# Tell me why! Explanations support learning relational and causal structure

arXiv: https://arxiv.org/abs/2112.03753 · HTML: https://mlsanity.com/paper/2112.03753/

## Key facts

- Cited by 36 papers in the map (data through 14 Oct 2026).
- Builds on 40 of its 88 references that are in the map.
- First posted on arXiv 7 Dec 2021.

## Builds on (most cited first, 30 of 40)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures](https://mlsanity.com/paper/1802.01561/) (2018)
- [e-SNLI: Natural Language Inference with Natural Language Explanations](https://mlsanity.com/paper/1812.01193/) (2018)
- [Reinforcement Learning with Unsupervised Auxiliary Tasks](https://mlsanity.com/paper/1611.05397/) (2016)
- [Efficient Off-Policy Meta-Reinforcement Learning via Probabilistic Context Variables](https://mlsanity.com/paper/1903.08254/) (2019)
- [A simple neural network module for relational reasoning](https://mlsanity.com/paper/1706.01427/) (2017)
- [Right for the Right Reasons: Training Differentiable Models by Constraining their Explanations](https://mlsanity.com/paper/1703.03717/) (2017)
- [Stabilizing Transformers for Reinforcement Learning](https://mlsanity.com/paper/1910.06764/) (2019)
- [Neurosymbolic AI: The 3rd Wave](https://mlsanity.com/paper/2012.05876/) (2020)
- [Learning to Explain: An Information-Theoretic Perspective on Model Interpretation](https://mlsanity.com/paper/1802.07814/) (2018)
- [The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence](https://mlsanity.com/paper/2002.06177/) (2020)
- [Measuring abstract reasoning in neural networks](https://mlsanity.com/paper/1807.04225/) (2018)
- [A Survey of Reinforcement Learning Informed by Natural Language](https://mlsanity.com/paper/1906.03926/) (2019)
- [Making deep neural networks right for the right scientific reasons by interacting with their explanations](https://mlsanity.com/paper/2001.05371/) (2020)
- [Language as an Abstraction for Hierarchical Deep Reinforcement Learning](https://mlsanity.com/paper/1906.07343/) (2019)
- [Explainable Deep Learning: A Field Guide for the Uninitiated](https://mlsanity.com/paper/2004.14545/) (2020)
- [What shapes feature representations? Exploring datasets, architectures, and training](https://mlsanity.com/paper/2006.12433/) (2020)
- [Using Natural Language for Reward Shaping in Reinforcement Learning](https://mlsanity.com/paper/1903.02020/) (2019)
- [Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability](https://mlsanity.com/paper/2107.06277/) (2021)
- [Grounded Language Learning in a Simulated 3D World](https://mlsanity.com/paper/1706.06551/) (2017)
- [Right for the Right Concept: Revising Neuro-Symbolic Concepts by Interacting with their Explanations](https://mlsanity.com/paper/2011.12854/) (2020)
- [Meta reinforcement learning as task inference](https://mlsanity.com/paper/1905.06424/) (2019)
- [Learning with Latent Language](https://mlsanity.com/paper/1711.00482/) (2017)
- [When Can Models Learn From Explanations? A Formal Framework for Understanding the Roles of Explanation Data](https://mlsanity.com/paper/2102.02201/) (2021)
- [Automatic Data Augmentation for Generalization in Deep Reinforcement Learning](https://mlsanity.com/paper/2006.12862/) (2020)
- [Theoretical Impediments to Machine Learning With Seven Sparks from the Causal Revolution](https://mlsanity.com/paper/1801.04016/) (2018)
- [Shaping Belief States with Generative Environment Models for RL](https://mlsanity.com/paper/1906.09237/) (2019)
- [Causal Reasoning from Meta-reinforcement Learning](https://mlsanity.com/paper/1901.08162/) (2019)
- [Explanation-Based Human Debugging of NLP Models: A Survey](https://mlsanity.com/paper/2104.15135/) (2021)

## Built on by (most cited first, 30 of 36)

- [Selection-Inference: Exploiting Large Language Models for Interpretable Logical Reasoning](https://mlsanity.com/paper/2205.09712/) (2022)
- [Can language models learn from explanations in context?](https://mlsanity.com/paper/2204.02329/) (2022)
- [REFINER: Reasoning Feedback on Intermediate Representations](https://mlsanity.com/paper/2304.01904/) (2023)
- [Scaling Instructable Agents Across Many Simulated Worlds](https://mlsanity.com/paper/2404.10179/) (2024)
- [A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains](https://mlsanity.com/paper/2402.00559/) (2024)
- [An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learning](https://mlsanity.com/paper/2302.04419/) (2023)
- [Feedback Descent: Open-Ended Text Optimization via Pairwise Comparison](https://mlsanity.com/paper/2511.07919/) (2025)
- [DriveLMM-o1: A Step-by-Step Reasoning Dataset and Large Multimodal Model for Driving Scenario Understanding](https://mlsanity.com/paper/2503.10621/) (2025)
- [Chain of Thought Imitation with Procedure Cloning](https://mlsanity.com/paper/2205.10816/) (2022)
- [Law Informs Code: A Legal Informatics Approach to Aligning Artificial Intelligence with Humans](https://mlsanity.com/paper/2209.13020/) (2022)
- [Mind the gap: Challenges of deep learning approaches to Theory of Mind](https://mlsanity.com/paper/2203.16540/) (2022)
- [Passive learning of active causal strategies in agents and language models](https://mlsanity.com/paper/2305.16183/) (2023)
- [Using Natural Language and Program Abstractions to Instill Human Inductive Biases in Machines](https://mlsanity.com/paper/2205.11558/) (2022)
- [Abstraction for Deep Reinforcement Learning](https://mlsanity.com/paper/2202.05839/) (2022)
- [Open Domain Question Answering with Conflicting Contexts](https://mlsanity.com/paper/2410.12311/) (2024)
- [Learning by Self-Explaining](https://mlsanity.com/paper/2309.08395/) (2023)
- [Structure in Deep Reinforcement Learning: A Survey and Open Problems](https://mlsanity.com/paper/2306.16021/) (2023)
- [Language and Culture Internalisation for Human-Like Autotelic AI](https://mlsanity.com/paper/2206.01134/) (2022)
- [Imagining and building wise machines: The centrality of AI metacognition](https://mlsanity.com/paper/2411.02478/) (2024)
- [Large Language Models as Fiduciaries: A Case Study Toward Robustly Communicating With Artificial Intelligence Through Legal Standards](https://mlsanity.com/paper/2301.10095/) (2023)
- [Towards a Framework for Evaluating Explanations in Automated Fact Verification](https://mlsanity.com/paper/2403.20322/) (2024)
- [Finetuning CLIP to Reason about Pairwise Differences](https://mlsanity.com/paper/2409.09721/) (2024)
- [Flex-Judge: Text-Only Reasoning Unleashes Zero-Shot Multimodal Evaluators](https://mlsanity.com/paper/2505.18601/) (2025)
- [Learning to Reason With Relational Abstractions](https://mlsanity.com/paper/2210.02615/) (2022)
- [Causality from Bottom to Top: A Survey](https://mlsanity.com/paper/2403.11219/) (2024)
- [Explanation Regeneration via Information Bottleneck](https://mlsanity.com/paper/2212.09603/) (2022)
- [Policy Learning with a Language Bottleneck](https://mlsanity.com/paper/2405.04118/) (2024)
- [Hierarchical Symbolic Reasoning in Hyperbolic Space for Deep Discriminative Models](https://mlsanity.com/paper/2207.01916/) (2022)
- [Hierarchical reinforcement learning with natural language subgoals](https://mlsanity.com/paper/2309.11564/) (2023)
- [Explaining Image Classification with Visual Debates](https://mlsanity.com/paper/2210.09015/) (2022)

## Abstract (arXiv)

Inferring the abstract relational and causal structure of the world is a major challenge for reinforcement-learning (RL) agents. For humans, language--particularly in the form of explanations--plays a considerable role in overcoming this challenge. Here, we show that language can play a similar role for deep RL agents in complex environments. While agents typically struggle to acquire relational and causal knowledge, augmenting their experience by training them to predict language descriptions and explanations can overcome these limitations. We show that language can help agents learn challenging relational tasks, and examine which aspects of language contribute to its benefits. We then show that explanations can help agents to infer not only relational but also causal structure. Language can shape the way that agents to generalize out-of-distribution from ambiguous, causally-confounded training, and explanations even allow agents to learn to perform experimental interventions to identify causal relationships. Our results suggest that language description and explanation may be powerful tools for improving agent learning and generalization.