directed acyclic graphs
Directed acyclic graphs (DAGs) are graphical structures used in AI to represent relationships between variables, where edges indicate direction and cycles are not allowed. DAGs are foundational in probabilistic graphical models, enabling efficient representation of complex dependencies.
- Aligning Compound AI Systems via System-level DPO
- Cyclic Counterfactuals under Shift–Scale Interventions
- DSCS: Fast CPDAG-Based Verification of Collapsible Submodels in High-Dimensional Bayesian Networks
- Decoding Causal Structure: End-to-End Mediation Pathways Inference
- Heterogeneous Swarms: Jointly Optimizing Model Roles and Weights for Multi-LLM Systems
- Median Selection with Noisy and Structural Information
- On the Universal Near Optimality of Hedge in Combinatorial Settings
- pLSTM: parallelizable Linear Source Transition Mark networks