-
The CAISAR Platform: Extending the Reach of Machine Learning Specification and Verification
Authors:
Michele Alberti,
François Bobot,
Julien Girard-Satabin,
Alban Grastien,
Aymeric Varasse,
Zakaria Chihani
Abstract:
The formal specification and verification of machine learning programs saw remarkable progress in less than a decade, leading to a profusion of tools. However, diversity may lead to fragmentation, resulting in tools that are difficult to compare, except for very specific benchmarks. Furthermore, this progress is heavily geared towards the specification and verification of a certain class of proper…
▽ More
The formal specification and verification of machine learning programs saw remarkable progress in less than a decade, leading to a profusion of tools. However, diversity may lead to fragmentation, resulting in tools that are difficult to compare, except for very specific benchmarks. Furthermore, this progress is heavily geared towards the specification and verification of a certain class of property, that is, local robustness properties. But while provers are becoming more and more efficient at solving local robustness properties, even slightly more complex properties, involving multiple neural networks for example, cannot be expressed in the input languages of winners of the International Competition of Verification of Neural Networks VNN-Comp. In this tool paper, we present CAISAR, an open-source platform dedicated to machine learning specification and verification. We present its specification language, suitable for modelling complex properties on neural networks, support vector machines and boosted trees. We show on concrete use-cases how specifications written in this language are automatically translated to queries to state-of-the-art provers, notably by using automated graph editing techniques, making it possible to use their off-the-shelf versions. The artifact to reproduce the paper claims is available at the following DOI: https://doi.org/10.5281/zenodo.15209510
△ Less
Submitted 10 June, 2025;
originally announced June 2025.
-
CaBRNet, an open-source library for developing and evaluating Case-Based Reasoning Models
Authors:
Romain Xu-Darme,
Aymeric Varasse,
Alban Grastien,
Julien Girard,
Zakaria Chihani
Abstract:
In the field of explainable AI, a vibrant effort is dedicated to the design of self-explainable models, as a more principled alternative to post-hoc methods that attempt to explain the decisions after a model opaquely makes them. However, this productive line of research suffers from common downsides: lack of reproducibility, unfeasible comparison, diverging standards. In this paper, we propose Ca…
▽ More
In the field of explainable AI, a vibrant effort is dedicated to the design of self-explainable models, as a more principled alternative to post-hoc methods that attempt to explain the decisions after a model opaquely makes them. However, this productive line of research suffers from common downsides: lack of reproducibility, unfeasible comparison, diverging standards. In this paper, we propose CaBRNet, an open-source, modular, backward-compatible framework for Case-Based Reasoning Networks: https://github.com/aiser-team/cabrnet.
△ Less
Submitted 25 September, 2024;
originally announced September 2024.
-
A More General Theory of Diagnosis from First Principles
Authors:
Alban Grastien,
Patrik Haslum,
Sylvie Thiébaux
Abstract:
Model-based diagnosis has been an active research topic in different communities including artificial intelligence, formal methods, and control. This has led to a set of disparate approaches addressing different classes of systems and seeking different forms of diagnoses. In this paper, we resolve such disparities by generalising Reiter's theory to be agnostic to the types of systems and diagnoses…
▽ More
Model-based diagnosis has been an active research topic in different communities including artificial intelligence, formal methods, and control. This has led to a set of disparate approaches addressing different classes of systems and seeking different forms of diagnoses. In this paper, we resolve such disparities by generalising Reiter's theory to be agnostic to the types of systems and diagnoses considered. This more general theory of diagnosis from first principles defines the minimal diagnosis as the set of preferred diagnosis candidates in a search space of hypotheses. Computing the minimal diagnosis is achieved by exploring the space of diagnosis hypotheses, testing sets of hypotheses for consistency with the system's model and the observation, and generating conflicts that rule out successors and other portions of the search space. Under relatively mild assumptions, our algorithms correctly compute the set of preferred diagnosis candidates. The main difficulty here is that the search space is no longer a powerset as in Reiter's theory, and that, as consequence, many of the implicit properties (such as finiteness of the search space) no longer hold. The notion of conflict also needs to be generalised and we present such a more general notion. We present two implementations of these algorithms, using test solvers based on satisfiability and heuristic search, respectively, which we evaluate on instances from two real world discrete event problems. Despite the greater generality of our theory, these implementations surpass the special purpose algorithms designed for discrete event systems, and enable solving instances that were out of reach of existing diagnosis approaches.
△ Less
Submitted 28 September, 2023;
originally announced September 2023.
-
Interval Predictability in Discrete Event Systems
Authors:
Alban Grastien
Abstract:
In this paper we study the problem of predictability in partially observable discrete event systems, i.e., the question whether an observer can predict the occurrence of a fault. We extend the definition of predictability to consider the time interval where the fault will occur: the $(i,j)$-predictability does not only specify that the fault will be predicted before it occurs, but also that the pr…
▽ More
In this paper we study the problem of predictability in partially observable discrete event systems, i.e., the question whether an observer can predict the occurrence of a fault. We extend the definition of predictability to consider the time interval where the fault will occur: the $(i,j)$-predictability does not only specify that the fault will be predicted before it occurs, but also that the predictor will be able to predict that its occurrence will occur in $i$ to $j$ observations from now. We also provide a quadratic algorithm that decides predictability of the system.
