| 1. |
- Lindgren, Erik, 1980, et al.
(author)
-
Analysis of industrial X-ray computed tomography data with deep neural networks
- 2021
-
In: Proceedings of SPIE - The International Society for Optical Engineering. - : SPIE. - 0277-786X .- 1996-756X. ; 11840
-
Conference paper (peer-reviewed)abstract
- X-ray computed tomography (XCT) is increasingly utilized industrially at material- and process development as well as in non-destructive quality control; XCT is important to many emerging manufacturing technologies, for example metal additive manufacturing. These trends lead to increased needs of safe automatic or semi-automatic data interpretation, considered an open research question for many critical high value industrial products such as within the aerospace industry. By safe, we mean that the interpretation is not allowed to unawarely or unexpectedly fail; specifically the algorithms must react sensibly to inputs dissimilar to the training data, so called out-of-distribution (OOD) inputs. In this work we explore data interpretation with deep neural networks to address: robust safe data interpretation which includes a confidence estimate with respect to OOD data, an OOD detector; generation of realistic synthetic material aw indications for the material science and nondestructive evaluation community. We have focused on industrial XCT related challenges, addressing difficulties with spatially correlated X-ray quantum noise. Results are reported on training auto-encoders (AE) and generative adversarial networks (GAN), on a publicly available XCT dataset of additively manufactured metal. We demonstrate that adding modeled X-ray noise during training reduces artefacts in the generated imperfection indications as well as improves the OOD detector performance. In addition, we show that the OOD detector can detect real and synthetic OOD data and still model the accepted in-distribution data down to the X-ray noise levels.
|
|
| 2. |
- Amoson, Jonas, 1973-
(author)
-
Building complex GUIs in Plan 9
- 2009
-
In: Proceedings 4th International Workshop on Plan9. - Athens, GA : University of Georgia. ; , s. 15-21
-
Conference paper (other academic/artistic)abstract
- How can non-trivial graphical user interfaces be designed in Plan 9 without them losing their minimalistic style? Different toolkits are discussed, and a proposal for a tabbed toolbar is suggested as a way to add functionality without cluttering the interface and avoiding the use of pop-up dialog boxes. A hypothetical port to the GUI in LyX is used as an example.
|
|
| 3. |
- Lindgren, Erik, et al.
(author)
-
Deep-learning-based out-of distribution data detection in visual inspection images
- 2023
-
In: Proceedings Of Spie 12489, NDE 4.0, Predictive Maintenance, Communication, and Energy Systems. - : Spie Digital Library. ; 12489
-
Conference paper (peer-reviewed)abstract
- Within quality critical industries, e.g. aerospace, quality control with non-destructive evaluation (NDE) is essential. The surface quality is often important and e.g. visual inspection is often applied. Part of the inspection is the data interpretation, not easily made automatic for critical products. Recent studies on the automatization have indicated promising results utilizing deep-learning-based artificial intelligence. However, many such algorithms are known to be overconfident when subjected to unexpected input (e.g. new/rare material defects) far from the training dataset, so-called out-of-distribution (OOD) data. We claim that safe computer-based interpretation of NDE data within quality critical applications, must respond sensible also to OOD data. A sensible response could be that the algorithms identify such OOD data and forward it to a human for further analysis. Such an OOD detector could facilitate a human-machine collaboration in a NDE 4.0 vision. In this work we have explored if a recently proposed (for industrial x-ray images) auto-encoder-based approach can be utilized as OOD detector (one-class classifier) for visual inspection data. The model is trained in an unsupervised manner on accepted input to reconstruct it at high precision. Simultaneously it is trained to remove synthetically added defect indications to generate a clean image patch, similar to denoising-auto-enoders. The difference between the input and reconstructed input is analyzed for OOD detection. We train and test the algorithm on a publicly available visual inspection dataset with surface defects. We achieve true positive rates at 0.90 with true negative rates at 0.99 and demonstrate detection of OOD data.
|
|
| 4. |
- Ljungdahl Eriksson, Martin, 1982-, et al.
