Goto

Collaborating Authors

Ribeiro

AAAI Conferences

We introduce a novel model-agnostic system that explains the behavior of complex models with high-precision rules called anchors, representing local, "sufficient" conditions for predictions. We propose an algorithm to efficiently compute these explanations for any black-box model with high-probability guarantees. We demonstrate the flexibility of anchors by explaining a myriad of different models for different domains and tasks. In a user study, we show that anchors enable users to predict how a model would behave on unseen instances with less effort and higher precision, as compared to existing linear explanations or no explanations.


Semantic Explanations of Predictions

arXiv.org Artificial Intelligence

The main objective of explanations is to transmit knowledge to humans. This work proposes to construct informative explanations for predictions made from machine learning models. Motivated by the observations from social sciences, our approach selects data points from the training sample that exhibit special characteristics crucial for explanation, for instance, ones contrastive to the classification prediction and ones representative of the models. Subsequently, semantic concepts are derived from the selected data points through the use of domain ontologies. These concepts are filtered and ranked to produce informative explanations that improves human understanding. The main features of our approach are that (1) knowledge about explanations is captured in the form of ontological concepts, (2) explanations include contrastive evidences in addition to normal evidences, and (3) explanations are user relevant.


Du

AAAI Conferences

ABox abduction is an important reasoning mechanism for description logic ontologies. It computes all minimal explanations (sets of ABox assertions) whose appending to a consistent ontology enforces the entailment of an observation while keeps the ontology consistent. We focus on practical computation for a general problem of ABox abduction, called the query abduction problem, where an observation is a Boolean conjunctive query and the explanations may contain fresh individuals neither in the ontology nor in the observation. However, in this problem there can be infinitely many minimal explanations. Hence we first identify a class of TBoxes called first-order rewritable TBoxes. It guarantees the existence of finitely many minimal explanations and is sufficient for many ontology applications. To reduce the number of explanations that need to be computed, we introduce a special kind of minimal explanations called representative explanations from which all minimal explanations can be retrieved. We develop a tractable method (in data complexity) for computing all representative explanations in a consistent ontology.


Fan

AAAI Conferences

Argumentation can be viewed as a process of generating explanations. However, existing argumentation semantics are developed for identifying acceptable arguments within a set, rather than giving concrete justifications for them. In this work, we propose a new argumentation semantics, related admissibility, designed for giving explanations to arguments in both Abstract Argumentation and Assumption-based Argumentation. We identify different types of explanations defined in terms of the new semantics. We also give a correct computational counterpart for explanations using dispute forests.


Anchors: High-Precision Model-Agnostic Explanations

AAAI Conferences

We introduce a novel model-agnostic system that explains the behavior of complex models with high-precision rules called anchors, representing local, "sufficient" conditions for predictions. We propose an algorithm to efficiently compute these explanations for any black-box model with high-probability guarantees. We demonstrate the flexibility of anchors by explaining a myriad of different models for different domains and tasks. In a user study, we show that anchors enable users to predict how a model would behave on unseen instances with less effort and higher precision, as compared to existing linear explanations or no explanations.