On the (complete) reasons behind decisions
A Darwiche, A Hirth - Journal of Logic, Language and Information, 2023 - Springer
Recent work has shown that the input-output behavior of some common machine learning
classifiers can be captured in symbolic form, allowing one to reason about the behavior of …
classifiers can be captured in symbolic form, allowing one to reason about the behavior of …
XAudit: A Theoretical Look at Auditing with Explanations
Responsible use of machine learning requires models to be audited for undesirable
properties. While a body of work has proposed using explanations for auditing, how to do so …
properties. While a body of work has proposed using explanations for auditing, how to do so …
On the complexity of enumerating prime implicants from decision-DNNF circuits
A de Colnet, P Marquis - arXiv preprint arXiv:2301.13328, 2023 - arxiv.org
We consider the problem EnumIP of enumerating prime implicants of Boolean functions
represented by decision decomposable negation normal form (dec-DNNF) circuits. We study …
represented by decision decomposable negation normal form (dec-DNNF) circuits. We study …
Towards Interpretable Machine-Learning-Based DDoS Detection
Q Zhou, R Li, L Xu, A Nallanathan, J Yang, A Fu - SN Computer Science, 2023 - Springer
The Internet is the most complex machine humankind has ever built, and thus, it is difficult to
defend it from attacks. The most common attack to the Internet is DDoS attacks. With the …
defend it from attacks. The most common attack to the Internet is DDoS attacks. With the …
Towards explainable meta-learning for DDoS detection
Q Zhou, R Li, L Xu, A Nallanathan, J Yang… - arXiv preprint arXiv …, 2022 - arxiv.org
The Internet is the most complex machine humankind has ever built, and how to defense it
from intrusions is even more complex. With the ever increasing of new intrusions, intrusion …
from intrusions is even more complex. With the ever increasing of new intrusions, intrusion …
Rin: Towards a semantic rigorous interpretable artificial immune system for intrusion detection
The Internet is the most complex machine humankind has ever built, and how to defense it
from intrusions is even more complex. With the ever increasing of new intrusions, intrusion …
from intrusions is even more complex. With the ever increasing of new intrusions, intrusion …
XAudit: A Learning-Theoretic Look at Auditing with Explanations
Responsible use of machine learning requires models to be audited for undesirable
properties. While a body of work has proposed using explanations for auditing, how to do so …
properties. While a body of work has proposed using explanations for auditing, how to do so …
Deriving Explanations for Decision Trees: The Impact of Domain Theories
We are interested in identifying the complexity of computing explanations of various types for
a decision tree, when the Boolean conditions used in the tree are not independent. When a …
a decision tree, when the Boolean conditions used in the tree are not independent. When a …
[PDF][PDF] Hard Functions in Knowledge Compilation: from Lower Bounds to Applications
A de Colnet - 2022 - theses.fr
I would like to thank the reviewers of this thesis: Adnan Darwiche, the father of DNNF, whose
many contributions to knowledge compilation are fundamental for many aspects of my …
many contributions to knowledge compilation are fundamental for many aspects of my …
Towards Tractable Formal Explanations for Neural Network Models
Q Zhou - … Conference on Addressing Socioethical Effects of … - openreview.net
Neural networks are fragile, it is widely reported that trivial disturbances on the inputs can
make a neural network model flip its decision, and as a black box model, it cannot provide …
make a neural network model flip its decision, and as a black box model, it cannot provide …