SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.03131 · ICLR · 2024

Constrained Decoding for Cross-lingual Label Projection

Minh Duong, Yang Le, Alan Chen, Wei Ritter, Xu

arXiv · PDF · Open in the Atlas

Code that ran

3 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 1 failure of our instrument; 1 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/2402.03131.json · badge: /api/ran/2402.03131.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
duonglm38/codec canonical (classifier label)
classifier basis: classifier_2025
3 ran · 1 no recorded run
FunctionStatusWhere it lives
Candidate Ran
ran on a synthesized input · constructed an object; no result was computed
duonglm38/codec/constrained_decoding.py
code served (permissive licence) · get_code("2bb799fc6782f023")
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 an instance of its own class, Candidate: the run constructed the object and did not call it, so no result was computed. 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)

# Create small tensors for text_ids and score
text_ids = torch.randint(0, 100, (2, 4))  # Shape (2, 4) - batch of 2, sequence length 4
score = torch.randn(2, 3)  # Shape (2, 3) - batch of 2, 3 scores

# Create accumulate_scores tensor
accumulate_scores = torch.randn(2, 5)  # Shape (2, 5) - batch of 2, 5 accumulated scores

ARGS = {
    "text_ids": text_ids,
    "score": score,
    "accumulate_scores": accumulate_scores,
}
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 Candidate — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Candidate 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
Node Ran
ran on a synthesized input · constructed an object; no result was computed
duonglm38/codec/constrained_decoding.py
code served (permissive licence) · get_code("e8ce6c1a54c20e57")
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 an instance of its own class, Node: the run constructed the object and did not call it, so no result was computed. 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
import numpy as np

torch.manual_seed(0)

ARGS = {
    'text': 'sample text',
    'log_prob': -2.5,
    'acc_log_prob': -5.0,
    'upperbound': -1.5
}
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 Node — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a Node 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
reorder_cache Ran
ran; our contract draft was wrong, not the code
duonglm38/codec/constrained_decoding.py
code served (permissive licence) · get_code("9ac5d343daaf013f")
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 6, 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 torch

torch.manual_seed(0)

# past_key_values is a tuple of tuples, where each inner tuple contains past states (tensors)
# that support index_select operation
# beam_idx is a tensor of indices used to reorder along dimension 0

# Create sample past_key_values structure
# Typically this is (num_layers,) where each layer has (key, value) tensors
# Shape: (batch_size, ...) where we'll reorder along dimension 0

batch_size = 2
num_heads = 3
head_dim = 4
seq_len = 5
num_layers = 2

# Create past_key_values with realistic shapes
past_key_values = ()
for layer_idx in range(num_layers):
    # Each layer has a tuple of (past_key, past_value)
    past_key = torch.randn(batch_size, num_heads, seq_len, head_dim)
    past_value = torch.randn(batch_size, num_heads, seq_len, head_dim)
    past_key_values += (tuple([past_key, past_value]),)

# beam_idx is used to reorder batch dimension (dimension 0)
# It selects which batch elements to keep in which order
beam_idx = torch.tensor([1, 0])  # Reorder: swap batch elements

ARGS = {
    "past_key_values": past_key_values,
    "beam_idx": beam_idx,
}
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 tuple, ndim 6, 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
bracket_constraint_decode Not yet run
unverified (harvested, no recorded run)
duonglm38/codec/constrained_decoding.py
code served (permissive licence) · get_code("f40014d4b63a95d7")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 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 4 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

Zero-shot cross-lingual transfer utilizing multilingual LLMs has become a popular learning paradigm for low-resource languages with no labeled training data. However, for NLP tasks that involve fine-grained predictions on words and phrases, the performance of zero-shot cross-lingual transfer learning lags far behind supervised fine-tuning methods. Therefore, it is common to exploit translation and label projection to further improve the performance by (1) translating training data that is available in a high-resource language (e.g., English) together with the gold labels into low-resource languages, and/or (2) translating test data in low-resource languages to a high-source language to run inference on, then projecting the predicted span-level labels back onto the original test data. However, state-of-the-art marker-based label projection methods suffer from translation quality degradation due to the extra label markers injected in the input to the translation model. In this work, we explore a new direction that leverages constrained decoding for label projection to overcome the aforementioned issues. Our new method not only can preserve the quality of translated texts but also has the versatility of being applicable to both translating training and translating test data strategies. This versatility is crucial as our experiments reveal that translating test data can lead to a considerable boost in performance compared to translating only training data. We evaluate on two cross-lingual transfer tasks, namely Named Entity Recognition and Event Argument Extraction, spanning 20 languages. The results demonstrate that our approach outperforms the state-of-the-art marker-based method by a large margin and also shows better performance than other label projection methods that rely on external word alignment 1 .

For agents

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

get_harvested_code_for_paper("2402.03131")
get_code_for_paper("2402.03131")
have("2402.03131")

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

curl https://syntology.ai/api/ran/2402.03131.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.03131.svg)](https://syntology.ai/paper/2402.03131)

Connect an agent — have() is free.