Högskolan i Skövde

Practitioners of multi-objective optimization currently lack open tools that provide decision support through knowledge discovery. There exist many software platforms for multi-objective optimization, but they often fall short of implementing methods for rigorous post-optimality analysis and knowledge discovery from the generated solutions. This paper presents Mimer, a multi-criteria decision support tool for solution exploration, preference elicitation, knowledge discovery, and knowledge visualization. Mimer is openly available as a web-based tool and uses state-of-the-art web-technologies based on WebAssembly to perform heavy computations on the client-side. Its features include multiple linked visualizations and input methods that enable the decision maker to interact with the solutions, knowledge discovery through interactive data mining and graph-based knowledge visualization. It also includes a complete Python programming interface for advanced data manipulation tasks that may be too specific for the graphical interface. Mimer is evaluated through a user study in which the participants are asked to perform representative tasks simulating practical analysis and decision making. The participants also complete a questionnaire about their experience and the features available in Mimer. The survey indicates that participants find Mimer useful for decision support. The participants also offered suggestions for enhancing some features and implementing new features to extend the capabilities of the tool.

Automatic recognition of grocery products can be used to improve customer flow at checkouts and reduce labor costs and store losses. Product recognition is, however, a challenging task for machine learning-based solutions due to the large number of products and their variations in appearance. In this work, we tackle the challenge of fine-grained product recognition by first extracting a large dataset from a grocery store containing products that are only differentiable by subtle details. Then, we propose a multimodal product recognition approach that uses product images with extracted OCR text from packages to improve fine-grained recognition of grocery products. We evaluate several image and text models separately and then combine them using different multimodal models of varying complexities. The results show that image and textual information complement each other in multimodal models and enable a classifier with greater recognition performance than unimodal models, especially when the number of training samples is limited. Therefore, this approach is suitable for many different scenarios in which product recognition is used to further improve recognition performance. The dataset can be found at https://github.com/Tubbias/finegrainocr.

A team of fighter pilots in a distributed environment with limited access to information rely on technology to pursue teamwork. In order to design systems that support distributed teamwork, it is, therefore, necessary to understand how access to information affects the team members. Certain factors, such as mutual performance monitoring, shared mental models, adaptability, and backup behavior are considered essential for effective teamwork. We investigate these factors in this work, focusing on how visually communicated information affects fighter pilots’ perception of these factors. For that, a questionnaire including the teamwork factors in relation to certain defined scenarios that contain various levels of information was distributed to fighter pilots. We show that the studied factors are affected by the level of information available to the pilots. Especially, mutual performance monitoring increases with the degree of available information. © 2023 Human Factors and Ergonomics Society.

Sales and Operations Planning (S&OP) is a process that aims to align dimensioning efforts in a company, based on one integrated plan and with clear decision milestones. The alignment is cross-functional and connects different operations functions with each other to set an overall delivery ability. There are always challenges connecting different functions in a company which most S&OP practitioners agree with, still, that is one of the things that the S&OP-process should bridge. Digital solutions such as Enterprise Resource Planning (ERP) and other more or less sophisticated tools have contributed to an improved cross functional communication over time. S&OP in an Engineer-to-order (ETO) context, especially where engineering is a major or an equal portion as e.g., make-to-stock (MTS) and make-to-order (MTO) contexts, may experience even further challenges. Technologies within Industry 4.0 are changing the way S&OP is carried out; one of the most relevant ones is Artificial Intelligence (AI), particularly, Machine Learning (ML) that analyses data collected during these processes to find patterns and extract knowledge. The intent with this paper is to, based on S&OP-challenges, see if ML can be used to improve these challenges.

In a brief literature review together with empiric data from a single industrial case (SIC), S&OP-challenges were defined and structured. Based on the challenges in several S&OP-sub-areas, classified into data quality, horizontal and vertical disconnects, specific tasks were specified and structured into anomaly detection, clustering and classification, and predictions. Which exact ML-method to use require further work and tests. Still, this is a good starting point to take the next step and the specified tasks could also be used for other practitioners that want to start using ML/AI in their daily activities.

