While Reinforcement Learning from Human Feedback (RLHF) has driven major advances—especially in large language models—its reliance on human-provided signals limits scalability, objectivity, and applicability. This workshop aims to push beyond human-centric feedback toward world feedback: signals arising directly from interactions with physical, digital, economic, or social environments. These include measurable outcomes such as efficiency, safety, latency, resource usage, system performance, and long-term impact. We invite submissions that explore how learning systems can leverage such feedback to achieve scalable, robust, and real-world-aligned intelligence.
We welcome submissions on topics including, but not limited to:
Learning from non-human feedback: RL methods leveraging environment-derived signals (e.g., energy, latency, safety, economic outcomes, system metrics)
Integrating heterogeneous feedback: Combining world feedback with human, synthetic, or self-generated signals under noise, delay, or partial observability
World feedback for foundation models: Training and aligning LLMs and multimodal models using interaction-based or outcome-driven signals
Modeling and generating feedback: Simulators, digital twins, generative world models, and diffusion-based approaches for inferring or amplifying feedback
Benchmarks and evaluation: Metrics, datasets, and real-world deployments in robotics, systems optimization, healthcare, economics, and multi-agent systems
We will use OpenReview to manage the submissions and the double-blind review process.
Format
Full-paper submission (up to 9 pages) in ICML or NeurIPS format with potentially large-scale experiments
Short submission (2-4 pages) in ICML or NeurIPS format with proof-of-concept demonstrations of the idea proposed. The proof-of-concept submission can include demos, code, and a blog post. This format is to account for underrepresented researchers who may not have access to large amounts of compute
Note
Page limits exclude references and appendices.
The workshop proceedings are non-archival, and we welcome submissions that are currently under review at other venues (e.g., NeurIPS 2026).
Key dates
Submission deadline: May 13, 2026, AoE
Author notification: May 31, 2026, AoE
Camera-ready deadline: June 30, 2026, AoE
Workshop date: July 10, 2026