IFIP Open Access Digital Library
Not a member yet
22614 research outputs found
Sort by
Advancing ScRNA-Seq Data Integration via a Novel Gene Selection Method
No full textPart 1: Biomedical/ClassificationInternational audienceCancer presents a formidable challenge in medical research, spurring efforts to demystify its underlying mechanisms towards advancing precision medicine, which aims at tailoring treatments to individuals’ genetic profiles. This study harnesses the power of single-cell RNA sequencing (scRNA-seq), a cutting-edge tool in next-generation sequencing, to delve into the transcriptomic intricacies of individual cells across diverse populations. Our methodology provides profound insights into gene expression patterns, significantly enhancing our understanding of cellular heterogeneity and its implications for cancer’s pathogenesis. To address the ’curse of dimensionality’ inherent in high-dimensional data, we introduce a sophisticated machine learning-based feature selection approach. This technique conceptualizes gene selection as a multi-label classification challenge, focusing on identifying genes critical for distinguishing between disease states and cell types. Importantly, our strategy underscores the value of data integration in reinforcing the statistical robustness of scRNA-seq analyses. By integrating disparate scRNA-seq datasets, we effectively mitigate batch effects, ensuring more accurate and reliable insights, thereby contributing significantly to the advancement of precision medicine in oncology
Are Mixture-of-Modality-Experts Transformers Robust to Missing Modality During Training and Inferring?
No full textPart 2: Natural Language ProcessingInternational audienceIt is commonly seen that the imperfect multi-modal data with missing modality appears in realistic application scenarios, which usually break the data completeness assumption of multi-modal analysis. Therefore, large efforts in multi-modal learning communities have been made on the robust solution for modality-missing data. Recently, pre-trained models based on Mixture-of-Modality-Experts (MoME) Transformers have been proposed, which achieved competitive performance in various downstream tasks, by utilizing different experts of feed-forward networks for single/multi modal inputs. One natural question arises: are Mixture-of-Modality-Experts Transformers robust to missing modality? To that end, in this paper, we conduct a deep investigation on MoME Transformer under the missing modality problem. Specifically, we propose a novel multi-task learning strategy, which leverages a uniform model to handle missing modalities during training and inference. In this way, the MoME Transformer will be empowered with robustness to missing modality. To validate the effectiveness of our proposed method, we conduct extensive experiments on three popular datasets, which indicate our method could outperform the state-of-the-art (SOTA) methods with a large margin
Stakeholder Engagement and the Use of Digital Tools in Strategy Processes in Swiss Municipalities
No full textPart 2: Digital TransformationInternational audienceLocal governments are striving to overcome their challenges by developing strategies. Some of them attempt to include leveraging digitalization and implementing data-centric procedures in strategy processes. However, there is a limited understanding of what these processes entail, such as who is involved and what digital tools are necessary for specific purposes and challenges. This paper aims to enhance and broaden this understanding. It employs a mixed-method approach by collecting both qualitative and quantitative data from Swiss municipalities and follows a convergent parallel design to analyze the data. The analysis reveals that larger municipalities tend to have established strategy practices, albeit in an ad-hoc manner without following any standard methods or best practices. The strategy development is significantly influenced by politicians and municipal administration. It mainly takes place around a limited group of internal individuals. So, it doesn’t meet the inclusivity component of Open Strategy. Citizen’s needs are considered by involving internal stakeholders; however, they mostly don’t directly participate in strategy processes. While municipalities do not perceive an immediate necessity for digital tools in their strategy processes, they acknowledge challenges in coordination, communication, and temporal resource management, for which they view digital tools as beneficial
Evaluating Learning Experiences-Comparison of Two Student Feedback Methods
No full textPart 2: Late-Breaking ResultsInternational audienceStudent feedback is crucial for the development of HCI learning experiences. Still, there is little research on how suitable the feedback methods are for gathering learning experiences and especially how the usage of the methods is experienced by the students. Hence, we have collected feedback using two different methods during an international two-week intensive course to understand the students’ experiences and to be able to compare the results. The methods were: the Retrospective Hand method and the Intensive Project Course Evaluation (IPCE) questionnaire method. Feedback was collected both during the course and on the last day of the course. By analyzing the feedback, we were able to iterate the course structure and content accordingly to meet the students’ needs. In this paper we describe the two methods and compare the results from both methods. We analyze the findings, discuss how these methods differ and discuss the usefulness of the feedback. Additionally, we advise how the two methods could be used in other courses for extending the communication between course teachers and students for improving the learning experiences
Urban Walkability: A Digital Twin Approach for Urban Regenerative Development
