Language and Experience: A Computational Model of Social Learning in Complex Tasks

2025/08/26 by Cédric Colas, Tracey Mills, Colas, Cédric +11
Neuroscience · Psychology · Social Sciences · #Artificial Intelligence (cs.AI) #Child and Animal Learning Development #Computation and Language (cs.CL) #Embodied and Extended Cognition #FOS: Computer and information sciences #Language and cultural evolution #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2509.00074

openalex publication_date 2025/08/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

The ability to combine linguistic guidance from others with direct experience is central to human development, enabling safe and rapid learning in new environments. How do people integrate these two sources of knowledge, and how might AI systems? We present a computational framework that models social learning as joint probabilistic inference over structured, executable world models given sensorimotor and linguistic data. We make this possible by turning a pretrained language model into a probabilistic model of how humans share advice conditioned on their beliefs, allowing our agents both to generate advice for others and to interpret linguistic input as evidence during Bayesian inference. Using behavioral experiments and simulations across 10 video games, we show how linguistic guidance can shape exploration and accelerate learning by reducing risky interactions and speeding up key discoveries in both humans and models. We further explore how knowledge can accumulate across generations through iterated learning experiments and demonstrate successful knowledge transfer between humans and models -- revealing how structured, language-compatible representations might enable human-machine collaborative learning.

Citations

• Code World Models for General Game Playing
• Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlo
• BALROG: Benchmarking Agentic LLM and VLM Reasoning On Games
• Building Machines that Learn and Think with People
• Pragmatic Instruction Following and Goal Assistance via Cooperative Language-Guided Inverse Planning
• WorldCoder, a Model-Based LLM Agent: Building World Models by Writing Code and Interacting with the Environment
• Machine Culture
• Can Large Language Models Really Improve by Self-critiquing Their Own Plans?
• Inferring the Goals of Communicating Agents from Actions and Instructions
• Sequential Monte Carlo Steering of Large Language Models using Probabilistic Programs
• Generalized Planning in PDDL Domains with Pretrained Large Language Models
• Do Embodied Agents Dream of Pixelated Sheep: Embodied Decision Making using Language Guided World Modelling
• Distributing Accountability, Not Capability: Phase Separation and the LLM Workflow Quadrant in Autonomous AI Agent Architectures
• Do As I Can, Not As I Say: Grounding Language in Robotic Affordances
• BNAI, NO-TOKEN, and MIND-UNITY: Pillars of a Systemic Revolution in Artificial Intelligence
• Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents
• Show Your Work: Scratchpads for Intermediate Computation with Language Models
• Skill Induction and Planning with Latent Language
• Learning to solve complex tasks by growing knowledge culturally across generations
• Human-Level Reinforcement Learning through Theory-Based Modeling, Exploration, and Planning
• Leveraging Language to Learn Program Abstractions and Search Heuristics
• Ask Your Humans: Using Human Instructions to Improve Generalization in\n Reinforcement Learning
• ALFWorld: Aligning Text and Embodied Environments for Interactive Learning
• DreamCoder: Growing generalizable, interpretable knowledge with wake-sleep Bayesian program learning
• Language Models are Few-Shot Learners
• Language as an Abstraction for Hierarchical Deep Reinforcement Learning
• A Survey of Reinforcement Learning Informed by Natural Language
• Hierarchical Decision Making by Generating and Following Natural Language Instructions
• Using Natural Language for Reward Shaping in Reinforcement Learning
• Deep Reinforcement Learning with Double Q-learning
• Sequential monte carlo samplers
• Planning and acting in partially observable stochastic domains
• Understanding Epistemic Language with a Language-augmented Bayesian Theory of Mind
• Reinforcement Learning: An Introduction

Related

• AI Pedagogy: Dialogic Social Learning for Artificial Agents
• Learning to solve complex tasks by growing knowledge culturally across generations
• Social Learning through Interactions with Other Agents: A Survey
• Language Game: Talking to Non-Human Systems
• Computational Language Acquisition with Theory of Mind