SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.03549 · 2026

FEA-SLT: A Gloss-Free End-to-End Framework for Facial-Expression-Aware Sign Language Translation

Di Weng, Guobin Tu

arXiv · PDF · Open in the Atlas

Code that ran

12 of 12 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; 1 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 10 failures of our instrument; 0 no recorded run), measured 2026-11-22; 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/2601.03549.json · badge: /api/ran/2601.03549.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
google-research/bleurt canonical (classifier label)
classifier basis: census
12 ran · 0 no recorded run
FunctionStatusWhere it lives
convert_to_unicode Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/lib/bert_tokenization.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from huawei-noah/Pretrained-Language-Model, is pointer only (licence: NONE) · get_code("1923fc05163d207d")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from huawei-noah/Pretrained-Language-Model, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

create_optimizer Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/lib/optimization.py
code served (permissive licence) · get_code("597f6dac08ea3245")
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 Operation. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create a simple computational graph with a loss
# We need: loss (a scalar tensor), init_lr, num_train_steps, num_warmup_steps, use_tpu

# Reset graph to ensure clean state
tf.reset_default_graph()

# Create simple variables and a loss
with tf.variable_scope("model"):
    # Create a small variable
    weights = tf.get_variable(
        name="weights",
        shape=[4, 3],
        dtype=tf.float32,
        initializer=tf.random_normal_initializer()
    )
    bias = tf.get_variable(
        name="bias",
        shape=[3],
        dtype=tf.float32,
        initializer=tf.zeros_initializer()
    )
    
    # Create a simple input and loss
    input_data = tf.constant(np.random.randn(2, 4).astype(np.float32))
    output = tf.matmul(input_data, weights) + bias
    
    # Create a simple loss (mean squared error)
    loss = tf.reduce_mean(tf.square(output))

# Define arguments for create_optimizer
ARGS = {
    "loss": loss,
    "init_lr": 0.001,
    "num_train_steps": 1000,
    "num_warmup_steps": 100,
    "use_tpu": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Operation — a Operation output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Operation 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
encode_example Ran
ran, honoured the contract we drafted
google-research/bleurt/bleurt/encoding.py
code served (permissive licence) · get_code("c59cb856dad5f089")
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 2, dtype family int; 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 torch
import numpy as np

class SimpleTokenizer:
    """Minimal tokenizer for testing encode_example."""
    
    def __init__(self):
        self.vocab = {
            '[CLS]': 101,
            '[SEP]': 102,
            '[PAD]': 0,
            'the': 1,
            'quick': 2,
            'brown': 3,
            'fox': 4,
            'jumps': 5,
            'over': 6,
            'lazy': 7,
            'dog': 8,
            'hello': 9,
            'world': 10,
        }
    
    def tokenize(self, text):
        """Simple whitespace tokenization."""
        return text.lower().split()
    
    def convert_tokens_to_ids(self, tokens):
        """Convert tokens to IDs."""
        ids = []
        for token in tokens:
            ids.append(self.vocab.get(token, 0))
        return ids

torch.manual_seed(0)

reference = "the quick brown fox"
candidate = "hello world"
tokenizer = SimpleTokenizer()
max_seq_length = 16

ARGS = {
    "reference": reference,
    "candidate": candidate,
    "tokenizer": tokenizer,
    "max_seq_length": max_seq_length,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family int — 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
gelu Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/lib/modeling.py
code served (permissive licence) · get_code("51d3bc0cfa6415a3")
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 SymbolicTensor. 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 numpy as np
import tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

# Create a small float32 tensor for x
x = tf.constant([[0.5, -0.3, 1.2, -0.8],
                  [0.1, 0.2, -0.5, 0.9]], dtype=tf.float32)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
get_activation Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/lib/modeling.py
code served (permissive licence) · get_code("217fe96a5945a42d")
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 function. 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 numpy as np
import six
import tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

ARGS = {
    'activation_string': 'relu'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, six, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
get_assignment_map_from_checkpoint Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/lib/modeling.py
code served (permissive licence) · get_code("7bc46f4950599a87")
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 collections
import re
import tensorflow.compat.v1 as tf
import tempfile
import os

# Disable eager execution for TensorFlow v1 compatibility
tf.disable_eager_execution()

# Create a temporary checkpoint file for testing
temp_dir = tempfile.mkdtemp()
checkpoint_path = os.path.join(temp_dir, "test_checkpoint")

# Create a simple TensorFlow checkpoint with some variables
with tf.Graph().as_default():
    with tf.Session() as sess:
        # Create some variables to save
        var1 = tf.Variable(tf.ones([2, 3]), name="layer1/weights")
        var2 = tf.Variable(tf.zeros([4]), name="layer1/bias")
        var3 = tf.Variable(tf.ones([3, 4]), name="layer2/weights")
        
        sess.run(tf.global_variables_initializer())
        
