SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.05045 · ICML · 2024

A Tensor Decomposition Perspective on Second-order RNNs

Guillaume Rabusseau, Maude Lizaire, Michael Rizvi-Martel, Marawan Gamal, Abdel Hameed

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 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; 0 failures of our instrument; 0 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/2406.05045.json · badge: /api/ran/2406.05045.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
MaudeLiz/cprnn marked official in Papers with Code
Papers with Code archive
2 ran · 0 no recorded run
FunctionStatusWhere it lives
CPRNN Ran
ran on a synthesized input · constructed an object; no result was computed
MaudeLiz/cprnn/models/cp_rnn.py
code served (permissive licence) · get_code("2230ceef1832685d")
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, CPRNN: 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 torch.nn as nn
import numpy as np
import math

torch.manual_seed(0)

class CharacterTokenizer:
    """Facilitates Conversion between vocab <=> integers (tokenization)

    Args:
        tokens: (iterable) previously extracted tokens (i.e. list of characters or words)
    """

    def __init__(self, tokens: list=None):
        self._tokens = list(set(tokens)) if tokens is not None else list()
        self._tokens = ['<UNK>', *self._tokens] if '<UNK>' not in self._tokens else self._tokens
        self.char_to_ix_dct = {ch: i for i, ch in enumerate(self._tokens)}
        self.ix_to_char_dct = {i: ch for i, ch in enumerate(self._tokens)}

    @property
    def vocab_size(self):
        return len(self.char_to_ix_dct)

    def add_token(self, token):
        if token not in self.char_to_ix_dct:
            self._tokens.append(token)
            self.char_to_ix_dct[token] = len(self.char_to_ix_dct)
            if len(self.ix_to_char_dct) in self.ix_to_char_dct:
                raise ValueError('Tokenizer is corrupted. Please check the integrity of the tokenizer.')
            self.ix_to_char_dct[len(self.ix_to_char_dct)] = token
        return self.char_to_ix_dct[token]

    @property
    def tokens(self):
        return self._tokens

    def char_to_ix(self, char: str=None):
        return self.char_to_ix_dct[char]

    def ix_to_char(self, ix):
        if isinstance(ix, np.ndarray):
            def f(x):
                return self.ix_to_char_dct[x]
            vf = np.vectorize(f)
            return vf(ix)
        elif any([isinstance(ix, d) for d in [int, np.int32, np.int64]]):
            return self.ix_to_char_dct[ix]
        else:
            raise ValueError('Tokenizer expected either int or array as input but got {}'.format(type(ix)))

    def tokenize(self, sentence: str=None):
        return np.array([self.char_to_ix_dct.get(ch, 0) for ch in sentence])


# Create a minimal tokenizer
tokens = ['a', 'b', 'c', 'd', 'e']
tokenizer = CharacterTokenizer(tokens=tokens)

# Build ARGS dictionary
ARGS = {
    'input_size': 4,
    'hidden_size': 3,
    'vocab_size': len(tokenizer.tokens),
    'use_embedding': False,
    'rank': 2,
    'tokenizer': tokenizer,
    'batch_first': True,
    'dropout': 0.5,
    'gate': 'tanh'
}
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 CPRNN — 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 CPRNN 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
CharacterTokenizer Ran
ran on a synthesized input · constructed an object; no result was computed
MaudeLiz/cprnn/models/cp_rnn.py
code served (permissive licence) · get_code("36275cee8d2c4b3d")
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, CharacterTokenizer: 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 numpy as np

ARGS = {
    "tokens": ['a', 'b', 'c', 'd', 'e']
}
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 CharacterTokenizer — 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 CharacterTokenizer 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 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 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 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.

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

Second-order Recurrent Neural Networks (2RNNs) extend RNNs by leveraging secondorder interactions for sequence modelling. These models are provably more expressive than their first-order counterparts and have connections to well-studied models from formal language theory. However, their large parameter tensor makes computations intractable. To circumvent this issue, one approach known as MIRNN consists in limiting the type of interactions used by the model. Another is to leverage tensor decomposition to diminish the parameter count. In this work, we study the model resulting from parameterizing 2RNNs using the CP decomposition, which we call CPRNN. Intuitively, the rank of the decomposition should reduce expressivity. We analyze how rank and hidden size affect model capacity and show the relationships between RNNs, 2RNNs, MIRNNs, and CPRNNs based on these parameters. We support these results empirically with experiments on the Penn Treebank dataset which demonstrate that, with a fixed parameter budget, CPRNNs outperforms RNNs, 2RNNs, and MIRNNs with the right choice of rank and hidden size.

For agents

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

get_harvested_code_for_paper("2406.05045")
get_code_for_paper("2406.05045")
have("2406.05045")

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

curl https://syntology.ai/api/ran/2406.05045.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.05045.svg)](https://syntology.ai/paper/2406.05045)

Connect an agent — have() is free.