The HYPER-AI project aims to revolutionize the field of distributed computing by addressing the complexities of integrating Internet of Things (IoT), Edge, and Cloud computing domains. It focuses on the creation of smart virtual computing nodes, optimizing data-processing applications across a distributed network to enhance performance and efficiency. HYPER-AI leverages autonomous, self-organizing computing swarms, enabling dynamic connectivity and cooperation among diverse computing resources. This approach facilitates the seamless integration of resources across all levels of the computing continuum, from edge devices to cloud infrastructure, through innovative semantic representation and orchestration techniques.
The project envisions a new era of computing ecosystems that are adaptable, efficient, and capable of supporting highly demanding applications in various domains such as healthcare and energy, thus contributing significantly to the advancement of digital technologies and sustainable development goals.
The DMML group from HES-SO participates in the project, to provide Reinforcement learning expertise. More particularly, our group will focus on Hierarchical Multi-Agent Deep Reinforcement Learning for Data Management.
Read more
CoORDinance - Shaping A Code of Conduct Concerning the ORD of the Indoor Positioning Research Field
The CoORDinance project aims to develop a comprehensive Code of Conduct for Open Research Data (ORD) in the Indoor Positioning (IP) field, addressing the current lack of standardized ORD practices and its fragmented ORD landscape. CoORDinance builds on the findings of the previous CoORDinates project, which performed a systematic study of the ORD landscape of the field, identified significant challenges of the IP community in terms of its ORD practices, and elaborated on lessons learned, on shortcomings, and identified best practices that emerged through its systematic study. CoORDinance aims to capitalize on these solid foundations to propose a clear Code of Conduct, in line with the Swiss National ORD Strategy and the FAIR principles. This initiative will involve collaborating with leading experts to ensure the guidelines are practical, comprehensive, and widely accepted. By promoting principles of Open Science, the project seeks to promote rigorous, transparent and reproducible (RTR) research practices in IP research, ultimately fostering a culture of collaboration and rigor, assisting the advancement of the field.
CoORDinance will integrate the lessons learned from CoORDinates project, to develop a concise and unambiguous Code of Conduct that will facilitate the motivated researcher of the field to engage with ORD, responsibly and adequately. CoORDinance aims to focus deeper on ethical considerations and scientific best practices, ensuring that the guidelines not only cover practical and technical aspects but also emphasize the ethical responsibilities of data providers and users. To achieve this, we will conduct a thorough review of existing codes of conduct and guidelines of other fields to inform our approach, ensuring that the Code of Conduct we develop is grounded in proven best practices.
The project is funded by the Swiss Academies of Arts and Sciences.
Read more
CoORDinates – Towards a Consensus-based ORD Standard for the Indoor Positioning Community
The community of indoor positioning research has identified the need for a paradigm shift towards more reproducible and open research. Over the last years, a slowly increasing tendency of openly sharing data and code has been observed in field, facilitating the establishment of baselines and benchmarks, assisting the reproducibility, verifiability, and comparability of scientific results. Nevertheless, accompanying research results with ORD still constitutes the exception rather than the rule in publications of the field. Moreover, since there exist no relevant standards or suggested guidelines, important aspects regarding the data format, the metadata and the overall documentation are typically decided individually by the authors. The data that might accompany scientific publications of the field usually do not undergo peer review as the published manuscripts does, and therefore the extent to which those data are well documented and reusable in practice relies on the rigour and the motivation of the authors.
In this project, we aim at exploring the existing landscape of ORD within the Indoor Positioning community, specifying guidelines leveraging experiences of existing ORD practices of other fields, moving towards a community consensus on the ORD practices that the Indoor Positioning community can adopt. To achieve these goals, we will engage active researchers and other stakeholders with track record on ORD sharing, planning research synergies which will culminate in a public event of this community. The outcome of the project will be a clear step towards a consensus-based ORD standard for the Indoor Positioning community.
The CoORDinates project has been funded by the Swiss Open Research Data Grants (CHORD) of SwissUniversities, under the Track A: Explore Projects funding tool. Moreover, HES-SO has also supported the project, guaranteeing the needed matching funds.
Results at a glance:
In the work "Evaluating Open Science Practices in Indoor Positioning and Indoor Navigation Research", we performed a comprehensive review of Open Science practices followed in recent publications in the field, analyzing all reference papers from the 2022 and 2023 ... Read more
Drug discovery is a well-known and challenging problem. Typically, one needs to navigate through a vast chemical space of up to \( 10^{60} \) small organic molecules to find a potential drug candidate with desired properties. Such a trial-and-error process is often cumbersome and error-prone. However, there is an abundance of molecule-related data in the real world, and data-driven approaches can be a promising direction to expedite the molecule design stage, which is a pivotal part of the entire chemical development.
Machine learning has revolutionized many fields including computer vision, natural language processing. Despite its current success, machine learning-based algorithms are proficient only at solving the forward problem, in which given a specific input structure there is an associated unique property that we want to predict. In contrast, generating a new molecule to meet specific property requirements is considered as an inverse problem, where there is a one-to-many mapping between the required property generation and structurally diverse molecules. This mapping is highly challenging and constitutes a fundamental aspect of molecule-related inverse problems. Furthermore, the underlying nature of molecules is based on graph structures, and the direct optimization over the discrete representation to find the target generated molecule property is a hard problem in the literature.
In this project, we aim to leverage generative models, a type of method that has demonstrated its effectiveness in solving the encountered problem as well as a more general setting of inverse modeling. The project spans over two overarching objectives:
to develop a single-step generation for a molecule given some expected properties;
to do style transfer.
The latter objective seeks to modify the structure of a given molecule with known properties to a new molecule by slightly adjusting the provided molecule structure to exhibit some desired characteristics. Previous studies in the literature solved these problems with a two-stage process that includes molecule property optimization in the second stage. However, this additional process is computationally prohibitive and requires retraining different target values for inference, which poses significant challenges.
Another limitation of such ... Read more
Infrastructure assets, such as bridges, need to be inspected regularly. Our objective is to reduce the need for human involvement, minimize risks to health and safety, decrease the impact of subjective engineering assessments, digitize asset management, and promote sustainable inspection practices. By achieving these goals, we aim to develop optimal maintenance strategies for infrastructure assets, leading to better long-term outcomes.
The project proposes an innovative solution to bridge inspection and condition assessment, which combines the use of UAV (drone) flights with automated defect detection using AI. This approach represents a globally emerging trend in infrastructure management.
The developed platform will allow inspections and condition assessments to take place directly on the bridge's digital twin, which will increase efficiency on both inspection and maintenance sides. Targeted interventions can then be made, leading to prolonged life spans and enhanced sustainability effects.
To achieve these benefits, the project will use raw images captured during UAV flights and automate the defect recognition process without any manual interventions.
The project is funded by Innosuisse, and is a collaboration of our team with the innovative company LeanBI and OST.
The role of our group within the automated bridge defect recognition are twofold:
Develop Self-Supervised learning models able to exploit the images acquired by the UAVs, reducing the annotation effort required for new downstream task (new bridge types, new defect types). The goal is the be able to train efficiently new segmentation models with a small number of annotated images. This would facilitate the development of new models for new types of infrastructure or defect. To this end, our research are focus on the latest methods in Contrastive Learning applied to image segmentation and Data Augmentation.
Develop automatic Out-Of-Distribution method working with image segmentation models. Here, we want first to develop methods able to detect if the sample analyzed (a bridge) is too different from the training samples of the model currently used and, secondly, to adapt them accordingly to the change.
Partners:
Read more