CogAtom: From Cognitive Atoms to Olympiad-level Mathematical Reasoning in Large Language Models

• URL: http://arxiv.org/abs/2509.17318v2
• Date: Wed, 24 Sep 2025 07:23:31 GMT
• Title: CogAtom: From Cognitive Atoms to Olympiad-level Mathematical Reasoning in Large Language Models
• Authors: Zhuofan Chen, Jiyuan He, Yichi Zhang, Xing Hu, Haoxing Wen, Jun Bai, Wenge Rong,
• Abstract summary: We introduce CogAtom, a novel cognitive atom-based framework for synthesizing rigorous and cognitively diverse problems.<n>Our work offers a cognitively grounded pathway toward scalable, high-quality math problem generation.
• Score: 16.00687492947949
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Mathematical reasoning poses significant challenges for Large Language Models (LLMs) due to its demand for multi-step reasoning and abstract conceptual integration. While recent test-time scaling techniques rely heavily on high-quality, challenging problems, the scarcity of Olympiad-level math problems remains a bottleneck. We introduce CogAtom, a novel cognitive atom-based framework for synthesizing mathematically rigorous and cognitively diverse problems. Unlike prior approaches, CogAtom models problem construction as a process of selecting and recombining fundamental reasoning units, cognitive atoms, extracted from human-authored solutions. A diversity-promoting random walk algorithm enables exploration of the cognitive atom space, while a constraint-based recombination mechanism ensures logical soundness and structural validity. The combinatorial nature of the graph structure provides a near-infinite space of reasoning paths, and the walk algorithm systematically explores this space to achieve large-scale synthesis of high-quality problems; meanwhile, by controlling the number of cognitive atoms, we can precisely adjust problem difficulty, ensuring diversity, scalability, and controllability of the generated problems. Experimental results demonstrate that CogAtom outperforms existing methods in accuracy, reasoning depth, and diversity, generating problems that closely match the difficulty of AIME while exceeding it in structural variation. Our work offers a cognitively grounded pathway toward scalable, high-quality math problem generation.Our code is publicly available at https://github.com/Icarus-1111/CogAtom.