△ Less
Submitted 4 August, 2015;
originally announced August 2015.
-
The Complexity of Switching and FACTS Maximum-Potential-Flow Problems
Authors:
Karsten Lehmann,
Alban Grastien,
Pascal Van Hentenryck
Abstract:
This papers considers the problem of maximizing the load that can be served by a power network. We use the commonly accepted Linear DC power network model and consider wo configuration options: switching lines and using FACTS devices. We present the first comprehensive complexity study of this optimization problem. Our results show hat the problem is NP-complete and that there is no fully polynomi…
▽ More
This papers considers the problem of maximizing the load that can be served by a power network. We use the commonly accepted Linear DC power network model and consider wo configuration options: switching lines and using FACTS devices. We present the first comprehensive complexity study of this optimization problem. Our results show hat the problem is NP-complete and that there is no fully polynomial-time approximation scheme. For switching, these results extend to planar networks with a aximum-node degree of 3. Additionally, we demonstrate that the optimization problems are still NP-hard if we restrict the network structure to cacti with a maximum degree of 3.
△ Less
Submitted 16 July, 2015;
originally announced July 2015.
-
The Complexity of DC-Switching Problems
Authors:
Karsten Lehmann,
Alban Grastien,
Pascal Van Hentenryck
Abstract:
This report provides a comprehensive complexity study of line switching in the Linear DC model for the feasibility problem and the optimization problems of maximizing the load that can be served (maximum switching flow, MSF) and minimizing generation cost (optimal transmission switching, OTS). Our results show that these problems are NP-complete and that there is no fully polynomial-time approxima…
▽ More
This report provides a comprehensive complexity study of line switching in the Linear DC model for the feasibility problem and the optimization problems of maximizing the load that can be served (maximum switching flow, MSF) and minimizing generation cost (optimal transmission switching, OTS). Our results show that these problems are NP-complete and that there is no fully polynomial-time approximation scheme for planar networks with a maximum-node degree of 3. Additionally, we demonstrate that the problems are still NP-hard if we restrict the network structure to cacti with a maximum degree of 3. We also show that the optimization problems can not be approximated within any constant factor.
△ Less
Submitted 17 November, 2014;
originally announced November 2014.
-
AC-Feasibility on Tree Networks is NP-Hard
Authors:
Karsten Lehmann,
Alban Grastien,
Pascal Van Hentenryck
Abstract:
Recent years have witnessed significant interest in convex relaxations of the power flows, several papers showing that the second-order cone relaxation is tight for tree networks under various conditions on loads or voltages. This paper shows that AC-feasibility, i.e., to find whether some generator dispatch can satisfy a given demand, is NP-Hard for tree networks.
Recent years have witnessed significant interest in convex relaxations of the power flows, several papers showing that the second-order cone relaxation is tight for tree networks under various conditions on loads or voltages. This paper shows that AC-feasibility, i.e., to find whether some generator dispatch can satisfy a given demand, is NP-Hard for tree networks.
△ Less
Submitted 30 October, 2014;
originally announced October 2014.
-
Predictability of Event Occurrences in Timed Systems
Authors:
Franck Cassez,
Alban Grastien
Abstract:
We address the problem of predicting events' occurrences in partially observable timed systems modelled by timed automata. Our contribution is many-fold: 1) we give a definition of bounded predictability, namely k-predictability, that takes into account the minimum delay between the prediction and the actual event's occurrence; 2) we show that 0-predictability is equivalent to the original notion…
▽ More
We address the problem of predicting events' occurrences in partially observable timed systems modelled by timed automata. Our contribution is many-fold: 1) we give a definition of bounded predictability, namely k-predictability, that takes into account the minimum delay between the prediction and the actual event's occurrence; 2) we show that 0-predictability is equivalent to the original notion of predictability of S. Genc and S. Lafortune; 3) we provide a necessary and sufficient condition for k-predictability (which is very similar to k-diagnosability) and give a simple algorithm to check k-predictability; 4) we address the problem of predictability of events' occurrences in timed automata and show that the problem is PSPACE-complete.
△ Less
Submitted 4 June, 2013;
originally announced June 2013.
-
An example illustrating the imprecision of the efficient approach for diagnosis of Petri nets via integer linear programming
Authors:
Alban Grastien
Abstract:
This document demonstrates that the efficient approach for diagnosis of Petri nets via integer linear programming may be unable to detect a fault even if the system is diagnosable.
This document demonstrates that the efficient approach for diagnosis of Petri nets via integer linear programming may be unable to detect a fault even if the system is diagnosable.
△ Less
Submitted 15 October, 2012;
originally announced October 2012.
