MINAR: Mechanistic Interpretability for Neural Algorithmic Reasoning

• URL: http://arxiv.org/abs/2602.21442v1
• Date: Tue, 24 Feb 2026 23:38:06 GMT
• Title: MINAR: Mechanistic Interpretability for Neural Algorithmic Reasoning
• Authors: Jesse He, Helen Jenne, Max Vargas, Davis Brown, Gal Mishne, Yusu Wang, Henry Kvinge,
• Abstract summary: We introduce Mechanistic Interpretability for Neural Algorithmic Reasoning (MINAR)<n>MINAR adapts attribution patching methods from mechanistic interpretability to the GNN setting.<n>We show through two case studies that MIAR recovers faithful neuron-level circuits from GNNs trained on algorithmic tasks.
• Score: 17.920835259832238
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The recent field of neural algorithmic reasoning (NAR) studies the ability of graph neural networks (GNNs) to emulate classical algorithms like Bellman-Ford, a phenomenon known as algorithmic alignment. At the same time, recent advances in large language models (LLMs) have spawned the study of mechanistic interpretability, which aims to identify granular model components like circuits that perform specific computations. In this work, we introduce Mechanistic Interpretability for Neural Algorithmic Reasoning (MINAR), an efficient circuit discovery toolbox that adapts attribution patching methods from mechanistic interpretability to the GNN setting. We show through two case studies that MINAR recovers faithful neuron-level circuits from GNNs trained on algorithmic tasks. Our study sheds new light on the process of circuit formation and pruning during training, as well as giving new insight into how GNNs trained to perform multiple tasks in parallel reuse circuit components for related tasks. Our code is available at https://github.com/pnnl/MINAR.

Related papers

• Which Algorithms Can Graph Neural Networks Learn? [14.935565203071528]
  We propose a general theoretical framework that characterizes the sufficient conditions under which MPNNs can learn an algorithm.<n>Our framework applies to a broad class of algorithms, including single-source shortest paths, minimum spanning trees, and general dynamic programming problems.
  arXiv  Detail & Related papers  (2026-02-13T17:09:50Z)
• Mind The Gap: Deep Learning Doesn't Learn Deeply [16.284360949127723]
  This paper aims to understand how neural networks learn algorithmic reasoning by addressing two questions.<n>How faithful are learned algorithms when they are effective, and why do neural networks fail to learn effective algorithms otherwise?
  arXiv  Detail & Related papers  (2025-05-24T10:11:36Z)
• Learning to Add, Multiply, and Execute Algorithmic Instructions Exactly with Neural Networks [13.351162559710124]
  We study the training dynamics of two-layer fully connected networks in the infinite-width limit.<n>We show how a sufficiently large ensemble of such models can be trained to execute exactly, with high probability.<n>We show how this can be efficiently achieved using only logarithmically many training data.
  arXiv  Detail & Related papers  (2025-02-24T00:50:02Z)
• NAR-*ICP: Neural Execution of Classical ICP-based Pointcloud Registration Algorithms [16.025166074715816]
  This study explores the intersection of neural networks and classical robotics algorithms through the Neural Algorithmic Reasoning (NAR) blueprint.<n>We propose a novel Graph Neural Network (GNN)-based framework, NAR-*ICP, that learns the intermediate computations of classical ICP-based registration algorithms.<n>We evaluate our approach across real-world and synthetic datasets, demonstrating its flexibility in handling complex inputs.
  arXiv  Detail & Related papers  (2024-10-14T19:33:46Z)
• Scalable Mechanistic Neural Networks for Differential Equations and Machine Learning [52.28945097811129]
  We propose an enhanced neural network framework designed for scientific machine learning applications involving long temporal sequences.<n>We reduce the computational time and space complexities from cubic and quadratic with respect to the sequence length, respectively, to linear.<n>Extensive experiments demonstrate that S-MNN matches the original MNN in precision while substantially reducing computational resources.
  arXiv  Detail & Related papers  (2024-10-08T14:27:28Z)
• How Graph Neural Networks Learn: Lessons from Training Dynamics [80.41778059014393]
  We study the training dynamics in function space of graph neural networks (GNNs) We find that the gradient descent optimization of GNNs implicitly leverages the graph structure to update the learned function. This finding offers new interpretable insights into when and why the learned GNN functions generalize.
  arXiv  Detail & Related papers  (2023-10-08T10:19:56Z)
• The Clock and the Pizza: Two Stories in Mechanistic Explanation of Neural Networks [59.26515696183751]
  We show that algorithm discovery in neural networks is sometimes more complex. We show that even simple learning problems can admit a surprising diversity of solutions.
  arXiv  Detail & Related papers  (2023-06-30T17:59:13Z)
• How neural networks learn to classify chaotic time series [77.34726150561087]
  We study the inner workings of neural networks trained to classify regular-versus-chaotic time series. We find that the relation between input periodicity and activation periodicity is key for the performance of LKCNN models.
  arXiv  Detail & Related papers  (2023-06-04T08:53:27Z)
• Permutation Equivariant Neural Functionals [92.0667671999604]
  This work studies the design of neural networks that can process the weights or gradients of other neural networks. We focus on the permutation symmetries that arise in the weights of deep feedforward networks because hidden layer neurons have no inherent order. In our experiments, we find that permutation equivariant neural functionals are effective on a diverse set of tasks.
  arXiv  Detail & Related papers  (2023-02-27T18:52:38Z)
• Neural Algorithmic Reasoning with Causal Regularisation [18.299363749150093]
  We make an important observation: there are many different inputs for which an algorithm will perform certain intermediate computations identically. This insight allows us to develop data augmentation procedures that, given an algorithm's intermediate trajectory, produce inputs for which the target algorithm would have exactly the same next trajectory step. We prove that the resulting method, which we call Hint-ReLIC, improves the OOD generalisation capabilities of the reasoner.
  arXiv  Detail & Related papers  (2023-02-20T19:41:15Z)
• Intelligence Processing Units Accelerate Neuromorphic Learning [52.952192990802345]
  Spiking neural networks (SNNs) have achieved orders of magnitude improvement in terms of energy consumption and latency. We present an IPU-optimized release of our custom SNN Python package, snnTorch.
  arXiv  Detail & Related papers  (2022-11-19T15:44:08Z)
• Graph Neural Networks are Dynamic Programmers [0.0]
  Graph neural networks (GNNs) are claimed to align with dynamic programming (DP) Here we show, using methods from theory and abstract algebra, that there exists an intricate connection between GNNs and DP.
  arXiv  Detail & Related papers  (2022-03-29T13:27:28Z)
• Recurrent Neural Network Learning of Performance and Intrinsic Population Dynamics from Sparse Neural Data [77.92736596690297]
  We introduce a novel training strategy that allows learning not only the input-output behavior of an RNN but also its internal network dynamics. We test the proposed method by training an RNN to simultaneously reproduce internal dynamics and output signals of a physiologically-inspired neural model. Remarkably, we show that the reproduction of the internal dynamics is successful even when the training algorithm relies on the activities of a small subset of neurons.
  arXiv  Detail & Related papers  (2020-05-05T14:16:54Z)

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.