Pith. sign in

REVIEW 11 cited by

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1912.09713 v2 pith:546MWNKW submitted 2019-12-20 cs.LG cs.CLstat.ML

Measuring Compositional Generalization: A Comprehensive Method on Realistic Data

classification cs.LG cs.CLstat.ML
keywords methodbenchmarkscompositionalgeneralizationdivergencerealisticabilitycompound
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

State-of-the-art machine learning methods exhibit limited compositional generalization. At the same time, there is a lack of realistic benchmarks that comprehensively measure this ability, which makes it challenging to find and evaluate improvements. We introduce a novel method to systematically construct such benchmarks by maximizing compound divergence while guaranteeing a small atom divergence between train and test sets, and we quantitatively compare this method to other approaches for creating compositional generalization benchmarks. We present a large and realistic natural language question answering dataset that is constructed according to this method, and we use it to analyze the compositional generalization ability of three machine learning architectures. We find that they fail to generalize compositionally and that there is a surprisingly strong negative correlation between compound divergence and accuracy. We also demonstrate how our method can be used to create new compositionality benchmarks on top of the existing SCAN dataset, which confirms these findings.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 11 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Learning to Theorize the World from Observation

    cs.LG 2026-05 unverdicted novelty 7.0

    NEO is a probabilistic neural model that induces compositional programs as a learned Language of Thought from non-textual observations and executes them via a shared transition model to enable explanation-driven gener...

  2. Training Transformers as a Universal Computer

    cs.AI 2026-04 unverdicted novelty 7.0

    A transformer trained on random meaningless MicroPy programs generalizes to execute diverse human-written programs, providing empirical evidence it can act as a universal computer.

  3. Understanding Generalization in Role-Playing Models via Information Theory

    cs.LG 2025-12 unverdicted novelty 7.0

    R-EMID metric with upper bound shows user shifts pose highest risk to role-playing model generalization, with co-evolving RL as most effective mitigation.

  4. From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning

    cs.LG 2026-06 unverdicted novelty 6.0

    Introduces a hierarchical latent selection model showing SFT supplies raw module materials in compound traces while RL decomposes them to identify atomic modules and enable recombination for new reasoning configurations.

  5. Why Prompt Optimization Works, and Why It Sometimes Doesn't: A Causal-Inspired Edit-Level Analysis

    cs.CL 2026-05 unverdicted novelty 6.0

    Observational causal-inspired analysis finds prompt optimization failures arise from systematic interactions between edit families and task characteristics rather than random artifacts.

  6. Learning to Theorize the World from Observation

    cs.LG 2026-05 unverdicted novelty 6.0

    NEO induces compositional latent programs as world theories from observations and executes them to enable explanation-driven generalization.

  7. Generalization in LLM Problem Solving: The Case of the Shortest Path

    cs.AI 2026-04 unverdicted novelty 6.0

    LLMs show strong spatial generalization to unseen maps in shortest-path tasks but fail length scaling due to recursive instability, with data coverage setting hard limits.

  8. How Do Language Models Compose Functions?

    cs.CL 2025-10 conditional novelty 6.0

    LLMs solve compositional factual recall either by computing intermediates or directly, with mechanism choice correlated to translation geometry in embedding spaces.

  9. Blind-Spots-Bench: Evaluating Blind Spots in Multimodal Models

    cs.AI 2026-07 conditional novelty 5.0

    A 235-item multimodal stress-test shows frontier closed models outpace open-weight peers by ~10% and leaves shared failures on counting, spatial, and character-level tasks.

  10. ClassicLogic: A Knowledge-Driven Benchmark of Classic Puzzle Games for Evaluating Compositional Generalization

    cs.AI 2026-07 conditional novelty 5.0

    ClassicLogic is an open-source benchmark using four logic puzzles with a hierarchical strategy knowledge base to evaluate three forms of compositional generalization in AI agents.

  11. Reasoning or Memorization? Direction-Aware Diversity Exploration in LLM Reinforcement Learning

    cs.AI 2026-06 unverdicted novelty 5.0

    DiRL extracts a reasoning-memorization direction from model representations inside GRPO to weight gradients and shape rewards so that exploration favors reasoning trajectories over memorization ones.