mlsanityData to 14 Oct 2026

CNNs found to jump around more skillfully than RNNs: Compositional generalization in seq2seq convolutional networks

arXiv 1905.08527 ↗First posted 21 May 2019cs.CL cs.AI cs.LGcited by 22

Not on the map. Why?

Abstract (arXiv)

Lake and Baroni (2018) introduced the SCAN dataset probing the ability of seq2seq models to capture compositional generalizations, such as inferring the meaning of "jump around" 0-shot from the component words. Recurrent networks (RNNs) were found to completely fail the most challenging generalization cases. We test here a convolutional network (CNN) on these tasks, reporting hugely improved performance with respect to RNNs. Despite the big improvement, the CNN has however not induced systematic rules, suggesting that the difference between compositional and non-compositional behaviour is not clear-cut.

Key facts. Cited by 22 papers in the map (data through 14 Oct 2026). First posted on arXiv 21 May 2019.

Builds on

No reference list yet.

Built on by

Cited by year

Citing papers per year: 2021: 6; 2022: 8; 2023: 2; 2024: 3; 2025: 3.

22 papers.

Sort:

  1. 2022State-of-the-art generalisation research in NLP: A taxonomy and reviewcited by 98
  2. 2021The Neural Data Router: Adaptive Control Flow in Transformers Improves Systematic Generalizationcited by 62
  3. 2021On Compositional Generalization of Neural Machine Translationcited by 34
  4. 2021Sequence-to-Sequence Learning with Latent Neural Grammarscited by 30
  5. 2022Emergent Communication: Generalization and Overfitting in Lewis Gamescited by 28
  6. 2022Equi-Tuning: Group Equivariant Fine-Tuning of Pretrained Modelscited by 22
  7. 2021Inducing Transformer's Compositional Generalization Ability via Auxiliary Sequence Prediction Taskscited by 19
  8. 2024A Survey on Compositional Learning of AI Models: Theoretical and Experimental Practicescited by 15
  9. 2021Can Transformers Jump Around Right in Natural Language? Assessing Performance Transfer from SCANcited by 14
  10. 2022Neural-Symbolic Recursive Machine for Systematic Generalizationcited by 12
  11. 2022Defending Compositionality in Emergent Languagescited by 8
  12. 2022Mutual Exclusivity Training and Primitive Augmentation to Induce Compositionalitycited by 8
10 more
  1. 2024From Frege to chatGPT: Compositionality in language, cognition, and deep neural networkscited by 7
  2. 2021A Minimalist Dataset for Systematic Generalization of Perception, Syntax, and Semanticscited by 6
  3. 2023Data Factors for Better Compositional Generalizationcited by 5
  4. 2022Deep learning for laboratory earthquake prediction and autoregressive forecasting of fault zone stresscited by 4
  5. 2022Understanding Robust Generalization in Learning Regular Languagescited by 3
  6. 2023Layer-wise Representation Fusion for Compositional Generalizationcited by 2
  7. 2025Systematic Generalization in Language Models Scales with Information Entropycited by 2
  8. 2024Inducing Systematicity in Transformers by Attending to Structurally Quantized Embeddingscited by 1
  9. 2025CompoST: A Benchmark for Analyzing the Ability of LLMs To Compositionally Interpret Questions in a QALD Settingcited by 1
  10. 2025Data Distributional Properties As Inductive Bias for Systematic Generalizationcited by 0