Related papers

• Lightning Grasp: High Performance Procedural Grasp Synthesis with Contact Fields [55.014365168982884]
  We present Lightning Grasp, a novel high-performance procedural grasp synthesis algorithm.<n>It achieves orders-of-magnitude speedups over state-of-the-art approaches.<n>We open-source our system to propel further innovation in robotic manipulation.
  arXiv  Detail & Related papers  (2025-11-10T18:59:44Z)
• Artificial intelligence for representing and characterizing quantum systems [49.29080693498154]
  Efficient characterization of large-scale quantum systems is a central challenge in quantum science.<n>Recent advances in artificial intelligence (AI) have emerged as a powerful tool to address this challenge.<n>This review discusses how each of these AI paradigms contributes to two core tasks in quantum systems characterization.
  arXiv  Detail & Related papers  (2025-09-05T08:41:24Z)
• Scaling Neuro-symbolic Problem Solving: Solver-Free Learning of Constraints and Objectives [3.2198807619678074]
  We introduce a differentiable neuro-symbolic architecture and a loss function dedicated to learning how to solve NP-hard reasoning problems.<n>We empirically show that it can efficiently learn how to solve NP-hard reasoning problems from natural inputs.
  arXiv  Detail & Related papers  (2025-08-28T16:33:27Z)
• EvolMathEval: Towards Evolvable Benchmarks for Mathematical Reasoning via Evolutionary Testing [26.594095530084996]
  This paper introduces EvolMathEval, an automated mathematical benchmark generation and evolution framework based on evolutionary testing.<n>By dynamically generating unique evaluation instances ab initio, the framework fundamentally eliminates the risk of data contamination.<n>EvolMathEval can not only generate a large volume of high-difficulty problems through continuous self-iteration, but it can also significantly enhance the complexity of public datasets.
  arXiv  Detail & Related papers  (2025-08-18T15:24:10Z)
• OMEGA: Can LLMs Reason Outside the Box in Math? Evaluating Exploratory, Compositional, and Transformative Generalization [88.76091817642963]
  Recent large-scale language models (LLMs) with long Chain-of-such reasoning as DeepSeek-R1-have achieved impressive results on Olympiad-level mathematics.<n>We introduce OMEGA-Out-of-distribution Math Problems Evaluation with 3 Generalization Axes-a benchmark designed to evaluate three axes of out-of-distribution generalization.
  arXiv  Detail & Related papers  (2025-06-23T17:51:40Z)
• Can Atomic Step Decomposition Enhance the Self-structured Reasoning of Multimodal Large Models? [68.72260770171212]
  We propose a paradigm of Self-structured Chain of Thought (SCoT), which is composed of minimal semantic atomic steps.<n>Our method can not only generate cognitive CoT structures for various complex tasks but also mitigates the phenomenon of overthinking.<n>We conduct extensive experiments to show that the proposed AtomThink significantly improves the performance of baseline MLLMs.
  arXiv  Detail & Related papers  (2025-03-08T15:23:47Z)
• PromptCoT: Synthesizing Olympiad-level Problems for Mathematical Reasoning in Large Language Models [59.920971312822736]
  We introduce PromptCoT, a novel approach for automatically generating high-quality Olympiad-level math problems.<n>The proposed method synthesizes complex problems based on mathematical concepts and the rationale behind problem construction.<n>Our method is evaluated on standard benchmarks including GSM8K, MATH-500, and AIME2024, where it consistently outperforms existing problem generation methods.
  arXiv  Detail & Related papers  (2025-03-04T06:32:30Z)
• AtomThink: Multimodal Slow Thinking with Atomic Step Reasoning [68.65389926175506]
  We propose a novel paradigm of Self-structured Chain of Thought (SCoT)<n>Our method can not only generate cognitive CoT structures for various complex tasks but also mitigates the phenomena of overthinking for easier tasks.<n>We conduct extensive experiments to show that the proposed AtomThink significantly improves the performance of baseline MLLMs.
  arXiv  Detail & Related papers  (2024-11-18T11:54:58Z)
• Unitary Complexity and the Uhlmann Transformation Problem [39.6823854861458]
  We introduce a framework for unitary synthesis problems, including notions of reductions and unitary complexity classes.<n>We use this framework to study the complexity of transforming one entangled state into another via local operations.<n>Our framework for unitary complexity thus provides new avenues for studying the computational complexity of many natural quantum information processing tasks.
  arXiv  Detail & Related papers  (2023-06-22T17:46:39Z)
• Scalable Coupling of Deep Learning with Logical Reasoning [0.0]
  We introduce a scalable neural architecture and loss function dedicated to learning the constraints and criteria of NP-hard reasoning problems. Our loss function solves one of the main limitations of Besag's pseudo-loglikelihood, enabling learning of high energies.
  arXiv  Detail & Related papers  (2023-05-12T17:09:34Z)
• Learning ground states of quantum Hamiltonians with graph networks [6.024776891570197]
  Solving for the lowest energy eigenstate of the many-body Schrodinger equation is a cornerstone problem. Variational methods approach this problem by searching for the best approximation within a lower-dimensional variational manifold. We use graph neural networks to define a structured variational manifold and optimize its parameters to find high quality approximations.
  arXiv  Detail & Related papers  (2021-10-12T22:56:16Z)
• Long-Time Error-Mitigating Simulation of Open Quantum Systems on Near Term Quantum Computers [38.860468003121404]
  We study an open quantum system simulation on quantum hardware, which demonstrates robustness to hardware errors even with deep circuits containing up to two thousand entangling gates. We simulate two systems of electrons coupled to an infinite thermal bath: 1) a system of dissipative free electrons in a driving electric field; and 2) the thermalization of two interacting electrons in a single orbital in a magnetic field -- the Hubbard atom. Our results demonstrate that algorithms for simulating open quantum systems are able to far outperform similarly complex non-dissipative algorithms on noisy hardware.
  arXiv  Detail & Related papers  (2021-08-02T21:36:37Z)
• Nature-Inspired Optimization Algorithms: Challenges and Open Problems [3.7692411550925673]
  Problems in science and engineering can be formulated as optimization problems, subject to complex nonlinear constraints. The solutions of highly nonlinear problems usually require sophisticated optimization algorithms, and traditional algorithms may struggle to deal with such problems. A current trend is to use nature-inspired algorithms due to their flexibility and effectiveness.
  arXiv  Detail & Related papers  (2020-03-08T13:00:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.