Diverging assessment maintains a common question set for all students but varies the input data so that each student has a unique problem to solve. It is an approach in student assessment that offers a unique and authentic learning experience. Although such assessments have been implemented in computing courses, their effectiveness and students' perceptions in different contexts remain unexplored. In this paper, we investigate student perspectives on diverging assessment. We surveyed students in four courses across three different universities. Each surveyed student was enrolled in one of the four courses on networking, operation systems, digital forensics or ethical hacking. Each course featured at least one diverging assessment. The students' overall perceptions about diverging assessments and three different aspects of diverging assessment, namely authenticity, assessment-as-learning, and academic integrity, are surveyed, reported, and analyzed.

In this experience paper, we introduce the concept of 'diverging assessments', process-based assessments designed so that they become unique for each student while all students see a common skeleton. We present experiences with diverging assessments in the contexts of computer networks, operating systems, ethical hacking, and software development. All the given examples allow the use of generative-AI-based tools, are authentic, and are designed to generate learning opportunities that foster students' meta-cognition. Finally, we reflect upon these experiences in five different courses across four universities, showing how diverging assessments enhance students' learning while respecting academic integrity.

Confluence is a novel non-Intersection over Union (IoU) alternative to Non-Maxima Suppression (NMS) in bounding box post-processing in object detection. It overcomes the inherent limitations of IoU-based NMS variants to provide a more stable, consistent predictor of bounding box clustering by using a normalized Manhattan Distance inspired proximity metric to represent bounding box clustering. Unlike Greedy and Soft NMS, it does not rely solely on classification confidence scores to select optimal bounding boxes, instead selecting the box which is closest to every other box within a given cluster and removing highly confluent neighboring boxes. Confluence is experimentally validated on the MS COCO and CrowdHuman benchmarks, improving Average Precision by 0.2--2.7% and 1--3.8% respectively and Average Recall by 1.3--9.3 and 2.4--7.3% when compared against Greedy and Soft-NMS variants. Quantitative results are supported by extensive qualitative analysis and threshold sensitivity analysis experiments support the conclusion that Confluence is more robust than NMS variants. Confluence represents a paradigm shift in bounding box processing, with potential to replace IoU in bounding box regression processes.

In order to better understand conditions that lead to methane explosions in underground coal mines, we apply machine learning to data collected in an industrial scale research project carried out at the University of Newcastle, Australia, 2014–2018 (VAM Abatement Safety Project). We present a comparison of five different methods (Decision Tree, Random Forest, Naïve Bayes, AdaBoostM1, and SVM with SMO) to classify the maximum pressure and maximum flame velocity in order to predict detonation and inform the design of capture ducts. All methods are evaluated with a tenfold cross validation technique. We found that tree-based classification methods provide the most accurate prediction of dangerous pressure and supersonic velocity.

This paper surveys the main computer search results for finding graceful labelling of trees. The paper is devoted to the memory of Mirka Miller, who made an outstanding contribution to the area of graph labelling.

We survey the most important results regarding the domination chain parameters, including the characterisation of the domination sequence, complexity of exact and parameterised algorithms, and approximation and inapproximability ratios. We provide a number of new results for the upper and lower irredundance and their complements, and a few results for other domination chain problems. In particular, we analyse the structure of maximum irredundant sets and we provide new bounds on the upper irredundance number. We show several approximability results for upper and lower irredundance and their complements on general graphs; all four problems remain NP-hard even on planar cubic graphs and APX-hard on cubic graphs. Finally, we give some results on everywhere dense graphs, and study some related extension problems.

Global warming is a long-term environmental hazard demonstrated by a gradual increase in the temperature of the Earth. It is caused by the accumulation of greenhouse gases in the atmosphere, including carbon dioxide and methane. Although, in terms of the volume, methane is considered secondary to carbon dioxide, it is about 21 times more damaging when compared over a 100-year period. Fugitive methane emissions from underground coal mines significantly contribute to global warming. Amongst all the known methods to reduce the fugitive methane, application of thermal oxidation (or, simply, burning) is deemed the most effective and practical. This process produces water vapour and carbon dioxide, which has significantly lower adverse impact on the atmosphere than methane. The thermal oxidisers operate at high temperatures, which may introduce a risk of fire and explosion to the mine. In order to mitigate such risk, a thorough understanding of the methane explosion characteristics is essential. Methane fire and explosion experiments under conditions pertinent to underground coal mines are expensive, risky and necessitate significant effort, and thus require enormous preparation and safety procedures. It is cheaper and safer to analyse existing data to discover patterns and predict explosions than to conduct new extensive experiments. In this paper, we present a comparative study of data mining and machine learning techniques used for these purposes.

This paper studies Upper Domination, i.e., the problem of computing the maximum cardinality of a minimal dominating set in a graph with respect to classical and parameterised complexity as well as approximability.

