Challenge: In-context learning is an inductive bias for compositional generalization, but many deep neural architectures struggle with this ability.
Approach: They propose to force a causal Transformer to in-context learn to promote compositional generalization by using earlier examples to generalize to later ones.
Outcome: The proposed model can solve 'ordinary' learning problems by utilizing earlier examples to generalize to later ones, i.e., in-context learning.

Similar Papers

How Do In-Context Examples Affect Compositional Generalization? (2023.acl-long)

Copied to clipboard

Challenge: In-context learning paradigms that focus on large corpus are limiting compositional generalization performance.
Approach: They propose a test suite to investigate in-context compositional generalization . they propose to use examples that are structurally similar to the test case .
Outcome: The proposed test suite investigates in-context compositional generalization performance . it finds that the performance can be affected by the selection of in-const examples .
Data Factors for Better Compositional Generalization (2023.emnlp-main)

Copied to clipboard

Challenge: Recent diagnostic datasets on compositional generalization expose severe problems . state-of-the-art models trained on larger and more general datasets show better generalization ability .
Approach: They conduct an empirical analysis by training Transformer models on a variety of training sets with different data factors including dataset scale, pattern complexity, example difficulty, etc.
Outcome: The proposed model training on larger datasets improves on compositional generalization tasks.
Analyzing the Inner Workings of Transformers in Compositional Generalization (2025.naacl-long)

Copied to clipboard

Challenge: Existing studies on compositional generalization abilities of neural models have focused on benchmarks, but the results do not reflect the underlying competence of the model.
Approach: They propose to find an existing subnetwork that contributes to the generalization performance and perform causal analyses on how the model utilizes syntactic features.
Outcome: The proposed model relies on syntactic features but the subnetwork with better generalization performance relies mainly on a non-compositional algorithm .
Making Transformers Solve Compositional Tasks (2022.acl-long)

Copied to clipboard

Challenge: Several studies have reported the inability of Transformer models to generalize compositionally . a key aspect of natural language is the ability to learn basic primitives .
Approach: They propose to use Transformers to generalize compositionally in a large range of tasks . they find that Transformers generalize significantly better than previous models .
Outcome: The proposed models generalize compositionally significantly better than previous models . a set of 12 datasets shows that the proposed models can be improved .
Can Input Attributions Explain Inductive Reasoning in In-Context Learning? (2025.findings-acl)

Copied to clipboard

Challenge: interpreting the internal process of neural models has long been a challenge . despite rapid progress, there are still questions bridging the IA and MI eras .
Approach: They propose to use input attribution methods to interpret in-context learning . they find that a certain simple IA method works best in large models .
Outcome: The proposed method is the best for interpreting LLM-based ICL, but the larger the model, the harder it is to interpret it.
Learning Syntax Without Planting Trees: Understanding Hierarchical Generalization in Transformers (2025.tacl-1)

Copied to clipboard

Challenge: Inductive biases in transformers can cause hierarchical generalization without explicitly encoding structural bias.
Approach: They investigate sources of inductive bias in transformer models and their training that could cause such preference for hierarchical generalization.
Outcome: The proposed model can generalize to novel syntactic forms without explicit bias . the proposed model is able to generalize on a dataset with a hierarchical grammar .
Consistency Regularization Training for Compositional Generalization (2023.acl-long)

Copied to clipboard

Challenge: Existing neural models have difficulty generalizing to unseen combinations of seen components.
Approach: They propose to improve the capability of Transformer on compositional generalization by consistency regularization training without modifying model architectures.
Outcome: The proposed model performs well on semantic parsing and machine translation benchmarks.
Compositional Generalization by Factorizing Alignment and Translation (2020.acl-srw)

Copied to clipboard

Challenge: a crucial property underlying the expressive power of human language is its systematicity.
Approach: They propose to make an analogous separation between alignment and translation in neural machine translation to capture compositional structure.
Outcome: The proposed architecture outperforms existing neural networks on a compositional generalization task without supervision.
When Can Transformers Ground and Compose: Insights from Compositional Generalization Benchmarks (2022.emnlp-main)

Copied to clipboard

Challenge: Recent benchmarks like ReaSCAN use navigation tasks grounded in a grid world to assess whether neural models exhibit compositional behaviour.
Approach: They propose a transformer-based model that outperforms specialized architectures on ReaSCAN and a modified version of gSCAN to test their performance.
Outcome: The proposed model outperforms specialized architectures on ReaSCAN and gSCAN on a grid world and can generalize to deeper input structures.
Semantic Training Signals Promote Hierarchical Syntactic Generalization in Transformers (2024.emnlp-main)

Copied to clipboard

Challenge: Neural networks without hierarchical biases struggle to learn linguistic rules that come naturally to humans . et al., 2018: Transformers trained on form and meaning favor hierarchically generalization more than those trained on forms alone.
Approach: They examine whether neural networks without hierarchical biases can generalize more like humans . they find that Transformers trained on form and meaning favor hierarchic generalization .
Outcome: The proposed neural networks perform better on syntactic evaluations when trained on form and meaning compared to those trained on forms alone.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations