# Translating Neuralese

arXiv: https://arxiv.org/abs/1704.06960 · HTML: https://mlsanity.com/paper/1704.06960/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Apr 2017.

## Built on by (most cited first, 16 of 16)

- [Scalable agent alignment via reward modeling: a research direction](https://mlsanity.com/paper/1811.07871/) (2018)
- [Post-hoc Interpretability for Neural NLP: A Survey](https://mlsanity.com/paper/2108.04840/) (2021)
- [Natural Language Descriptions of Deep Visual Features](https://mlsanity.com/paper/2201.11114/) (2022)
- [Inferring Rewards from Language in Context](https://mlsanity.com/paper/2204.02515/) (2022)
- [The Who in XAI: How AI Background Shapes Perceptions of AI Explanations](https://mlsanity.com/paper/2107.13509/) (2021)
- [Linking Emergent and Natural Languages via Corpus Transfer](https://mlsanity.com/paper/2203.13344/) (2022)
- [Communicating Activations Between Language Model Agents](https://mlsanity.com/paper/2501.14082/) (2025)
- [Emergent Communication in Multi-Agent Reinforcement Learning for Future Wireless Networks](https://mlsanity.com/paper/2309.06021/) (2023)
- [First Contact: Unsupervised Human-Machine Co-Adaptation via Mutual Information Maximization](https://mlsanity.com/paper/2205.12381/) (2022)
- [Interpretable Learned Emergent Communication for Human-Agent Teams](https://mlsanity.com/paper/2201.07452/) (2022)
- [Sometimes We Want Translationese](https://mlsanity.com/paper/2104.07623/) (2021)
- [TexRel: a Green Family of Datasets for Emergent Communications on Relations](https://mlsanity.com/paper/2105.12804/) (2021)
- [Unsupervised Translation of Emergent Communication](https://mlsanity.com/paper/2502.07552/) (2025)
- [CB2: Collaborative Natural Language Interaction Research Platform](https://mlsanity.com/paper/2303.08127/) (2023)
- [Compositionality Through Language Transmission, using Artificial Neural Networks](https://mlsanity.com/paper/2101.11739/) (2021)
- [Co-Alignment: Rethinking Alignment as Bidirectional Human-AI Cognitive Adaptation](https://mlsanity.com/paper/2509.12179/) (2025)

## Abstract (arXiv)

Several approaches have recently been proposed for learning decentralized deep multiagent policies that coordinate via a differentiable communication channel. While these policies are effective for many tasks, interpretation of their induced communication strategies has remained a challenge. Here we propose to interpret agents' messages by translating them. Unlike in typical machine translation problems, we have no parallel data to learn from. Instead we develop a translation model based on the insight that agent messages and natural language strings mean the same thing if they induce the same belief about the world in a listener. We present theoretical guarantees and empirical evidence that our approach preserves both the semantics and pragmatics of messages by ensuring that players communicating through a translation layer do not suffer a substantial loss in reward relative to players with a common language.