# Prospective Learning: Principled Extrapolation to the Future

arXiv: https://arxiv.org/abs/2201.07372 · HTML: https://mlsanity.com/paper/2201.07372/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 57 references that are in the map.
- First posted on arXiv 19 Jan 2022.

## Builds on (most cited first, 20 of 20)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Towards Deep Learning Models Resistant to Adversarial Attacks](https://mlsanity.com/paper/1706.06083/) (2017)
- [Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks](https://mlsanity.com/paper/1703.03400/) (2017)
- [Three scenarios for continual learning](https://mlsanity.com/paper/1904.07734/) (2019)
- [Generalizing from a Few Examples: A Survey on Few-Shot Learning](https://mlsanity.com/paper/1904.05046/) (2019)
- [Towards Out-Of-Distribution Generalization: A Survey](https://mlsanity.com/paper/2108.13624/) (2021)
- [Online Learning: A Modern Introduction Using Convex Optimization](https://mlsanity.com/paper/1912.13213/) (2019)
- [Meta-Dataset: A Dataset of Datasets for Learning to Learn from Few Examples](https://mlsanity.com/paper/1903.03096/) (2019)
- [Meta-Learning: A Survey](https://mlsanity.com/paper/1810.03548/) (2018)
- [Mind the Gap: Assessing Temporal Generalization in Neural Language Models](https://mlsanity.com/paper/2102.01951/) (2021)
- [Online Learning: A Comprehensive Survey](https://mlsanity.com/paper/1802.02871/) (2018)
- [REMIND Your Neural Network to Prevent Catastrophic Forgetting](https://mlsanity.com/paper/1910.02509/) (2019)
- [A Tour of Reinforcement Learning: The View from Continuous Control](https://mlsanity.com/paper/1806.09460/) (2018)
- [Strike (with) a Pose: Neural Networks Are Easily Fooled by Strange Poses of Familiar Objects](https://mlsanity.com/paper/1811.11553/) (2018)
- [OoD-Bench: Quantifying and Understanding Two Dimensions of Out-of-Distribution Generalization](https://mlsanity.com/paper/2106.03721/) (2021)
- [Out of Distribution Generalization in Machine Learning](https://mlsanity.com/paper/2103.02667/) (2021)
- [Model Zoo: A Growing "Brain" That Learns Continually](https://mlsanity.com/paper/2106.03027/) (2021)
- [Meta-learning in natural and artificial intelligence](https://mlsanity.com/paper/2011.13464/) (2020)
- [Online Structured Meta-learning](https://mlsanity.com/paper/2010.11545/) (2020)
- [Towards a theory of out-of-distribution learning](https://mlsanity.com/paper/2109.14501/) (2021)

## Built on by (most cited first, 6 of 6)

- [Latent learning: episodic memory complements parametric learning by enabling flexible reuse of experiences](https://mlsanity.com/paper/2509.16189/) (2025)
- [Transformers for Supervised Online Continual Learning](https://mlsanity.com/paper/2403.01554/) (2024)
- [Prospective Learning: Learning for a Dynamic Future](https://mlsanity.com/paper/2411.00109/) (2024)
- [Causality in the human niche: lessons for machine learning](https://mlsanity.com/paper/2506.13803/) (2025)
- [Prospective Learning in Retrospect](https://mlsanity.com/paper/2507.07965/) (2025)
- [Learning Model Successors](https://mlsanity.com/paper/2502.00197/) (2025)

## Abstract (arXiv)

Learning is a process which can update decision rules, based on past experience, such that future performance improves. Traditionally, machine learning is often evaluated under the assumption that the future will be identical to the past in distribution or change adversarially. But these assumptions can be either too optimistic or pessimistic for many problems in the real world. Real world scenarios evolve over multiple spatiotemporal scales with partially predictable dynamics. Here we reformulate the learning problem to one that centers around this idea of dynamic futures that are partially learnable. We conjecture that certain sequences of tasks are not retrospectively learnable (in which the data distribution is fixed), but are prospectively learnable (in which distributions may be dynamic), suggesting that prospective learning is more difficult in kind than retrospective learning. We argue that prospective learning more accurately characterizes many real world problems that (1) currently stymie existing artificial intelligence solutions and/or (2) lack adequate explanations for how natural intelligences solve them. Thus, studying prospective learning will lead to deeper insights and solutions to currently vexing challenges in both natural and artificial intelligences.