(author)
-
My Sound Space : An attentional shield for immersive redirection
- 2018
-
In: Audio Mostly 2018. - New York : Association for Computing Machinery (ACM). - 9781450366090
-
Conference paper (peer-reviewed)abstract
- In the context of extended reality, the term immersion is commonly used as a property denoting to which extent a technology can deliver an illusion of reality while occluding the users’ sensory access to the physical environment. In this paper we discuss an alternative interpretation of immersion, used in the My Sound Space project. The project is a research endeavor aiming to develop a sound environment system that enables a personalized sound space suitable for individual work places. The medium, which in our case is sound, is transparent and thus becomes an entangled part of the surrounding environment. This type of immersion is only partly occluding the users sensory access to physical reality. The purpose of using the sound space is not to become immersed by the sounds, rather to use the sounds to direct cognitive attention to get immersed in another cognitive activity.
|
|
| 5. |
- Hansen, Kjetil Falkenberg, Docent, 1972-, et al.
(author)
-
Sound design through large audience interaction
- 2019
-
In: Proceedings of the Sound and Music Computing Conferences. - Malaga : CERN. - 9788409085187 ; , s. 119-126, s. 119-126
-
Conference paper (peer-reviewed)abstract
- In collaboration with Volvo Cars, we presented a novel design tool to a large public of approximately three million people at the three leading motor shows in 2017 in Geneva, Shanghai and New York. The purpose of the tool was to explore the relevance of interactive audio-visual strategies for supporting the development of sound environments in future silent cars, i.e., a customised sonic identity that would alter the sonic ambience for the driver and by-passers. This new tool should be able to efficiently collect non-experts’ sonic preferences for different given contexts. The design process should allow for a high-level control of complex synthesised sounds. The audience interacted individually using a single-touch selection of colour from five palettes and applying it by pointing to areas in a colour-book painting showing a road scene. Each palette corresponded to a sound, and the colour nuance in the palette corresponded to certain tweaking of the sound. In effect, the user selected and altered each sound, added it to the composition, and finally would hear a mix of layered sounds based on the colouring of the scene. The installation involved large touch screens with high quality headphones. In the study presented here, we examine differences in sound preferences between two audiences and a control group, and evaluate the feasibility of the tool based on the sound designs that emerged. Copyright: © 2019 Kjetil Falkenberg Hansen et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 Unported License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
|
|
| 6. |
- Norström, Livia, 1975-, et al.
(author)
-
New Logics of Ethics in the Age of Digital Platforms : Design Fictions of Autonomous Cars
- 2019
-
In: Proceedings of the 17th European Conference on Computer-Supported Cooperative Work-Demos and Posters. - : European Society for Socially Embedded Technologies (EUSSET).
-
Conference paper (peer-reviewed)abstract
- Autonomous cars are the first major examples of Artificial Intelligence (AI) in everyday life. When cars are transformed into platforms,new service relationships emerge between car companies and the car users. These relationships generate gains and catches for both parts related to how physical and non-physical resources are exchanged in the sharing economy; how integrity is negotiated; and how responsibility is delegated when AI enables the car to take over most of the driving. With a "car as a platform approach", in this paper, we present a design fiction on ethical implications for citizens' daily lives with autonomous cars
|
|
| 7. |
- Gunnarsson, Maj, et al.
(author)
-
Less is more : When IT comes to knowledge management
- 2000
-
In: Proceedings of IRIS23.
-
Conference paper (peer-reviewed)abstract
- In this work we take the point of departure from two thoroughly conducted casestudies, in which we have analysed certain knowledge work activities with followingimplications for design. Both studies clearly stated the call for a ICT-support whensearching information and sharing knowledge. For example, sharing bookmarksamong knowledge workers seemed to be a necessary and useful activity. Importantand valuable bookmarks often mirror what kind of information and knowledgepeople are looking for, and also what kind of forum they usually go to on the worldwide web. In this work we have developed a prototype, KnowMan, which managesbookmarks on the world wide web. The prototype is designed as a user-friendly"pad-interface" in which bookmarks can be collaboratively collected, shared andfurther explored and annotated by other knowledge workers. We end up with ageneral discussion about the possible effects of using KnowMan and what furtherdesign aspects that should be considered.
|
|
| 8. |
- Lundh Snis, Ulrika, 1970-, et al.
