# Pragmatic Instruction Following and Goal Assistance via Cooperative Language-Guided Inverse Planning

arXiv: https://arxiv.org/abs/2402.17930 · HTML: https://mlsanity.com/paper/2402.17930/

## Key facts

- Cited by 28 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 84 references that are in the map.
- First posted on arXiv 27 Feb 2024.

## Builds on (most cited first, 27 of 27)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Do As I Can, Not As I Say: Grounding Language in Robotic Affordances](https://mlsanity.com/paper/2204.01691/) (2022)
- [VirtualHome: Simulating Household Activities via Programs](https://mlsanity.com/paper/1806.07011/) (2018)
- [Language to Rewards for Robotic Skill Synthesis](https://mlsanity.com/paper/2306.08647/) (2023)
- [Cooperative Inverse Reinforcement Learning](https://mlsanity.com/paper/1606.03137/) (2016)
- [Constrained Language Models Yield Few-Shot Semantic Parsers](https://mlsanity.com/paper/2104.08768/) (2021)
- [Watch-And-Help: A Challenge for Social Perception and Human-AI Collaboration](https://mlsanity.com/paper/2010.09890/) (2020)
- [Reward-rational (implicit) choice: A unifying formalism for reward learning](https://mlsanity.com/paper/2002.04833/) (2020)
- [Minding Language Models' (Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker](https://mlsanity.com/paper/2306.00924/) (2023)
- [MMToM-QA: Multimodal Theory of Mind Question Answering](https://mlsanity.com/paper/2401.08743/) (2024)
- [Online algorithms for POMDPs with continuous state, action, and observation spaces](https://mlsanity.com/paper/1709.06196/) (2017)
- [Sequential Monte Carlo Steering of Large Language Models using Probabilistic Programs](https://mlsanity.com/paper/2306.03081/) (2023)
- [Online Bayesian Goal Inference for Boundedly-Rational Planning Agents](https://mlsanity.com/paper/2006.07532/) (2020)
- [From Language to Goals: Inverse Reinforcement Learning for Vision-Based Instruction Following](https://mlsanity.com/paper/1902.07742/) (2019)
- [Unified Pragmatic Models for Generating and Following Instructions](https://mlsanity.com/paper/1711.04987/) (2017)
- [Inferring Rewards from Language in Context](https://mlsanity.com/paper/2204.02515/) (2022)
- [Rao-Blackwellised Particle Filtering for Dynamic Bayesian Networks](https://mlsanity.com/paper/1301.3853/) (2013)
- [Pragmatic-Pedagogic Value Alignment](https://mlsanity.com/paper/1707.06354/) (2017)
- [The Neuro-Symbolic Inverse Planning Engine (NIPE): Modeling Probabilistic Social Inferences from Linguistic Inputs](https://mlsanity.com/paper/2306.14325/) (2023)
- [HandMeThat: Human-Robot Communication in Physical and Social Environments](https://mlsanity.com/paper/2310.03779/) (2023)
- [Modeling the Mistakes of Boundedly Rational Agents Within a Bayesian Theory of Mind](https://mlsanity.com/paper/2106.13249/) (2021)
- [Inferring the Goals of Communicating Agents from Actions and Instructions](https://mlsanity.com/paper/2306.16207/) (2023)
- [Literal or Pedagogic Human? Analyzing Human Model Misspecification in Objective Learning](https://mlsanity.com/paper/1903.03877/) (2019)
- [Reward Learning from Narrated Demonstrations](https://mlsanity.com/paper/1804.10692/) (2018)
- [Monte Carlo Sampling Methods for Approximating Interactive POMDPs](https://mlsanity.com/paper/1401.3455/) (2014)
- [Planning With Uncertain Specifications (PUnS)](https://mlsanity.com/paper/1906.03218/) (2019)
- [Solving the Baby Intuitions Benchmark with a Hierarchically Bayesian Theory of Mind](https://mlsanity.com/paper/2208.02914/) (2022)