Sales and Operations Planning (S&OP) is a process that aims to align dimensioning efforts in a company, based on the "One Plan" and with clear decision milestones, where “One Plan” relates to the ultimate outcome of S&OP by integrating multiple plans. This alignment is cross functional and connects, not only sales and operations, but also different operations functions with each other, to set an overall delivery ability. There are always challenges when connecting different functions in a company, something most S&OP practitioners agree with, still, cross functional integration is one of the things that the S&OP-process addresses. For S&OP in an Engineer-to-order (ETO) context, especially where engineering is a major or an equal portion of the product as e.g., make-to-stock (MTS) or make-to-order (MTO) contexts, further complexity is added. If these businesses also have long lead times and low volumes, another perspective to the S&OP-process is given when it comes to the balance between demand and supply (DS). Digital solutions such as Enterprise Resource Planning (ERP) and other more or less sophisticated tools are a pre-requisite for the S&OP-process and improves cross functional integration. Technologies within Industry 4.0 are changing the way S&OP is carried out; one of the most relevant one is Artificial Intelligence (AI), particularly, Machine Learning (ML) that analyses data collected during these processes to find patterns and extract knowledge.

 Therefore, in this paper, the purpose is to investigate and define the main sub-areas of the S&OP-process in an ETO-context and discuss how AI, in particular ML, currently supports the sub-areas. To be able to fulfil the purpose, a literature study of the two main fields, S&OP and AI, has been carried out.

 The results are pointing at an underuse of ML-techniques for S&OP. Forecasting in MTS- context is where ML is mostly used, and the most common ML-technique is Artificial Neutral Networks (ANN) which is considered as Supervised Learning. The results of this paper will serve as a starting point for further research on the efforts and effects required for improving the S&OP-process in an ETO-context and with what ML-techniques.

A common approach for analyzing large-scale molecular data is to cluster objects sharing similar characteristics. This assumes that genes with highly similar expression profiles are likely participating in a common molecular process. Biological systems are extremely complex and challenging to understand, with proteins having multiple functions that sometimes need to be activated or expressed in a time-dependent manner. Thus, the strategies applied for clustering of these molecules into groups are of key importance for translation of data to biologically interpretable findings. Here we implemented a multi-assignment clustering (MAsC) approach that allows molecules to be assigned to multiple clusters, rather than single ones as in commonly used clustering techniques. When applied to high-throughput transcriptomics data, MAsC increased power of the downstream pathway analysis and allowed identification of pathways with high biological relevance to the experimental setting and the biological systems studied. Multi-assignment clustering also reduced noise in the clustering partition by excluding genes with a low correlation to all of the resulting clusters. Together, these findings suggest that our methodology facilitates translation of large-scale molecular data into biological knowledge. The method is made available as an R package on GitLab (https://gitlab.com/wolftower/masc).

The growing neural gas (GNG) is an unsupervised topology learning algorithm that models a data space through interconnected units that stand on the populated areas of that space. Its output is a graph that can be visually represented on a two-dimensional plane, and be used as means to disclose cluster patterns in datasets. GNG, however, creates highly connected graphs when trained on high dimensional data, which in turn leads to highly clutter representations that fail to disclose any meaningful patterns. Moreover, its sequential learning limits its potential for faster executions on local datasets, and, more importantly, its potential for training on distributed datasets while leveraging from the computational resources of the infrastructures in which they reside.

This paper presents two methods that improve GNG for the visualization of cluster patterns in large and high-dimensional datasets. The first one focuses on providing more meaningful and accurate cluster pattern representations of high-dimensional datasets, by avoiding connections that lead to high-dimensional graphs in the modeled topology, which may, in turn, lead to visual cluttering in 2D representations. The second method presented in this paper enables the use of GNG on big and distributed datasets with faster execution times, by modeling and merging separate parts of a dataset using the MapReduce model.

