Episodic Context and Persistent 3D World Models Enable Curiosity-Driven Exploration in Photorealistic Environments
This paper addresses the failure modes of curiosity-driven RL in complex 3D environments, where agents revisit forgotten states and get trapped in local loops due to lacking spatial persistence and episodic memory. The authors combine an online 3D reconstruction as a persistent world model with a sequence-model policy over RGB observations to maintain episodic trajectory context. Trained purely via intrinsic curiosity on HM3D, the agent outperforms RL-based active mapping baselines and zero-shot generalizes to Gibson and AI-generated environments. The approach also enables efficient downstream task adaptation for apple picking and image-goal navigation.
Related guides (2)
Related events (8)
Large-scale Study of Curiosity-Driven Learning
OpenAI published research on curiosity-driven learning, exploring intrinsic motivation as a reward signal for reinforcement learning agents at scale. The study investigates how curiosity-based exploration can enable agents to learn useful behaviors without extrinsic rewards. This represents an early foundational contribution to reward-free and self-supervised RL research.
Mem-π: Adaptive Memory for LLM Agents via On-Demand Generation and Decoupled RL
Mem-π introduces a framework where a dedicated language or vision-language model generates context-specific guidance for LLM agents on demand, rather than retrieving static entries from episodic memory banks. The system is trained with a decision-content decoupled reinforcement learning objective that jointly learns when to generate guidance and what to generate, enabling abstention when generation would not help. Evaluated across web navigation, terminal-based tool use, and text-based embodied interaction benchmarks, Mem-π achieves over 30% relative improvement on web navigation tasks compared to retrieval-based and prior RL-optimized memory baselines.
LongTraceRL: Reinforcement Learning for Long-Context Reasoning via Search Agent Trajectories and Rubric Rewards
LongTraceRL is a new RL training framework for improving long-context reasoning in LLMs, addressing limitations of existing RLVR methods. It constructs challenging training data using multi-hop questions from knowledge graph random walks and tiered distractors derived from search agent trajectories (high-confusability: read but uncited; low-confusability: seen but unopened). A rubric reward provides entity-level process supervision along reasoning chains, applied only to correct responses to prevent reward hacking. Experiments across three LLMs (4B–30B parameters) on five long-context benchmarks show consistent improvements over strong baselines.
ContextRL: Context-aware reinforcement learning improves grounding in agentic and multimodal LLMs
Researchers introduce ContextRL, a reinforcement learning method that trains LLMs to select the context that supports a given query-answer pair from two highly similar candidates, rather than supervising only final answers. The approach constructs contrastive context pairs in two domains: coding agent trajectories (1k pairs) and multimodal image pairs (7k pairs). ContextRL achieves +2.2% average gains over standard GRPO on 5 long-horizon benchmarks and +1.8% across 12 visual QA benchmarks, with ablations showing the gains stem from the context-selection objective rather than the contrastive data alone.
Some considerations on learning to explore via meta-reinforcement learning
OpenAI published a research post examining exploration strategies learned through meta-reinforcement learning. The work investigates how agents can acquire exploration behaviors through meta-learning rather than having them hand-designed. This is an early OpenAI contribution to the intersection of meta-learning and RL, predating the current frontier model era.
GRASP: Gradient-based Planning for World Models at Longer Horizons
Researchers from Berkeley, Meta, and collaborators introduce GRASP, a gradient-based planner designed to make long-horizon planning with learned world models more robust. The method addresses three core failure modes: ill-conditioned computation graphs from backpropagation through time, non-greedy loss landscapes with many local minima, and brittle gradients through high-dimensional vision models. GRASP lifts trajectory optimization into virtual states for parallel optimization across time, injects stochasticity into state iterates for exploration, and reshapes gradients to avoid problematic state-input gradient paths. The work is positioned in the context of scaling world models toward general-purpose simulators usable for control and planning.
RePro: Retrospective Progress-Aware Self-Refinement for LLM Agent Training
Researchers introduce RePro (Retrospective Progress-Aware Training), a framework addressing the gap between step-wise RL optimization and metacognitive task-progress awareness in LLM agents. The approach uses a forward-then-reflect rollout paradigm where agents execute actions online and then retrospectively assess step-wise progress given the completed trajectory and known outcome. Evaluated on WebShop, ALFWorld, and Sokoban, RePro achieves up to 12% absolute success rate gains over baseline Qwen-family models without requiring continuous external supervision.
Genie 3: A new frontier for world models
DeepMind has announced Genie 3, a world model capable of generating interactive, navigable 3D environments in real time at 24 fps and 720p resolution. The system maintains consistency for several minutes, representing a significant step up from prior Genie iterations. This positions Genie 3 as a frontier capability demonstration in generative world modeling for interactive applications.