The concept of face-antimagic labeling of plane graphs was introduced by Mirka Miller in 2003. This survey aims to give an overview of the recent results obtained in this topic.

We consider Upper Domination, the problem of finding a maximum cardinality minimal dominating set in a graph. We show that this problem does not admit an \(n^{1-\epsilon }\) approximation for any \(\epsilon >0\), making it significantly harder than Dominating Set, while it remains hard even on severely restricted special cases, such as cubic graphs (APX-hard), and planar subcubic graphs (NP-hard). We complement our negative results by showing that the problem admits an \(O(\varDelta )\) approximation on graphs of maximum degree \(\varDelta \), as well as an EPTAS on planar graphs. Along the way, we also derive essentially tight \(n^{1-\frac{1}{d}}\) upper and lower bounds on the approximability of the related problem Maximum Minimal Hitting Set on d-uniform hypergraphs, generalising known results for Maximum Minimal Vertex Cover.

In this paper, we survey and supplement the complexity landscape of the domination chain parameters as a whole, including classifications according to approximability and parameterised complexity. Moreover, we provide clear pointers to yet open questions. As this posed the majority of hitherto unsettled problems, we focus on Upper Irredundance and Lower Irredundance that correspond to finding the largest irredundant set and resp. the smallest maximal irredundant set. The problems are proved NP-hard even for planar cubic graphs. While Lower Irredundance is proved not clog (n) -approximable in polynomial time unless NP⊆DTIME(n^loglog n) , no such result is known for Upper Irredundance. Their complementary versions are constant-factor approximable in polynomial time. All these four versions are APX-hard even on cubic graphs.

Providing a global understanding of privacy is crucial, because everything is connected. Nowadays companies are providing their customers with more services that will give them more access to their data and daily activity; electricity companies are marketing the new smart meters as a new service with great benefit to reduce the electricity usage by monitoring the electricity reading in real time. Although the users might benefit from this extra service, it will compromise the privacy of the users by having constant access to the readings. Since the smart meters will provide the users with real electricity readings, they will be able to decide and identify which devices are consuming energy in that specific moment and how much it will cost. This kind of information can be exploited by numerous types of people. Unauthorized use of this information is an invasion of privacy and may lead to much more severe consequences. This paper will propose an algorithm approach for the comparison and analysis of Smart Meter data readings, considering the time and temperature factors at each second to identify the use patterns at each house by identifying the appliances activities at each second in time complexity O(log(m)).

This paper studies Upper Domination, i.e., the problem of computing the maximum cardinality of a minimal dominating set in a graph with respect to classical and parameterised complexity as well as approximability.

DNA sequencing is the process of determining the exact order of the nucleotide bases of an individual's genome in order to catalogue sequence variation and understand its biological implications. Whole-genome sequencing techniques produce masses of data in the form of short sequences known as reads. Assembling these reads into a whole genome constitutes a major algorithmic challenge. Most assembly algorithms utilize de Bruijn graphs constructed from reads for this purpose. A critical step of these algorithms is to detect typical motif structures in the graph caused by sequencing errors and genome repeats, and filter them out; one such complex subgraph class is a so-called superbubble. In this paper, we propose an O(n+m)-time algorithm to detect all superbubbles in a directed acyclic graph with n nodes and m (directed) edges, improving the best-known O(m log m)-time algorithm by Sung et al.

The Minimum Vertex Cover (MinVC) problem is a well-known NP-hard problem. Recently there has been great interest in solving this problem on real-world massive graphs. For such graphs, local search is a promising approach to finding optimal or near-optimal solutions. In this paper we propose a local search algorithm that exploits reduction rules and data structures to solve the MinVC problem in such graphs. Experimental results on a wide range of real-word massive graphs show that our algorithm finds better covers than state-of-the-art local search algorithms for MinVC. Also we present interesting results about the complexities of some well-known heuristics.

It is well recognised that data mining and statistical analysis pose a serious treat to privacy. This is true for financial, medical, criminal and marketing research. Numerous techniques have been proposed to protect privacy, including restriction and data modification. Recently proposed privacy models such as differential privacy and k-anonymity received a lot of attention and for the latter there are now several improvements of the original scheme, each removing some security shortcomings of the previous one. However, the challenge lies in evaluating and comparing privacy provided by various techniques. In this paper we propose a novel entropy based security measure that can be applied to any generalisation, restriction or data modification technique. We use our measure to empirically evaluate and compare a few popular methods, namely query restriction, sampling and noise addition.

In order to protect privacy of social network participants, network graph data should be anonymised prior to its release. Most proposals in the literature aim to achieve k -anonymity under specific assumptions about the background information available to the attacker. Our method is based on randomizing the location of the triangles in the graph. We show that this simple method preserves the main structural parameters of the graph to a high extent, while providing a high re-identification confusion.