-
Graph polynomials: some questions on the edge
Authors:
Graham Farr,
Kerri Morgan
Abstract:
We raise some questions about graph polynomials, highlighting concepts and phenomena that may merit consideration in the development of a general theory. Our questions are mainly of three types: When do graph polynomials have reduction relations (simple linear recursions based on local operations), perhaps in a wider class of combinatorial objects? How many levels of reduction relations does a gra…
▽ More
We raise some questions about graph polynomials, highlighting concepts and phenomena that may merit consideration in the development of a general theory. Our questions are mainly of three types: When do graph polynomials have reduction relations (simple linear recursions based on local operations), perhaps in a wider class of combinatorial objects? How many levels of reduction relations does a graph polynomial need in order to express it in terms of trivial base cases? For a graph polynomial, how are properties such as equivalence and factorisation reflected in the structure of a graph? We illustrate our discussion with a variety of graph polynomials and other invariants. This leads us to reflect on the historical origins of graph polynomials. We also introduce some new polynomials based on partial colourings of graphs and establish some of their basic properties.
△ Less
Submitted 22 June, 2024;
originally announced June 2024.
-
Anytime Approximate Formal Feature Attribution
Authors:
**qiang Yu,
Graham Farr,
Alexey Ignatiev,
Peter J. Stuckey
Abstract:
Widespread use of artificial intelligence (AI) algorithms and machine learning (ML) models on the one hand and a number of crucial issues pertaining to them warrant the need for explainable artificial intelligence (XAI). A key explainability question is: given this decision was made, what are the input features which contributed to the decision? Although a range of XAI approaches exist to tackle t…
▽ More
Widespread use of artificial intelligence (AI) algorithms and machine learning (ML) models on the one hand and a number of crucial issues pertaining to them warrant the need for explainable artificial intelligence (XAI). A key explainability question is: given this decision was made, what are the input features which contributed to the decision? Although a range of XAI approaches exist to tackle this problem, most of them have significant limitations. Heuristic XAI approaches suffer from the lack of quality guarantees, and often try to approximate Shapley values, which is not the same as explaining which features contribute to a decision. A recent alternative is so-called formal feature attribution (FFA), which defines feature importance as the fraction of formal abductive explanations (AXp's) containing the given feature. This measures feature importance from the view of formally reasoning about the model's behavior. It is challenging to compute FFA using its definition because that involves counting AXp's, although one can approximate it. Based on these results, this paper makes several contributions. First, it gives compelling evidence that computing FFA is intractable, even if the set of contrastive formal explanations (CXp's) is provided, by proving that the problem is #P-hard. Second, by using the duality between AXp's and CXp's, it proposes an efficient heuristic to switch from CXp enumeration to AXp enumeration on-the-fly resulting in an adaptive explanation enumeration algorithm effectively approximating FFA in an anytime fashion. Finally, experimental results obtained on a range of widely used datasets demonstrate the effectiveness of the proposed FFA approximation approach in terms of the error of FFA approximation as well as the number of explanations computed and their diversity given a fixed time limit.
△ Less
Submitted 11 December, 2023;
originally announced December 2023.
-
A survey of repositories in graph theory
Authors:
Srinibas Swain,
C. Paul Bonnington,
Graham Farr,
Kerri Morgan
Abstract:
Since the pioneering work of R. M. Foster in the 1930s, many graph repositories have been created to support research in graph theory. This survey reviews many of these graph repositories and summarises the scope and contents of each repository. We identify opportunities for the development of repositories that can be queried in more flexible ways.
Since the pioneering work of R. M. Foster in the 1930s, many graph repositories have been created to support research in graph theory. This survey reviews many of these graph repositories and summarises the scope and contents of each repository. We identify opportunities for the development of repositories that can be queried in more flexible ways.
△ Less
Submitted 23 June, 2020;
originally announced June 2020.
-
Factorisation of Greedoid Polynomials of Rooted Digraphs
Authors:
Kai Siong Yow,
Kerri Morgan,
Graham Farr
Abstract:
Gordon and McMahon defined a two-variable greedoid polynomial $ f(G;t,z) $ for any greedoid $ G $. They studied greedoid polynomials for greedoids associated with rooted graphs and rooted digraphs. They proved that greedoid polynomials of rooted digraphs have the multiplicative direct sum property. In addition, these polynomials are divisible by $ 1 + z $ under certain conditions. We compute the g…
▽ More
Gordon and McMahon defined a two-variable greedoid polynomial $ f(G;t,z) $ for any greedoid $ G $. They studied greedoid polynomials for greedoids associated with rooted graphs and rooted digraphs. They proved that greedoid polynomials of rooted digraphs have the multiplicative direct sum property. In addition, these polynomials are divisible by $ 1 + z $ under certain conditions. We compute the greedoid polynomials for all rooted digraphs up to order six. A greedoid polynomial $ f(D) $ of a rooted digraph $ D $ of order $ n $ GM-factorises if $ f(D) = f(G) \cdot f(H) $ such that $ G $ and $ H $ are rooted digraphs of order at most $ n $ and $ f(G),f(H) \ne 1 $. We study the GM-factorability of greedoid polynomials of rooted digraphs, particularly those that are not divisible by $ 1 + z $. We give some examples and an infinite family of rooted digraphs that are not direct sums but their greedoid polynomials GM-factorise.