        # Save the checkpoint
        saver = tf.train.Saver()
        saver.save(sess, checkpoint_path)

# Create variables for the current graph
with tf.Graph().as_default():
    with tf.Session() as sess:
        # Create variables with matching names
        tvars = [
            tf.Variable(tf.ones([2, 3]), name="layer1/weights"),
            tf.Variable(tf.zeros([4]), name="layer1/bias"),
            tf.Variable(tf.ones([3, 4]), name="layer2/weights"),
            tf.Variable(tf.ones([5]), name="layer3/bias"),  # This one won't be in checkpoint
        ]
        
        sess.run(tf.global_variables_initializer())
        
        ARGS = {
            "tvars": tvars,
            "init_checkpoint": checkpoint_path,
        }
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — 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
preserve_token Ran
ran, violated the contract we drafted
google-research/bleurt/bleurt/lib/bert_tokenization.py
code served (permissive licence) · get_code("600830144a7c6590")
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 bool, ndim 0, dtype family float, 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.

import re

PRESERVE_UNUSED_TOKENS = False

_UNUSED_TOKEN_RE = re.compile('^\\[unused\\d+\\]$')

def preserve_token(token, vocab):
    """Returns True if the token should forgo tokenization and be preserved."""
    if not PRESERVE_UNUSED_TOKENS:
        return False
    if token not in vocab:
        return False
    return bool(_UNUSED_TOKEN_RE.search(token))


# Build test arguments
token = "[unused42]"
vocab = {"[unused0]", "[unused1]", "[unused42]", "hello", "world"}

ARGS = {
    "token": token,
    "vocab": vocab,
}
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 bool, ndim 0, dtype family float, 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
printable_text Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/lib/bert_tokenization.py
code served (permissive licence) · get_code("0e5615f8994003cf")
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 six

# The function takes a single parameter 'text' which can be:
# - str (works in both Python 2 and 3)
# - bytes (in Python 3, decoded to utf-8)
# - unicode (in Python 2, encoded to utf-8)

# Since we're running on Python 3, we'll test with a string
text_input = "Hello, World!"

ARGS = {
    "text": text_input
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
six — 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
ranking_loss Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/model.py
code served (permissive licence) · get_code("928f493bec0ffc69")
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 SymbolicTensor. 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 tensorflow.compat.v1 as tf
import numpy as np

tf.disable_v2_behavior()

# Create small sample tensors
batch_size = 2
predictions = tf.constant([[0.3], [0.7]], dtype=tf.float32)
labels = tf.constant([[0.2], [0.9]], dtype=tf.float32)

ARGS = {
    "predictions": predictions,
    "labels": labels,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
read_bleurt_config Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/checkpoint.py
code served (permissive licence) · get_code("f57743217d50baa3")
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 json
import os
import tempfile
import shutil

# Create a temporary directory structure for BLEURT checkpoint
temp_dir = tempfile.mkdtemp()

# Create the checkpoint directory
checkpoint_dir = os.path.join(temp_dir, 'bleurt_checkpoint')
os.makedirs(checkpoint_dir)

# Create variables directory
variables_dir = os.path.join(checkpoint_dir, 'variables')
os.makedirs(variables_dir)

# Create a dummy variables file
variables_file = os.path.join(variables_dir, 'variables')
open(variables_file, 'w').close()

# Create config file
config_data = {
    'name': 'test_bleurt_model',
    'vocab_file': 'vocab.txt',
    'sp_model': 'sentencepiece_model',
    'some_other_file': 'other.txt',
    'some_dir': 'subdir'
}

config_path = os.path.join(checkpoint_dir, 'bleurt_config.json')
with open(config_path, 'w') as f:
    json.dump(config_data, f)

# Create the referenced files and directories
vocab_file = os.path.join(checkpoint_dir, 'vocab.txt')
open(vocab_file, 'w').close()

other_file = os.path.join(checkpoint_dir, 'other.txt')
open(other_file, 'w').close()

subdir = os.path.join(checkpoint_dir, 'subdir')
os.makedirs(subdir)

# Create sentencepiece model files
sp_model_file = os.path.join(checkpoint_dir, 'sentencepiece_model.model')
open(sp_model_file, 'w').close()

sp_vocab_file = os.path.join(checkpoint_dir, 'sentencepiece_model.vocab')
open(sp_vocab_file, 'w').close()

ARGS = {
    'path': checkpoint_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — 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
serialize_example Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/encoding.py
code served (permissive licence) · get_code("f71891eba726ed24")
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 bytes. 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 collections
import tensorflow.compat.v1 as tf

