\n")
    f.write("walk stride [hypernym] 0.90\n")

ARGS = {
    "verbocean_filename": verbocean_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
make_tree Ran
ran on a synthesized input
mynlp/ccg2lambda/en/easyccg2jigg.py
code served (permissive licence) · get_code("3d6eb8869685a2e0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tree. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import logging
import string
from nltk import Tree

def substitute_chars(sin):
    sout = []
    inside_tag = False
    prev_char = None
    for i, s in enumerate(sin):
        if s == '<' and prev_char == '(':
            inside_tag = True
            t = s
        elif s == '>' and (prev_char in string.punctuation or prev_char.isupper() or prev_char in string.digits or (prev_char == 'j')):
            inside_tag = False
            t = s
        elif inside_tag:
            if s == ' ':
                t = '|||'
            elif s == '(':
                t = '-lb-'
            elif s == ')':
                t = '-rb-'
            else:
                t = s
        else:
            t = s
        sout.append(t)
        prev_char = s
    return ''.join(sout)

def make_tree(line):
    tree_str = substitute_chars(line.strip())
    try:
        tree = Tree.fromstring(tree_str)
    except ValueError:
        tree = None
        logging.warning('Failed to Tree parse line: {0}'.format(line))
    return tree

# Valid NLTK parse tree string format: (ROOT (S (NP (DT The) (NN dog)) (VP (VBZ runs))))
valid_tree_line = "(S (NP (DT the) (NN dog)) (VP (VBZ runs)))\n"

ARGS = {
    "line": valid_tree_line
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tree — a Tree output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
substitute_chars Ran
ran on a synthesized input
mynlp/ccg2lambda/en/easyccg2jigg.py
code served (permissive licence) · get_code("e2d0be2cf9b3ae20")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned str. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import string

def substitute_chars(sin):
    sout = []
    inside_tag = False
    prev_char = None
    for i, s in enumerate(sin):
        if s == '<' and prev_char == '(':
            inside_tag = True
            t = s
        elif s == '>' and (prev_char in string.punctuation or prev_char.isupper() or prev_char in string.digits or (prev_char == 'j')):
            inside_tag = False
            t = s
        elif inside_tag:
            if s == ' ':
                t = '|||'
            elif s == '(':
                t = '-lb-'
            elif s == ')':
                t = '-rb-'
            else:
                t = s
        else:
            t = s
        sout.append(t)
        prev_char = s
    return ''.join(sout)

# Test inputs to understand the function behavior
test_cases = [
    "hello world",
    "(<tag>)",
    "(<OPEN>)",
    "(<123>)",
    "(<j>)",
    "(< inside >)",
    "normal text (<tag with spaces>) end",
]

ARGS = {
    "sin": "(<hello world>)"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compare_answers Not yet run
unverified (harvested, no recorded run)
mynlp/ccg2lambda/en/report_results.py
code served (permissive licence) · get_code("a348a12b155cc698")
How it ran

no invocation recorded.

get_text_from_node Not yet run
unverified (harvested, no recorded run)
mynlp/ccg2lambda/en/extract_entailment_problems.py
code served (permissive licence) · get_code("ca93c525b7a659f6")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 14 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

How to identify, extract, and use phrasal knowledge is a crucial problem for the task of Recognizing Textual Entailment (RTE). To solve this problem, we propose a method for detecting paraphrases via natural deduction proofs of semantic relations between sentence pairs. Our solution relies on a graph reformulation of partial variable unifications and an algorithm that induces subgraph alignments between meaning representations. Experiments show that our method can automatically detect various paraphrases that are absent from existing paraphrase databases. In addition, the detection of paraphrases using proof information improves the accuracy of RTE tasks.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1804.07656")
get_code_for_paper("1804.07656")
have("1804.07656")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/1804.07656.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/1804.07656.svg)](https://syntology.ai/paper/1804.07656)

Connect an agent — have() is free.