Compiling Bayesian Network Classifiers into Decision Graphs

Shih, Andy; Choi, Arthur; Darwiche, Adnan

doi:10.1609/aaai.v33i01.33017966

Cited by 57 publications

(84 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Past work on computing explanations has mostly addressed local (or instance-dependent) explanations [15,16,38,51,69,70,75,76]. Exceptions include for example approaches that distill ML models, e.g.…”

Section: Introductionmentioning

confidence: 99%

“…In contrast, local (or global) model-based explanations are referred to as rigorous, since these offer the strongest formal guarantees with respect to the underlying ML model. Concrete examples of such rigorous approaches include [15,16,35,[38][39][40][41]43,52,64,[75][76][77].…”

Section: Introductionmentioning

confidence: 99%

“…For rigorous approaches, the answer to a 'Why prediction π?' question has been referred to as PI-explanations [75,76], abductive explanations [38], but also as (minimal) sufficient reasons [15,16]. (Hereinafter, we use the term abductive explanation because of the other forms of explanations studied in the paper.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

From Contrastive to Abductive Explanations and Back Again

Ignatiev

Narodytska

Asher

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Explanations of Machine Learning (ML) models often address a 'Why?' question. Such explanations can be related with selecting feature-value pairs which are sufficient for the prediction. Recent work has investigated explanations that address a 'Why Not?' question, i.e. finding a change of feature values that guarantee a change of prediction. Given their goals, these two forms of explaining predictions of ML models appear to be mostly unrelated. However, this paper demonstrates otherwise, and establishes a rigorous formal relationship between 'Why?' and 'Why Not?' explanations. Concretely, the paper proves that, for any given instance, 'Why?' explanations are minimal hitting sets of 'Why Not?' explanations and vice-versa. Furthermore, the paper devises novel algorithms for extracting and enumerating both forms of explanations.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

From Contrastive to Abductive Explanations and Back Again

Ignatiev

Narodytska

Asher

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Hence, the proposal in [9] amounted to compiling a Naive Bayes classifier into a tractable NNF circuit that precisely captures its input-output behavior. This compilation algorithm was recently extended to Bayesian network classifiers with tree structures [82] and later to Bayesian network classifiers with arbitrary structures [83]. 17 Certain classes of neural networks can also be compiled into tractable circuits, which include SDD circuits as shown in [15,80,84].…”

Section: Logic For Meta Reasoningmentioning

confidence: 99%

Three Modern Roles for Logic in AI

Darwiche

2020

Proceedings of the 39th ACM SIGMOD-SIGACT-SIGAI Symposium on Principles of Database Systems

Self Cite

View full text Add to dashboard Cite

We consider three modern roles for logic in artificial intelligence, which are based on the theory of tractable Boolean circuits: (1) logic as a basis for computation, (2) logic for learning from a combination of data and knowledge, and (3) logic for reasoning about the behavior of machine learning systems. CCS CONCEPTS • Computing methodologies → Learning in probabilistic graphical models; Logical and relational learning; • Theory of computation → Automated reasoning; Complexity classes; Problems, reductions and completeness.

show abstract

“…A more descriptive term for the underlying probability model would be the "independent feature model" [21]. Bayesian has shown as an accurate model for various problem [22] [23]. Naïve Bayes was used by [24] in mapping out the potentially poor family in Indonesia to planning the right method in preventing such occurrence towards the family.…”

Section: Related Workmentioning

confidence: 99%

Household Overspending Model Amongst B40, M40 and T20 using Classification Algorithm

Othman¹,

Bakar²,

Sani³

et al. 2020

IJACSA

View full text Add to dashboard Cite

The family economy is a critical indicator of the well-being of a family institution. It can be seen by the total income and how well the household finances is managed. In Malaysia, the household income level is categorized as B40, M40 and T20. These categories can also indicate the poverty level of the household. Overspending is a phenomenon where the monthly expenses are more than the household's total income, which affects economic wellbeing. Finding important factors that affect the spending patterns among the household can reveal the causes of overspending. It will assist the government in mitigating such problems. Availability of 4 million household expenditure records obtained from the survey conducted in 2016 by the Department of Statistics Malaysia eases the aim of this study to develop a household overspending model by using machine learning. The model is developed using 12 household demographic attributes with 14451 household records. The attributes are the number of households, area, state, strata, race, highest certificate, marital status, gender, housing, income, total expenditure, and category as attributes class. The model development employs five machine learning algorithms namely decision tree, Naïve Bayes, Neural network, Support Vector Machines, Nearest Neighbour. The results show that the decision tree through J48 algorithm has produced the easiest rule to be interpreted. The model shows four attributes which were income, state, races and number of households that highly influence the overspending problem. Based on the research finding, it can be concluded that these attributes are essential for improving the indicator measure for Malaysian Family Wellbeing Index in the aspect of overspending.

show abstract

Compiling Bayesian Network Classifiers into Decision Graphs

Cited by 57 publications

References 9 publications

From Contrastive to Abductive Explanations and Back Again

From Contrastive to Abductive Explanations and Back Again

Three Modern Roles for Logic in AI

Household Overspending Model Amongst B40, M40 and T20 using Classification Algorithm

Contact Info

Product

Resources

About