| 4361. |
- Heyn, Hans-Martin, 1987, et al.
(författare)
-
An Investigation of Challenges Encountered When Specifying Training Data and Runtime Monitors for Safety Critical ML Applications
- 2023
-
Ingår i: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). - : Springer. - 1611-3349 .- 0302-9743. ; 13975 LNCS, s. 206-222
-
Konferensbidrag (refereegranskat)abstract
- [Context and motivation] The development and operation of critical software that contains machine learning (ML) models requires diligence and established processes. Especially the training data used during the development of ML models have major influences on the later behaviour of the system. Runtime monitors are used to provide guarantees for that behaviour. [Question/problem] We see major uncertainty in how to specify training data and runtime monitoring for critical ML models and by this specifying the final functionality of the system. In this interview-based study we investigate the underlying challenges for these difficulties. [Principal ideas/results] Based on ten interviews with practitioners who develop ML models for critical applications in the automotive and telecommunication sector, we identified 17 underlying challenges in 6 challenge groups that relate to the challenge of specifying training data and runtime monitoring. [Contribution] The article provides a list of the identified underlying challenges related to the difficulties practitioners experience when specifying training data and runtime monitoring for ML models. Furthermore, interconnection between the challenges were found and based on these connections recommendation proposed to overcome the root causes for the challenges.
|
|
| 4362. |
- Heyn, Hans-Martin, 1987, et al.
(författare)
-
Automotive Perception Software Development : An Empirical Investigation into Data, Annotation, and Ecosystem Challenges
- 2023
-
Ingår i: Proceedings - 2023 IEEE/ACM 2nd International Conference on AI Engineering - Software Engineering for AI, CAIN 2023. - : Institute of Electrical and Electronics Engineers Inc.. - 9798350301137 ; , s. 13-24
-
Konferensbidrag (refereegranskat)abstract
- Software that contains machine learning algorithms is an integral part of automotive perception, for example, in driving automation systems. The development of such software, specifically the training and validation of the machine learning components, requires large annotated datasets. An industry of data and annotation services has emerged to serve the development of such data-intensive automotive software components. Wide-spread difficulties to specify data and annotation needs challenge collaborations between OEMs (Original Equipment Manufacturers) and their suppliers of software components, data, and annotations.This paper investigates the reasons for these difficulties for practitioners in the Swedish automotive industry to arrive at clear specifications for data and annotations. The results from an interview study show that a lack of effective metrics for data quality aspects, ambiguities in the way of working, unclear definitions of annotation quality, and deficits in the business ecosystems are causes for the difficulty in deriving the specifications. We provide a list of recommendations that can mitigate challenges when deriving specifications and we propose future research opportunities to overcome these challenges. Our work contributes towards the on-going research on accountability of machine learning as applied to complex software systems, especially for high-stake applications such as automated driving.
|
|
| 4363. |
- Heyn, Hans-Martin, 1987, et al.
(författare)
-
Setting AI in Context: A Case Study on Defining the Context and Operational Design Domain for Automated Driving
- 2022
-
Ingår i: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). - Cham : Springer International Publishing. - 1611-3349 .- 0302-9743. ; 13216 LNCS, s. 199-215
-
Konferensbidrag (refereegranskat)abstract
- [Context and motivation] For automated driving systems, the operational context needs to be known in order to state guarantees on performance and safety. The operational design domain (ODD) is an abstraction of the operational context, and its definition is an integral part of the system development process. [Question/problem] There are still major uncertainties in how to clearly define and document the operational context in a diverse and distributed development environment such as the automotive industry. This case study investigates the challenges with context definitions for the development of perception functions that use machine learning for automated driving. [Principal ideas/results] Based on qualitative analysis of data from semi-structured interviews, the case study shows that there is a lack of standardisation for context definitions across the industry, ambiguities in the processes that lead to deriving the ODD, missing documentation of assumptions about the operational context, and a lack of involvement of function developers in the context definition. [Contribution] The results outline challenges experienced by an automotive supplier company when defining the operational context for systems using machine learning. Furthermore, the study collected ideas for potential solutions from the perspective of practitioners.
|
|
| 4364. |
- Hezari, H., et al.
(författare)
-
The Dirichlet isospectral problem for trapezoids
- 2021
-
Ingår i: Journal of Mathematical Physics. - : AIP Publishing. - 0022-2488 .- 1089-7658. ; 62:5
-
Tidskriftsartikel (refereegranskat)abstract
- We show that non-obtuse trapezoids are uniquely determined by their Dirichlet Laplace spectrum. This extends our previous result [Hezari et al., Ann. Henri Poincare 18(12), 3759-3792 (2017)], which was only concerned with the Neumann Laplace spectrum.
|
|
| 4365. |
- Hezari, Hamid, et al.