## Built on by (most cited first, 28 of 28)

- [Building Machines that Learn and Think with People](https://mlsanity.com/paper/2408.03943/) (2024)
- [Large Language Models Assume People are More Rational than We Really are](https://mlsanity.com/paper/2406.17055/) (2024)
- [Beyond Preferences in AI Alignment](https://mlsanity.com/paper/2408.16984/) (2024)
- [Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlo](https://mlsanity.com/paper/2504.13139/) (2025)
- [Hypothesis-Driven Theory-of-Mind Reasoning for Large Language Models](https://mlsanity.com/paper/2502.11881/) (2025)
- [On Benchmarking Human-Like Intelligence in Machines](https://mlsanity.com/paper/2502.20502/) (2025)
- [AutoToM: Scaling Model-based Mental Inference via Automated Agent Modeling](https://mlsanity.com/paper/2502.15676/) (2025)
- [GOMA: Proactive Embodied Cooperative Communication via Goal-Oriented Mental Alignment](https://mlsanity.com/paper/2403.11075/) (2024)
- [M3HF: Multi-agent Reinforcement Learning from Multi-phase Human Feedback of Mixed Quality](https://mlsanity.com/paper/2503.02077/) (2025)
- [AssistanceZero: Scalably Solving Assistance Games](https://mlsanity.com/paper/2504.07091/) (2025)
- [Resource Rational Contractualism Should Guide AI Alignment](https://mlsanity.com/paper/2506.17434/) (2025)
- [Structured Preference Optimization for Vision-Language Long-Horizon Task Planning](https://mlsanity.com/paper/2502.20742/) (2025)
- [Improving Large Language Model Planning with Action Sequence Similarity](https://mlsanity.com/paper/2505.01009/) (2025)
- [Modeling Others' Minds as Code](https://mlsanity.com/paper/2510.01272/) (2025)
- [Infinite Ends from Finite Samples: Open-Ended Goal Inference as Top-Down Bayesian Filtering of Bottom-Up Proposals](https://mlsanity.com/paper/2407.16770/) (2024)
- [Understanding Epistemic Language with a Language-augmented Bayesian Theory of Mind](https://mlsanity.com/paper/2408.12022/) (2024)
- [Pragmatic Embodied Spoken Instruction Following in Human-Robot Collaboration with Theory of Mind](https://mlsanity.com/paper/2409.10849/) (2024)
- [Retrieval-augmented systems can be dangerous medical communicators](https://mlsanity.com/paper/2502.14898/) (2025)
- [Goal Inference from Open-Ended Dialog](https://mlsanity.com/paper/2410.13957/) (2024)
- [Towards Machine Theory of Mind with Large Language Model-Augmented Inverse Planning](https://mlsanity.com/paper/2507.03682/) (2025)
- [Representational Alignment Supports Effective Machine Teaching](https://mlsanity.com/paper/2406.04302/) (2024)
- [Gricean Norms as a Basis for Effective Collaboration](https://mlsanity.com/paper/2503.14484/) (2025)
- [Language and Experience: A Computational Model of Social Learning in Complex Tasks](https://mlsanity.com/paper/2509.00074/) (2025)
- [Grounding Social Perception in Intuitive Physics](https://mlsanity.com/paper/2603.27410/) (2026)
- [Planning Ahead with RSA: Efficient Signalling in Dynamic Environments by Projecting User Awareness across Future Timesteps](https://mlsanity.com/paper/2510.23340/) (2025)
- [Realist and Pluralist Conceptions of Intelligence and Their Implications on AI Research](https://mlsanity.com/paper/2511.15282/) (2025)
- [When Contextual Inference Fails: Cancelability in Interactive Instruction Following](https://mlsanity.com/paper/2603.19997/) (2026)
- [MindZero: Learning Online Mental Reasoning With Zero Annotations](https://mlsanity.com/paper/2606.00240/) (2026)

## Abstract (arXiv)

People often give instructions whose meaning is ambiguous without further context, expecting that their actions or goals will disambiguate their intentions. How can we build assistive agents that follow such instructions in a flexible, context-sensitive manner? This paper introduces cooperative language-guided inverse plan search (CLIPS), a Bayesian agent architecture for pragmatic instruction following and goal assistance. Our agent assists a human by modeling them as a cooperative planner who communicates joint plans to the assistant, then performs multimodal Bayesian inference over the human's goal from actions and language, using large language models (LLMs) to evaluate the likelihood of an instruction given a hypothesized plan. Given this posterior, our assistant acts to minimize expected goal achievement cost, enabling it to pragmatically follow ambiguous instructions and provide effective assistance even when uncertain about the goal. We evaluate these capabilities in two cooperative planning domains (Doors, Keys & Gems and VirtualHome), finding that CLIPS significantly outperforms GPT-4V, LLM-based literal instruction following and unimodal inverse planning in both accuracy and helpfulness, while closely matching the inferences and assistive judgments provided by human raters.