# Create a minimal stand-in tokenizer class
class StandInTokenizer:
    def tokenize(self, text):
        # Simple whitespace tokenization
        return text.lower().split()
    
    def convert_tokens_to_ids(self, tokens):
        # Map tokens to simple integer ids
        # Special tokens get low ids, others get higher ids
        token_to_id = {
            '[CLS]': 101,
            '[SEP]': 102,
        }
        ids = []
        for token in tokens:
            if token in token_to_id:
                ids.append(token_to_id[token])
            else:
                # Map word tokens to sequential ids
                ids.append(1000 + hash(token) % 1000)
        return ids

# Instantiate the tokenizer
tokenizer = StandInTokenizer()

# Define arguments for serialize_example
ARGS = {
    'reference': 'The quick brown fox jumps over the lazy dog',
    'candidate': 'A fast brown fox leaps over a lazy dog',
    'tokenizer': tokenizer,
    'max_seq_length': 128,
    'score': 0.85
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bytes — a bytes output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a bytes 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
serialize_raw_example Ran
ran; our contract draft was wrong, not the code
google-research/bleurt/bleurt/encoding.py
code served (permissive licence) · get_code("74a38d5d8b44259d")
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 bytes. 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 collections
import tensorflow.compat.v1 as tf

# Disable TF v2 behavior
tf.disable_v2_behavior()

ARGS = {
    "reference": "This is a reference sentence for testing.",
    "candidate": "This is a candidate sentence for testing."
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bytes — a bytes output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a bytes 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

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 10 are also present in other papers whose code we hold (10 with at least one paper dated earlier, 0 the same month, 0 later) and 2 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.

FunctionAlso present inEarliest-datedOthers named
convert_to_unicode 36 other papers
36 dated earlier
1609.08144 (2016-09)
Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.04085 (2019-01)
Passage Re-ranking with BERT
and 33 more not named here
printable_text 33 other papers
33 dated earlier
1609.08144 (2016-09)
Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.04085 (2019-01)
Passage Re-ranking with BERT
and 30 more not named here
get_activation 32 other papers
32 dated earlier · 1 only as an assembly brick
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.04085 (2019-01)
Passage Re-ranking with BERT
and 29 more not named here
gelu 26 other papers
26 dated earlier · 4 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
and 23 more not named here
get_assignment_map_from_checkpoint 26 other papers
26 dated earlier
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
1808.09781 (2018-08)
Self-Attentive Sequential Recommendation
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
and 23 more not named here
ranking_loss 2 other papers
2 dated earlier
2310.20703 (2023-10)
Vanishing Gradients in Reinforcement Finetuning of Language Models
2024.findings-emnlp.27 (2024 (year only))
TRANSLLAMA: LLM-based Simultaneous Translation System
encode_example 1 other paper
1 dated earlier
2310.20703 (2023-10)
Vanishing Gradients in Reinforcement Finetuning of Language Models
read_bleurt_config 1 other paper
1 dated earlier
2310.20703 (2023-10)
Vanishing Gradients in Reinforcement Finetuning of Language Models
serialize_example 1 other paper
1 dated earlier
2310.20703 (2023-10)
Vanishing Gradients in Reinforcement Finetuning of Language Models
serialize_raw_example 1 other paper
1 dated earlier
2310.20703 (2023-10)
Vanishing Gradients in Reinforcement Finetuning of Language Models

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

Sign Language Translation (SLT) is a challenging cross-modal task requiring joint modeling of manual articulations and non-manual signals. Existing gloss-free SLT methods effectively capture gestural dynamics but often underutilize facial expressions, which play crucial grammatical and disambiguating roles. This limitation can cause semantic degradation when distinct concepts share similar manual configurations. To address this issue, we propose FEA-SLT (Facial-Expression-Aware Sign Language Translation), a gloss-free end-toend framework that uses facial dynamics as semantic anchors for resolving manual ambiguity. FEA-SLT employs a domain-transferred facial encoder to extract expression-sensitive representations and integrates them with manual features through a linguistically constrained Facial-Expression-Aware Fusion (FEAF) module. FEAF captures reciprocal dependencies between manual and facial channels via bidirectional modulation, enhancing syntactic fidelity. Experiments on PHOENIX14T and CSL-Daily show that FEA-SLT achieves stateof-the-art BLEU performance among glossfree methods, while targeted analyses confirm improved translation of facial-sensitive utterances. Code is available at https://github. com/TuGuobin/FEA-SLT.

For agents

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

get_harvested_code_for_paper("2601.03549")
get_code_for_paper("2601.03549")
have("2601.03549")

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

curl https://syntology.ai/api/ran/2601.03549.json

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

[![Syntology run record](https://syntology.ai/api/ran/2601.03549.svg)](https://syntology.ai/paper/2601.03549)

Connect an agent — have() is free.