(författare)
-
The Neumann Isospectral Problem for Trapezoids
- 2017
-
Ingår i: Annales Henri Poincare. - : Springer Science and Business Media LLC. - 1424-0637 .- 1424-0661. ; 18:12, s. 3759-3792
-
Tidskriftsartikel (refereegranskat)abstract
- We show that non-obtuse trapezoids with identical Neumann spectra are congruent up to rigid motions of the plane. The proof is based on heat trace invariants and some new wave trace invariants associated to certain diffractive billiard trajectories. We use the method of reflections to express the Dirichlet and Neumann wave kernels in terms of the wave kernel of the double polygon. Using Hillairet's trace formulas for isolated diffractive geodesics and one-parameter families of regular geodesics with geometrically diffractive boundaries for Euclidean surfaces with conical singularities (Hillairet in J Funct Anal 226(1):48-89, 2005), we obtain the new wave trace invariants for trapezoids. To handle the reflected term, we use another result of Hillairet (J Funct Anal 226(1):48-89, 2005), which gives a Fourier integral operator representation for the Keller and Friedlander parametrix (Keller in Proc Symp Appl Math 8:27-52, 1958; Friedlander in Math Proc Camb Philos Soc 90(2):335-341, 1981) of the wave propagator near regular diffractive geodesics. The reason we can only treat the Neumann case is that the wave trace is "more singular" for the Neumann case compared to the Dirichlet case. This is a new observation which is of independent interest.
|
|
| 4366. |
- Heß, Stefanie, et al.
(författare)
-
Sewage from Airplanes Exhibits High Abundance and Diversity of Antibiotic Resistance Genes
- 2019
-
Ingår i: Environmental Science & Technology. - : American Chemical Society (ACS). - 0013-936X .- 1520-5851. ; 53:23, s. 13898-13905
-
Tidskriftsartikel (refereegranskat)abstract
- Airplane sanitary facilities are shared by an international audience. We hypothesized the corresponding sewage to be an extraordinary source of antibiotic-resistant bacteria (ARB) and resistance genes (ARG) in terms of diversity and quantity. Accordingly, we analyzed ARG and ARB in airplane-borne sewage using complementary approaches: Metagenomics, quantitative polymerase chain reaction (qPCR), and cultivation. For the purpose of comparison, we also quantified ARG and ARB in the inlets of municipal treatment plants with and without connection to airports. As expected, airplane sewage contained an extraordinarily rich set of mobile ARG, and the relative abundances of genes were mostly increased compared to typical raw sewage of municipal origin. Moreover, combined resistance against third-generation cephalosporins, fluorochinolones, and aminoglycosides was unusually common (28.9%) among Escherichia coli isolated from airplane sewage. This percentage exceeds the one reported for German clinical isolates by a factor of 8. Our findings suggest that airplane-borne sewage can effectively contribute to the fast and global spread of antibiotic resistance.
|
|
| 4367. |
- Hickey, P. J., et al.
(författare)
-
Spin-crossover structural transition in the layered perovskite Gd2SrCo2O7
- 2007
-
Ingår i: Physical Review B. ; 75:2
-
Tidskriftsartikel (refereegranskat)abstract
- The structure of the double-layer Ruddlesden-Popper phase Gd2SrCo2O7 has been determined in the temperature range 300-973 K using powder neutron diffraction. The room-temperature structure is tetragonal and displays small tilts of the oxygen sublattice that are described in P4(2)/mnm symmetry instead of the idealized I4/mmm structure. A structural transition at T approximate to 580 K is characterized by reduction of unit cell symmetry and a marked axial elongation of the CoO6 octahedra; this behavior is associated with a spin crossover at the octahedrally coordinated cobalt (III) ions, most likely from an intermediate to a high spin configuration.
|
|
| 4368. |
- Hieronymus, Fredrik, 1986, et al.