(author)
-
Knowman : To Go Where No Man Has Gone Before
- 2000
-
In: Proceedings of WebNet World Conference on the WWW and Internet 2000. - San Antonio, Texas : Association for the Advancement of Computing in Education (AACE). ; , s. 815-816
-
Conference paper (peer-reviewed)abstract
- IT-support for knowledge work should be designed to support and enhance the human interaction and knowledge sharing processes. The specific usage of web-based technology gives the possibilities of putting crucial, knowledge-related interaction on-line. For example, sharing bookmarks among knowledge workers seems to be a necessary and useful activity. Important and valuable bookmarks often mirror what kind of information and knowledge people are looking for and also what kind of forum they use to go to on the world wide web. In this work we have developed a prototype, KNOWMAN, which manages bookmarks on the world wide web. The prototype is designed as a user-friendly "pad-interface" in which bookmarks can be collaboratively collected, shared and further annotated and explored by other knowledge workers.
|
|
| 9. |
- Cederlund, Katarina, 1968-, et al.
(author)
-
Assessing multimodal texts : practice development and support for teachers´ learning
- 2016
-
In: The 8th international conference on Modality.
-
Conference paper (other academic/artistic)abstract
- As digital technology is becoming increasingly more common in the classrooms, students are creating screen-based texts that incorporate various modalities. Previous studies show that even if students’ texts are multimodal, the complexity in multimodal text composition are not the subject to established assessment practices privileging verbal and written language. Assessment practices signal which knowledge is legitimate and influence how teaching is organized. How teachers acknowledge and assess multimodal aspects and how practice development in this matter can be supported thereby becomes important issues to address.This study was conducted within the project Digital Arenas in Literacy Practices in Early Primary School (DILS) where teachers from three classes in different schools worked together in workshops to develop literacy teaching. The empirical material consists primarily of the participating teachers written assessments of student texts, before and after an intervention where the teachers were introduced to multimodal perspectives and concepts for analyzing and assessing multimodal texts. The text material was further complemented by video-recordings of this workshop and the teachers’ discussions. Exploring how the teachers understanding of students’ multimodal texts is manifested in their assessments, the study highlights the challenges education and teachers face. The results shows how a lack of a meta-language to describe quality and development in multimodal text-composing results in assessments influenced by established assement practices focusing formal language aspects. Access to such meta-language provided more focus on the use, balance and interrelations of modalities for communicative effect. The implications of what knowledge and practices that are made available to teachers are discussed, and further research on how to implement multimodal assessment practices in the pedagogical practice is suggested.
|
|
| 10. |
- Christiansson, Anna-Karin, 1947-, et al.
(author)
-
Low order sampled-data H∞ control
- 2003
-
In: Decision and Control, 2003. Proceedings. 42nd IEEE Conference on. ; 3, s. 2308 - 2313
-
Conference paper (peer-reviewed)abstract
- A method for obtaining low order sampled-data H∞ controllers is presented. The method is mainly based on a parametric static feedback controller for a plant that is augmented with the controller dynamics. The design of a full-order controller is a convex problem, while the optimisation problem for lower order controllers is non convex. The proposed method starts with design of a full-order sampled-data controller using Riccati equations. Then this controller is reduced by an ordinary model reduction technique, and the reduced controller is used as an initial value for an iterative procedure using linear matrix inequalities (LMIs) in the search for an optimal controller. The matrix inequalities are in fact linear in either the Lyapunov matrix or the static controller matrix, why the solution to the non convex problem fundamentally is given by a bilinear matrix inequality (BMI). The order of the controller is reduced until the closed loop performance degrades too much. Simulations are shown for the control of a time delayed SISO-plant where the controller order can be reduced from 8th to 3rd order. Results are also shown from control of a MIMO-model of a jet engine where the reduction is successful from 15th to 4th order. It is argued that the non convexity is handled efficiently since the procedure uses a model reduction of the full-order controller as initial value.
|
|
