VisualPatchWorld: Code World Models as Latent Structured Representations for Planning
Categories
15 pages
Publication
PatchWorld: Learning Executable World Models Without Gradients
Automating the Grounding Layer: A Survey of Operational Knowledge for Enterprise Agents
The Agent Harness: Infrastructure That Makes LLM Agents Actually Work
[ICLR'26] Taming the Wild Hypothesis: Introducing Controllable Abductive Reasoning in Knowledge Graphs
![[ICLR'26] Taming the Wild Hypothesis: Introducing Controllable Abductive Reasoning in Knowledge Graphs](/p/iclr26-taming-the-wild-hypothesis-introducing-controllable-abductive-reasoning-in-knowledge-graphs/ctrlhgen_hu16494910985212281503.png)
[WWW'26] Unifying Logic Reasoning in Knowledge Graphs with Diffusion Models
![[WWW'26] Unifying Logic Reasoning in Knowledge Graphs with Diffusion Models](/p/www26-unifying-logic-reasoning-in-knowledge-graphs-with-diffusion-models/dark_hu13192193941403510688.png)
Token Efficient RAG
[ICLR'26] AtlasKV: Teaching a Language Model a Billion Facts, Without Breaking the Memory Bank
![[ICLR'26] AtlasKV: Teaching a Language Model a Billion Facts, Without Breaking the Memory Bank](/p/iclr26-atlaskv-teaching-a-language-model-a-billion-facts-without-breaking-the-memory-bank/framework_hu15432790057858931294.png)