Quantitative and qualitative evaluations show that the first method leads to the creation of lower-dimensional graph structures, which in turn provide more accurate and meaningful cluster representations; and that the second method preserves the accuracy and meaning of the cluster representations while enabling its execution in distributed settings.

The increasing use of artificial intelligence (AI) technologies across application domains has prompted our society to pay closer attention to AI's trustworthiness, fairness, interpretability, and accountability. In order to foster trust in AI, it is important to consider the potential of interactive visualization, and how such visualizations help build trust in AI systems. This manifesto discusses the relevance of interactive visualizations and makes the following four claims: i) trust is not a technical problem, ii) trust is dynamic, iii) visualization cannot address all aspects of trust, and iv) visualization is crucial for human agency in AI.

We are largely used to hearing explanations. For example, if someone thinks you are sad today, they might reply to your “why?” with “because you were so Hmmmmm-mmm-mmm”. Today’s Artificial Intelligence (AI), however, is – if at all – largely providing explanations of decisions in a visual or textual manner. While such approaches are good for communication via visual media such as in research papers or screens of intelligent devices, they may not always be the best way to explain; especially when the end user is not an expert. In particular, when the AI’s task is about Audio Intelligence, visual explanations appear less intuitive than audible, sonified ones. Sonification has also great potential for explainable AI (XAI) in systems that deal with non-audio data – for example, because it does not require visual contact or active attention of a user. Hence, sonified explanations of AI decisions face a challenging, yet highly promising and pioneering task. That involves incorporating innovative XAI algorithms to allow pointing back at the learning data responsible for decisions made by an AI, and to include decomposition of the data to identify salient aspects. It further aims to identify the components of the preprocessing, feature representation, and learnt attention patterns that are responsible for the decisions. Finally, it targets decision-making at the model-level, to provide a holistic explanation of the chain of processing in typical pattern recognition problems from end-to-end. Sonified AI explanations will need to unite methods for sonification of the identified aspects that benefit decisions, decomposition and recomposition of audio to sonify which parts in the audio were responsible for the decision, and rendering attention patterns and salient feature representations audible. Benchmarking sonified XAI is challenging, as it will require a comparison against a backdrop of existing, state-of-the-art visual and textual alternatives, as well as synergistic complementation of all modalities in user evaluations. Sonified AI explanations will need to target different user groups to allow personalisation of the sonification experience for different user needs, to lead to a major breakthrough in comprehensibility of AI via hearing how decisions are made, hence supporting tomorrow’s humane AI’s trustability. Here, we introduce and motivate the general idea, and provide accompanying considerations including milestones of realisation of sonifed XAI and foreseeable risks.

This article presents an empirical user study that compares eight multidimensional projection techniques for supporting the estimation of the number of clusters, k, embedded in six multidimensional data sets. The selection of the techniques was based on their intended design, or use, for visually encoding data structures, that is, neighborhood relations between data points or groups of data points in a data set. Concretely, we study: the difference between the estimates of k as given by participants when using different multidimensional projections; the accuracy of user estimations with respect to the number of labels in the data sets; the perceived usability of each multidimensional projection; whether user estimates disagree with k values given by a set of cluster quality measures; and whether there is a difference between experienced and novice users in terms of estimates and perceived usability. The results show that: dendrograms (from Ward's hierarchical clustering) are likely to lead to estimates of k that are different from those given with other multidimensional projections, while Star Coordinates and Radial Visualizations are likely to lead to similar estimates; t-Stochastic Neighbor Embedding is likely to lead to estimates which are closer to the number of labels in a data set; cluster quality measures are likely to produce estimates which are different from those given by users using Ward and t-Stochastic Neighbor Embedding; U-Matrices and reachability plots will likely have a low perceived usability; and there is no statistically significant difference between the answers of experienced and novice users. Moreover, as data dimensionality increases, cluster quality measures are likely to produce estimates which are different from those perceived by users using any of the assessed multidimensional projections. It is also apparent that the inherent complexity of a data set, as well as the capability of each visual technique to disclose such complexity, has an influence on the perceived usability.