No full textPart 3: PhD Student Discussion ForumInternational audienceTraditional urban planning methods fall short in addressing the complexities of modern urbanization and environmental challenges. By adopting Generative Design, Virtual Reality, and Digital Twins, the aim is to make urban planning more interactive, accessible, and inclusive. The research is grounded in a comprehensive literature review across related fields, including Digital Twins’ dynamic modeling capabilities and Virtual Reality’s immersive visualization potential. Initial findings from sit-down interviews with Eindhoven residents reveal critical insights into pedestrian preferences. In my next steps in the PhD, I will focus on developing Virtual Reality prototypes to further examine walkability and Generative Design prototypes to generate urban planning aligning with regenerative development toward walkability. At the end of my PhD, the knowledge from previous prototypes is used to create a tool to simplify urban planning, including multiple stakeholders, to improve walkability in an urban context. This innovative approach promises to contribute significantly to the creation of livable, sustainable urban environments by leveraging cutting-edge technologies to address key challenges in urban walkability and planning
The Dual Nature of Organizational Policies
No full textPart 4: Security, Compliance, and Configuration in Enterprise ModelingInternational audienceOrganizational decisions are usually constrained by policies and rules, sometimes up to the point of completely automated decision making. Policies exist on multiple levels within the organization and require organizational power to be created. They are typically expressed in a policy document that has multiple practical functions. Drawing on and extending the UFO-L ontology on legal positions, we offer a critical analysis of the relationship between policy and policy document and make an argument for the dual nature of policies. We present four ontological patterns that we claim to be fundamental for organizations as social and economic phenomena. These patterns address: (1) organizational coordination and policy, (2) policy documents, (3) delegation, (4) community ruling. For an initial evaluation, the patterns are exemplified in the university domain
Enriching Business Process Event Logs with Multimodal Evidence
No full textPart 3: Process Mining and Business Process AnalysisInternational audienceProcess mining uses data from event logs to understand which activities were undertaken, their timing, and the involved entities, providing a data trail for process analysis and improvement. However, a significant challenge involves ensuring that these logs accurately reflect the actual processes. Some processes leave few digital traces, and their event logs often lack details about manual and physical work that does not involve computers or simple sensors. We introduce the Business-knowledge Integration Cycles (BICycle) method and mm_proc_miner tool to convert raw and unstructured data from various modalities, such as video, audio, and sensor data, into a structured and unified event log, while keeping human-in-the-loop. Our method analyzes the semantic distance between visible, audible, and textual evidence within a self-hosted joint embedding space. Our approach is designed to consider (1) preserving the privacy of evidence data, (2) achieving real-time performance and scalability, and (3) preventing AI hallucinations. We also publish a dataset consisting of over 2K processes with 16K steps to facilitate domain inference-related tasks. For the evaluation, we created a novel test dataset in the domain of DNA home kit testing, for which we can guarantee that it was not encountered during the training of the employed AI foundational models. We show positive insights in both event log enrichment with multimodal evidence and human-in-the-loop contribution
Cellular Automata and Discrete Complex Systems: 30th IFIP WG 1.5 International Workshop, AUTOMATA 2024, Durham, UK, July 22–24, 2024, Proceedings
No full textInternational audienceBook Front Matter of LNCS 1478
Roots in the semiring of finite deterministic dynamical systems
No full textPart 2: Contributed PapersInternational audienceFinite discrete-time dynamical systems (FDDS) model phenomena that evolve deterministically in discrete time. It is possible to define sum and product operations on these systems (disjoint union and direct product, respectively) giving a commutative semiring. This algebraic structure led to several works employing polynomial equations to model hypotheses on phenomena modelled using FDDS. To solve these equations, algorithms for performing the division and computing k-th roots are needed. In this paper, we propose two polynomial algorithms for these tasks, under the condition that the result is a connected FDDS. This ultimately leads to an efficient solution to equations of the type AXk = B for connected X. These results are some of the important final steps for solving more general polynomial equations on FDDS
Mining Profitability in Bitcoin: Calculations of User-Miner Equilibria and Cost of Mining
No full textInternational audienceThis paper examines the equilibrium between user transaction fees and miner profitability within proof-of-work-based blockchains, specifically focusing on Bitcoin. We analyze the dependency of mining profit on factors such as transaction fee adjustments and operational costs, particularly electricity. By applying a multidimensional profitability model and performing a sensitivity analysis, we evaluate the potential for profit maximization through operational cost reduction versus fee increases. Our model integrates variable electricity costs, market-driven Bitcoin prices, mining hardware efficiency, network hash rate, and transaction fee elasticity. We show that mining strategies aimed at reducing electricity expenses are far more profitable than pursuing transactions with higher fees