(författare)
-
A mega-analysis of fixed-dose trials reveals dose-dependency and a rapid onset of action for the antidepressant effect of three selective serotonin reuptake inhibitors
- 2016
-
Ingår i: Translational Psychiatry. - : Springer Science and Business Media LLC. - 2158-3188. ; 6:6
-
Tidskriftsartikel (refereegranskat)abstract
- The possible dose-dependency for the antidepressant effect of selective serotonin reuptake inhibitors (SSRIs) remains controversial. We believe we have conducted the first comprehensive patient-level mega-analysis exploring this issue, one incentive being to address the possibility that inclusion of low-dose arms in previous meta-analyses may have caused an underestimation of the efficacy of these drugs. All company-sponsored, acute-phase, placebo-controlled, fixed-dose trials using the Hamilton Depression Rating Scale (HDRS) and conducted to evaluate the effect of citalopram, paroxetine or sertraline in adult major depression were included (11 trials, n = 2859 patients). The single-item depressed mood, which has proven a more sensitive measure to detect an antidepressant signal than the sum score of all HDRS items, was designated the primary effect parameter. Doses below or at the lower end of the usually recommended dose range (citalopram: 10-20 mg, paroxetine: 10 mg; sertraline: 50 mg) were superior to placebo but inferior to higher doses, hence confirming a dose-dependency to be at hand. In contrast, among doses above these, there was no indication of a dose-response relationship. The effect size (ES) after exclusion of suboptimal doses was of a more respectable magnitude (0.5) than that usually attributed to the antidepressant effect of the SSRIs. In conclusion, the observation that low doses are less effective than higher ones challenges the oft-cited view that the effect of the SSRIs is not dose-dependent and hence not caused by a specific, pharmacological antidepressant action. Moreover, we suggest that inclusion of suboptimal doses in previous meta-analyses has led to an underestimation of the efficacy of these drugs.
|
|
| 4369. |
- Hieronymus, Fredrik, 1986, et al.
(författare)
-
Consistent superiority of selective serotonin reuptake inhibitors over placebo in reducing depressed mood in patients with major depression.
- 2016
-
Ingår i: Molecular psychiatry. - : Springer Science and Business Media LLC. - 1476-5578 .- 1359-4184. ; 21:4, s. 523-30
-
Tidskriftsartikel (refereegranskat)abstract
- The recent questioning of the antidepressant effect of selective serotonin reuptake inhibitors (SSRIs) is partly based on the observation that approximately half of company-sponsored trials have failed to reveal a significant difference between active drug and placebo. Most of these have applied the Hamilton depression rating scale to assess symptom severity, the sum score for its 17 items (HDRS-17-sum) serving as effect parameter. In this study, we examined whether the negative outcomes of many SSRI trials may be partly caused by the use of this frequently questioned measure of response. We undertook patient-level post-hoc analyses of 18 industry-sponsored placebo-controlled trials regarding paroxetine, citalopram, sertraline or fluoxetine, and including in total 6669 adults with major depression, the aim being to assess what the outcome would have been if the single item depressed mood (rated 0-4) had been used as a measure of efficacy. In total, 32 drug-placebo comparisons were reassessed. While 18 out of 32 comparisons (56%) failed to separate active drug from placebo at week 6 with respect to reduction in HDRS-17-sum, only 3 out of 32 comparisons (9%) were negative when depressed mood was used as an effect parameter (P<0.001). The observation that 29 out of 32 comparisons detected an antidepressant signal from the tested SSRI suggests the effect of these drugs to be more consistent across trials than previously assumed. Further, the frequent use of the HDRS-17-sum as an effect parameter may have distorted the current view on the usefulness of SSRIs and hampered the development of novel antidepressants.Molecular Psychiatry advance online publication, 28 April 2015; doi:10.1038/mp.2015.53.
|
|
| 4370. |
- Hieronymus, Fredrik, 1986, et al.
(författare)
-
Efficacy of selective serotonin reuptake inhibitors in the absence of side effects: a mega-analysis of citalopram and paroxetine in adult depression
- 2018
-
Ingår i: Molecular Psychiatry. - : Springer Science and Business Media LLC. - 1476-5578 .- 1359-4184. ; 23:8, s. 1731-1736
-
Tidskriftsartikel (refereegranskat)abstract
- It has been suggested that the superiority of antidepressants over placebo in controlled trials is merely a consequence of side effects enhancing the expectation of improvement by making the patient realize that he/she is not on placebo. We explored this hypothesis in a patient-level post hoc-analysis including all industry-sponsored, Food and Drug Administration-registered placebo-controlled trials of citalopram or paroxetine in adult major depression that used the Hamilton Depression Rating Scale (HDRS) and included a week 6 symptom assessment (n=15). The primary analyses, which compared completers on active treatment without early adverse events to completers on placebo (with or without adverse events) with respect to reduction in the HDRS depressed mood item showed larger symptom reduction in patients given active treatment, the effect sizes being 0.48 for citalopram and 0.33 for paroxetine. In actively treated subjects reporting early adverse events, who also outperformed those given placebo, the severity of the adverse events did not predict response. Several sensitivity analyses, for example, including (i) those using change of the sum of all HDRS-17 items as effect parameter, (ii) those excluding all subjects with adverse events (that is, also those on placebo) and (iii) those based on the intention-to-treat population, were all in line with the primary analyses. The finding that both paroxetine and citalopram are clearly superior to placebo also when not producing adverse events, as well as the lack of association between adverse event severity and response, argue against the theory that antidepressants outperform placebo solely or largely because of their side effects.
|
|
