SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1909.11874 · 2019

Compact Trilinear Interaction for Visual Question Answering

arXiv · PDF · Open in the Atlas

Code that ran

3 of 5 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, 1 violated it and 2 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/1909.11874.json · badge: /api/ran/1909.11874.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
aioz-ai/ICCV19_VQA-CTI canonical (classifier label)
classifier basis: census
3 ran · 2 no recorded run
FunctionStatusWhere it lives
ModeProduct Ran
ran on a synthesized input
aioz-ai/ICCV19_VQA-CTI/src/Tensor.py
code served (permissive licence) · get_code("dae936b303032b0f")
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 Tensor, ndim 5, dtype family float, shape (2, 7, 8, 9, 10) 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 torch

torch.manual_seed(0)

# Analyze the function signature and body:
# - tensor: 5D tensor, used with .transpose(), .size(), .view(), .contiguous()
# - matrix_1, matrix_2, matrix_3, matrix_4: 2D matrices used with torch.matmul()
# - n_way: int parameter that controls conditional branches

# From the code, tensor has 5 dimensions (indices 0-4)
# tensor.size(0) = batch dimension
# tensor.size(1), tensor.size(2), tensor.size(3), tensor.size(4) are spatial dims

# First operation: tensor.transpose(3, 2).contiguous().view(tensor.size(0), tensor.size(1), tensor.size(2) * tensor.size(3) * tensor.size(4))
# This means we reshape to (batch, dim1, dim2*dim3*dim4)
# So matrix_1 should be compatible: torch.matmul(matrix_1, (batch, dim1, dim2*dim3*dim4))
# matrix_1 shape: (out_features, dim1)

# Let's use small distinct sizes:
batch_size = 2
dim1 = 3
dim2 = 4
dim3 = 5
dim4 = 6

# Create the 5D input tensor
tensor = torch.randn(batch_size, dim1, dim2, dim3, dim4)

# matrix_1: for torch.matmul(matrix_1, tensor_1) where tensor_1 is (batch, dim1, dim2*dim3*dim4)
# Result should be (batch, out1, dim2*dim3*dim4)
out1 = 7
matrix_1 = torch.randn(out1, dim1)

# After first matmul and reshape, tensor_product is (batch, out1, dim4, dim3, dim2)
# Then tensor_2 reshape for second matmul needs matrix_2 compatible with (batch, dim2, out1*dim3*dim4)
out2 = 8
matrix_2 = torch.randn(out2, dim2)

# For n_way > 2, matrix_3 is used
# tensor_3 reshape for third matmul: (batch, dim3, dim4*dim2*out1)
out3 = 9
matrix_3 = torch.randn(out3, dim3)

# For n_way > 3, matrix_4 is used
# tensor_4 reshape: (batch, dim4, out3*dim2*dim1)
out4 = 10
matrix_4 = torch.randn(out4, dim4)

ARGS = {
    'tensor': tensor,
    'matrix_1': matrix_1,
    'matrix_2': matrix_2,
    'matrix_3': matrix_3,
    'matrix_4': matrix_4,
    'n_way': 4
}
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 Tensor, ndim 5, dtype family float, shape (2, 7, 8, 9, 10) — shape (2, 7, 8, 9, 10) 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
answer_filter Ran
ran, violated the contract we drafted
aioz-ai/ICCV19_VQA-CTI/src/dataset.py
pointer only (licence: MIT); the stored copy get_code serves, from KaveeshaSIlva/PathVQA, is pointer only (licence: NONE) · get_code("cf2efa825f4afaf0")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from KaveeshaSIlva/PathVQA, 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_glove_embedding_init Ran
ran on a synthesized input
aioz-ai/ICCV19_VQA-CTI/src/dataset.py
code served (permissive licence) · get_code("71aa385eadeb1226")
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. 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 tempfile
import os

# Create a temporary GloVe file for testing
temp_dir = tempfile.mkdtemp()
glove_file = os.path.join(temp_dir, "glove.txt")

# Write a small GloVe file with 3 words and 4-dimensional embeddings
glove_content = """apple 0.1 0.2 0.3 0.4
banana 0.5 0.6 0.7 0.8
cherry 0.9 1.0 1.1 1.2
"""

with open(glove_file, 'w', encoding='utf-8') as f:
    f.write(glove_content)

# Create idx2word mapping
idx2word = {
    0: 'apple',
    1: 'banana',
    2: 'cherry',
    3: 'unknown_word'
}

ARGS = {
    'idx2word': idx2word,
    'glove_file': glove_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple 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
load_json Not yet run
unverified (harvested, no recorded run)
aioz-ai/ICCV19_VQA-CTI/src/evaluate_TDIUC.py
code served (permissive licence) · get_code("6652b5d834042107")
How it ran

no invocation recorded.

tfidf_from_questions Not yet run
unverified (harvested, no recorded run)
aioz-ai/ICCV19_VQA-CTI/src/dataset.py
code served (permissive licence) · get_code("fccda75b0263c740")
How it ran

no invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 2 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 2 later) and 3 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
answer_filter 5 other papers
2 dated earlier, 3 later
1805.07932 (2018-05)
Bilinear Attention Networks
1903.12314 (2019-03)
Relation-Aware Graph Attention Network for Visual Question Answering
2003.10286 (2020-03)
PathVQA: 30000+ Questions for Medical Visual Question Answering
and 2 more not named here
create_glove_embedding_init 1 other paper
1 later
2110.02526 (2021-10)
Coarse-to-Fine Reasoning for Visual Question Answering

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

In Visual Question Answering (VQA), answers have a great correlation with question meaning and visual contents. Thus, to selectively utilize image, question and answer information, we propose a novel trilinear interaction model which simultaneously learns high level associations between these three inputs. In addition, to overcome the interaction complexity, we introduce a multimodal tensor-based PARALIND decomposition which efficiently parameterizes trilinear interaction between the three inputs. Moreover, knowledge distillation is first time applied in Free-form Opened-ended VQA. It is not only for reducing the computational cost and required memory but also for transferring knowledge from trilinear interaction model to bilinear interaction model. The extensive experiments on benchmarking datasets TDIUC, VQA-2.0, and Visual7W show that the proposed compact trilinear interaction model achieves state-of-the-art results when using a single model on all three datasets.

For agents

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

get_harvested_code_for_paper("1909.11874")
get_code_for_paper("1909.11874")
have("1909.11874")

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

curl https://syntology.ai/api/ran/1909.11874.json

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

[![Syntology run record](https://syntology.ai/api/ran/1909.11874.svg)](https://syntology.ai/paper/1909.11874)

Connect an agent — have() is free.