mlsanityData to 14 Oct 2026

System 2 Reasoning for Human-AI Alignment: Generality and Adaptivity via ARC-AGI

arXiv 2410.07866 ↗First posted 10 Oct 2024cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

Despite their broad applicability, transformer-based models still fall short in System~2 reasoning, lacking the generality and adaptivity needed for human--AI alignment. We examine weaknesses on ARC-AGI tasks, revealing gaps in compositional generalization and novel-rule adaptation, and argue that closing these gaps requires overhauling the reasoning pipeline and its evaluation. We propose three research axes: (1) Symbolic representation pipeline for compositional generality, (2) Interactive feedback-driven reasoning loop for adaptivity, and (3) Test-time task augmentation balancing both qualities. Finally, we demonstrate how ARC-AGI's evaluation suite can be adapted to track progress in symbolic generality, feedback-driven adaptivity, and task-level robustness, thereby guiding future work on robust human--AI alignment.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 18 of its 34 references that are in the map. First posted on arXiv 10 Oct 2024.

Builds on

18 of 34 references.

References · 18

  1. 2023Large Language Models as General Pattern Machinescited by 197
  2. 2023Hypothesis Search: Inductive Reasoning with Language Modelscited by 120
  3. 2023LLMs and the Abstraction and Reasoning Corpus: Successes, Failures, and the Importance of Object-based Representationscited by 82
  4. 2020Thinking Fast and Slow in AIcited by 70
  5. 2024Code Repair with LLMs gives an Exploration-Exploitation Tradeoffcited by 51
  6. 2022Graphs, Constraints, and Search for the Abstraction and Reasoning Corpuscited by 32
  7. 2024Reasoning Abilities of Large Language Models: In-Depth Analysis on the Abstraction and Reasoning Corpuscited by 32
  8. 2021Neural-guided, Bidirectional Program Search for Abstraction and Reasoningcited by 23
  9. 2024CodeIt: Self-Improving Language Models with Prioritized Hindsight Replaycited by 23
  10. 2024HYSYNTH: Context-Free LLM Approximation for Guiding Program Synthesiscited by 15
  11. 2023Counting and Algorithmic Generalization with Transformerscited by 12
  12. 2024Generalized Planning for the Abstraction and Reasoning Corpuscited by 10
6 more
  1. 2023A Neurodiversity-Inspired Solver for the Abstraction & Reasoning Corpus (ARC) Using Visual Imagery and Program Synthesiscited by 9
  2. 2023Tackling the Abstraction and Reasoning Corpus (ARC) with Object-centric Models and the MDL Principlecited by 7
  3. 2023Infusing Lattice Symmetry Priors in Attention Mechanisms for Sample-Efficient Abstract Geometric Reasoningcited by 4
  4. 2024Speak It Out: Solving Symbol-Related Problems with Symbol-to-Language Conversion for Language Modelscited by 4
  5. 2024Program Synthesis using Inductive Logic Programming for the Abstraction and Reasoning Corpuscited by 4
  6. 2024Fast Analysis of the OpenAI O1-Preview Model in Solving Random K-SAT Problem: Does the LLM Solve the Problem Itself or Call an External SAT Solver?cited by 3

Built on by

Not cited yet.