△ Less
Submitted 3 May, 2019; v1 submitted 9 September, 2018;
originally announced September 2018.
-
Tutte Invariants for Alternating Dimaps
Authors:
Kai Siong Yow,
Graham Farr,
Kerri Morgan
Abstract:
An alternating dimap is an orientably embedded Eulerian directed graph where the edges incident with each vertex are directed inwards and outwards alternately. Three reduction operations for alternating dimaps were investigated by Farr. A minor of an alternating dimap can be obtained by reducing some of its edges using the reduction operations. Unlike classical minor operations, these reduction op…
▽ More
An alternating dimap is an orientably embedded Eulerian directed graph where the edges incident with each vertex are directed inwards and outwards alternately. Three reduction operations for alternating dimaps were investigated by Farr. A minor of an alternating dimap can be obtained by reducing some of its edges using the reduction operations. Unlike classical minor operations, these reduction operations do not commute in general. A Tutte invariant for alternating dimaps is a function $ P $ defined on every alternating dimap and taking values in a field such that $ P $ is invariant under isomorphism and obeys a linear recurrence relation involving reduction operations. It is well known that if a graph $ G $ is planar, then the Tutte polynomial $ T $ satisfies $ T(G;x,y)=T(G^{*};y,x) $. We note an analogous relation for the extended Tutte invariants for alternating dimaps introduced by Farr. We then characterise the Tutte invariant for alternating dimaps of genus zero under several conditions. As a result of the non-commutativity of the reduction operations, the recursions based on them cannot always be satisfied. We investigate the properties of alternating dimaps of genus zero that are required in order to obtain a well defined Tutte invariant. Some excluded minor characterisations for these alternating dimaps are also given.
△ Less
Submitted 23 March, 2020; v1 submitted 14 March, 2018;
originally announced March 2018.
-
Topological containment of the 5-clique minus an edge in 4-connected graphs
Authors:
Rebecca Robinson,
Graham Farr
Abstract:
The topological containment problem is known to be polynomial-time solvable for any fixed pattern graph $H$, but good characterisations have been found for only a handful of non-trivial pattern graphs. The complete graph on five vertices, $K_5$, is one pattern graph for which a characterisation has not been found. The discovery of such a characterisation would be of particular interest, due to the…
▽ More
The topological containment problem is known to be polynomial-time solvable for any fixed pattern graph $H$, but good characterisations have been found for only a handful of non-trivial pattern graphs. The complete graph on five vertices, $K_5$, is one pattern graph for which a characterisation has not been found. The discovery of such a characterisation would be of particular interest, due to the Hajós Conjecture. One step towards this may be to find a good characterisation of graphs that do not topologically contain the simpler pattern graph $K_5^-$, obtained by removing a single edge from $K_5$.
This paper makes progress towards achieving this, by showing that every 4-connected graph must contain a $K_5^-$-subdivision.
△ Less
Submitted 3 May, 2017; v1 submitted 2 May, 2017;
originally announced May 2017.
-
Search strategies for develo** characterizations of graphs without small wheel subdivisions
Authors:
Rebecca Robinson,
Graham Farr
Abstract:
Practical algorithms for solving the Subgraph Homeomorphism Problem are known for only a few small pattern graphs: among these are the wheel graphs with four, five, six, and seven spokes. The length and difficulty of the proofs leading to these algorithms increase greatly as the size of the pattern graph increases. Proving a result for the wheel with six spokes requires extensive case analysis on…
▽ More
Practical algorithms for solving the Subgraph Homeomorphism Problem are known for only a few small pattern graphs: among these are the wheel graphs with four, five, six, and seven spokes. The length and difficulty of the proofs leading to these algorithms increase greatly as the size of the pattern graph increases. Proving a result for the wheel with six spokes requires extensive case analysis on many small graphs, and even more such analysis is needed for the wheel with seven spokes. This paper describes algorithms and programs used to automate the generation and testing of the graphs that arise as cases in these proofs. The main algorithm given may be useful in a more general context, for develo** other characterizations of SHP-related properties.
△ Less
Submitted 3 November, 2013;
originally announced November 2013.
-
Graphs with no 7-wheel subdivision
Authors:
Rebecca Robinson,
Graham Farr
Abstract:
The subgraph homeomorphism problem, SHP($H$), has been shown to be polynomial-time solvable for any fixed pattern graph $H$, but practical algorithms have been developed only for a few specific pattern graphs. Among these are the wheels with four, five, and six spokes. This paper examines the subgraph homeomorphism problem where the pattern graph is a wheel with seven spokes, and gives a result th…
▽ More
The subgraph homeomorphism problem, SHP($H$), has been shown to be polynomial-time solvable for any fixed pattern graph $H$, but practical algorithms have been developed only for a few specific pattern graphs. Among these are the wheels with four, five, and six spokes. This paper examines the subgraph homeomorphism problem where the pattern graph is a wheel with seven spokes, and gives a result that describes graphs with no $W_{7}$-subdivision, showing how they can be built up, using certain operations, from smaller `pieces' that meet certain conditions. We also discuss algorithmic aspects of the problem.
△ Less
Submitted 16 December, 2013; v1 submitted 3 November, 2013;
originally announced November 2013.
