Discovering causal graphs with cycles and latent confounders: An exact branch-and-bound approach

Rantanen, Kari; Hyttinen, Antti; Järvisalo, Matti

doi:10.1016/j.ijar.2019.10.009

Cited by 13 publications

(8 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These approaches include many algorithms designed for the linear case, including LLC [78], score-based approaches [58], and BackShift [133]. Algorithms for the general case include SAT-based approaches [81], exact score-based approaches [129], and constraintbased approaches [48,113,154].…”

Section: Discussion and Open Problemsmentioning

confidence: 99%

Causal Structure Learning: a Combinatorial Perspective

Squires¹,

Uhler²

2022

Preprint

View full text Add to dashboard Cite

In this review, we discuss approaches for learning causal structure from data, also called causal discovery. In particular, we focus on approaches for learning directed acyclic graphs (DAGs) and various generalizations which allow for some variables to be unobserved in the available data. We devote special attention to two fundamental combinatorial aspects of causal structure learning. First, we discuss the structure of the search space over causal graphs. Second, we discuss the structure of equivalence classes over causal graphs, i.e., sets of graphs which represent what can be learned from observational data alone, and how these equivalence classes can be refined by adding interventional data.

show abstract

Section: Discussion and Open Problemsmentioning

confidence: 99%

Causal Structure Learning: a Combinatorial Perspective

Squires¹,

Uhler²

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…En las reglas de la fase 2 y fase 3 se especifican valores y porcentajes que hemos propuesto para que el método evite la evaluación exhaustiva de los nodos y realice la ponderación con una precisión adecuada. Aunque esta es una propuesta de los autores, se ha tomado en cuenta la relación causal entre los diferentes niveles de conocimiento en un AD con su grado de complejidad respecto a los conocimientos previos y posteriores (Nie, Shi & Li, 2020;Rantanen, Hyttinen & Järvisalo, 2020;Shin & Jeong, 2021). Figura 2.…”

Section: Salidaunclassified

“…Por lo tanto, el conocimiento pueden representarse mediante colecciones interrelacionadas de temas en un ámbito determinado. En este sentido, el GC establece relaciones causales entre los temas previos y posteriores de un AD (Rantanen, Hyttinen & Järvisalo, 2020;Shin & Jeong, 2021). Así, la no evaluación exhaustiva se justifica al evitar la evaluación de todos los temas de un AD; es decir, al suponer tres temas interrelacionados por su complejidad A>B>C, si al evaluar el tema B se determina 100% de dominio, entonces se asume que el tema A es dominado totalmente, lo que determinaría también parte del conocimiento que se tiene del tema C.…”

Section: Introductionunclassified

Knowledge Graph to determine the domain of learning in Higher Education

Almazán

Vega

Amaya

et al. 2021

View full text Add to dashboard Cite

The representing of a student’s knowledge in an academic discipline plays an important role in boosting the student’s skills. To support stakeholders in the educational domain, it is necessary to provide them with robust assessment strategies that facilitate the teaching-learning process. Student´s mastery is determined by the degree of knowledge, which demonstrates objectively, on the topics included in the different areas that make up an academic discipline. Although there is a wide variety of techniques to represent knowledge, particularly Knowledge Graph technique is becoming relevant due to the structured approach and benefits it offers. This paper proposes a method that classifies and weights the nodes (topics) of a Knowledge Graph of a disciplinary area, which is analyzed through a case study. The method has two approaches: avoid exhaustive evaluation of the nodes and weight the nodes with adequate precision. Method´s application is illustrated by a case study. As results, a Knowledge Graph is obtained with its classified and weighted nodes through the application of the proposed method, in which 100% of the topics have been impacted through the objective evaluation of 20.8% representing 10 nodes. It is concluded that the proposed method has potential to be used in the representation and management of knowledge, being necessary to improve phases’ iteration to condition number of objective nodes.

show abstract

“…The existing works on cyclic causal models primarily focus independently and identically distributed (i.i.d) data instances (Richardson, 1996(Richardson, , 1997Strobl, 2019;Rantanen et al, 2020). However, in many real-world systems units are often interconnected in a complex network.…”

Section: Introductionmentioning

confidence: 99%

Relational Causal Models with Cycles:Representation and Reasoning

Ahsan¹,

Arbour²,

Zheleva³

2022

Preprint

View full text Add to dashboard Cite

Causal reasoning in relational domains is fundamental to studying real-world social phenomena in which individual units can influence each other's traits and behavior. Dynamics between interconnected units can be represented as an instantiation of a relational causal model; however, causal reasoning over such instantiation requires additional templating assumptions that capture feedback loops of influence. Previous research has developed lifted representations to address the relational nature of such dynamics but has strictly required that the representation has no cycles. To facilitate cycles in relational representation and learning, we introduce relational σ-separation, a new criterion for understanding relational systems with feedback loops. We also introduce a new lifted representation, σ-abstract ground graph which helps with abstracting statistical independence relations in all possible instantiations of the cyclic relational model. We show the necessary and sufficient conditions for the completeness of σ-AGG and that relational σ-separation is sound and complete in the presence of one or more cycles with arbitrary length. To the best of our knowledge, this is the first work on representation of and reasoning with cyclic relational causal models.

show abstract

Discovering causal graphs with cycles and latent confounders: An exact branch-and-bound approach

Cited by 13 publications

References 22 publications

Causal Structure Learning: a Combinatorial Perspective

Causal Structure Learning: a Combinatorial Perspective

Knowledge Graph to determine the domain of learning in Higher Education

Relational Causal Models with Cycles:Representation and Reasoning

Contact Info

Product

Resources

About