# Evolutionary Training and Abstraction Yields Algorithmic Generalization of Neural Computers

arXiv: https://arxiv.org/abs/2105.07957 · HTML: https://mlsanity.com/paper/2105.07957/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 28 of its 57 references that are in the map.
- First posted on arXiv 17 May 2021.

## Builds on (most cited first, 28 of 28)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Continual Lifelong Learning with Neural Networks: A Review](https://mlsanity.com/paper/1802.07569/) (2018)
- [Deep Learning in Neural Networks: An Overview](https://mlsanity.com/paper/1404.7828/) (2014)
- [Neural Turing Machines](https://mlsanity.com/paper/1410.5401/) (2014)
- [Deep learning for time series classification: a review](https://mlsanity.com/paper/1809.04356/) (2018)
- [Evolution Strategies as a Scalable Alternative to Reinforcement Learning](https://mlsanity.com/paper/1703.03864/) (2017)
- [Deep Learning for Generic Object Detection: A Survey](https://mlsanity.com/paper/1809.02165/) (2018)
- [Memory Networks](https://mlsanity.com/paper/1410.3916/) (2014)
- [Neural Logic Machines](https://mlsanity.com/paper/1904.11694/) (2019)
- [Neural GPUs Learn Algorithms](https://mlsanity.com/paper/1511.08228/) (2015)
- [Neural Execution of Graph Algorithms](https://mlsanity.com/paper/1910.10593/) (2019)
- [Ask Me Anything: Dynamic Memory Networks for Natural Language Processing](https://mlsanity.com/paper/1506.07285/) (2015)
- [Improving Exploration in Evolution Strategies for Deep Reinforcement Learning via a Population of Novelty-Seeking Agents](https://mlsanity.com/paper/1712.06560/) (2017)
- [Neural Programmer-Interpreters](https://mlsanity.com/paper/1511.06279/) (2015)
- [Building machines that adapt and compute like brains](https://mlsanity.com/paper/1711.04203/) (2017)
- [Inferring Algorithmic Patterns with Stack-Augmented Recurrent Nets](https://mlsanity.com/paper/1503.01007/) (2015)
- [Neural Arithmetic Logic Units](https://mlsanity.com/paper/1808.00508/) (2018)
- [Unsupervised Predictive Memory in a Goal-Directed Agent](https://mlsanity.com/paper/1803.10760/) (2018)
- [Learning to Transduce with Unbounded Memory](https://mlsanity.com/paper/1506.02516/) (2015)
- [Making Neural Programming Architectures Generalize via Recursion](https://mlsanity.com/paper/1704.06611/) (2017)
- [Neural Programmer: Inducing Latent Programs with Gradient Descent](https://mlsanity.com/paper/1511.04834/) (2015)
- [Neural Random-Access Machines](https://mlsanity.com/paper/1511.06392/) (2015)
- [Neural Arithmetic Units](https://mlsanity.com/paper/2001.05016/) (2020)
- [Neural Stored-program Memory](https://mlsanity.com/paper/1906.08862/) (2019)
- [Learning Simple Algorithms from Examples](https://mlsanity.com/paper/1511.07275/) (2015)
- [Learning Algorithmic Solutions to Symbolic Planning Tasks with a Neural Computer Architecture](https://mlsanity.com/paper/1911.00926/) (2019)
- [Feed-Forward Neural Networks Need Inductive Bias to Learn Equality Relations](https://mlsanity.com/paper/1812.01662/) (2018)
- [HyperENTM: Evolving Scalable Neural Turing Machines through HyperNEAT](https://mlsanity.com/paper/1710.04748/) (2017)

## Built on by (most cited first, 5 of 5)

- [Neuro-Symbolic Imitation Learning: Discovering Symbolic Abstractions for Skill Learning](https://mlsanity.com/paper/2503.21406/) (2025)
- [Learning Type-Generalized Actions for Symbolic Planning](https://mlsanity.com/paper/2308.04867/) (2023)
- [DNCs Require More Planning Steps](https://mlsanity.com/paper/2406.02187/) (2024)
- [Local Pairwise Distance Matching for Backpropagation-Free Reinforcement Learning](https://mlsanity.com/paper/2507.11367/) (2025)
- [Embodied Neurocomputation: A Framework for Interfacing Biological Neural Cultures with Scaled Task-Driven Validation](https://mlsanity.com/paper/2605.13315/) (2026)

## Abstract (arXiv)

A key feature of intelligent behaviour is the ability to learn abstract strategies that scale and transfer to unfamiliar problems. An abstract strategy solves every sample from a problem class, no matter its representation or complexity -- like algorithms in computer science. Neural networks are powerful models for processing sensory data, discovering hidden patterns, and learning complex functions, but they struggle to learn such iterative, sequential or hierarchical algorithmic strategies. Extending neural networks with external memories has increased their capacities in learning such strategies, but they are still prone to data variations, struggle to learn scalable and transferable solutions, and require massive training data. We present the Neural Harvard Computer (NHC), a memory-augmented network based architecture, that employs abstraction by decoupling algorithmic operations from data manipulations, realized by splitting the information flow and separated modules. This abstraction mechanism and evolutionary training enable the learning of robust and scalable algorithmic solutions. On a diverse set of 11 algorithms with varying complexities, we show that the NHC reliably learns algorithmic solutions with strong generalization and abstraction: perfect generalization and scaling to arbitrary task configurations and complexities far beyond seen during training, and being independent of the data representation and the task domain.