Chalmers Research
Not a member yet
88095 research outputs found
Sort by
Excess Mortality and the Effect of the Covid-19 Vaccines. Part 1: European Data.
Get PDFUsing publicly available data for 28 EU/EES countries from Eurostat and Our World in Data, we investigate how the current rate of Covid vaccination in acountry compares to its average rate of excess mortality (EM) in the pandemic to date. We find that, in the linear regression, the correlation betweenaverage EM and vaccination rate is strongly negative, a priori evidence to support the claim that the Covid vaccines have saved many lives. However, acloser analysis of the timeline suggests otherwise. The correlation was already strongly negative before the vaccines were rolled out and is onlyweakly negative thereafter. In theory, survivor bias could still explain this shift, especially since waves of EM closely align with Covid waves. However,we find in addition that about half of our 28 countries experienced higher EM in 2022 than in 2021, and all that did so have higher than average vaccination rates. This is something which survivor bias cannot explain and raises the real possibility that the vaccines have not just failed to savemany lives, but may have already caused net harm. Moreover, any such harm may be ongoing since we find that EM and vaccination rates have been consistently positively correlated since April 2022. We show that all these findings are robust to several different ways of measuring EM and/or vaccination rates. Finally, using public data from Worldometers, we show that the correlation over time of official Covid mortality rates with current vaccination rates closely tracks that of EM rates, even as Covid mortality has waxed and waned and even in the post-omicron period
An Empirical Study of Multitask Learning to Improve Open Domain Dialogue Systems
No full textAutoregressive models used to generate responses in open-domain dialogue systems often struggle to take long-term context into account and to maintain consistency over a dialogue. Previous research in open-domain dialogue generation has shown that the use of auxiliary tasks can introduce inductive biases that encourage the model to improve these qualities. However, most previous research has focused on encoder-only or encoder/decoder models, while the use of auxiliary tasks in encoder-only autoregressive models is under-explored. This paper describes an investigation where four different auxiliary tasks are added to small and medium-sized GPT-2 models fine-tuned on the PersonaChat and DailyDialog datasets. The results show that the introduction of the new auxiliary tasks leads to small but consistent improvement in evaluations of the investigated models
Nonlinear Model Inversion-Based Output Tracking Control for Battery Fast Charging
Get PDFWe propose a novel nonlinear control approach for fast charging of lithium-ion batteries, where health- and safety-related variables, or their time derivatives, are expressed in an input-polynomial form. By converting a constrained optimal control problem into an output tracking problem with multiple tracking references, the required control input, i.e., the charging current, is obtained by computing a series of candidate currents associated with different tracking references. Consequently, an optimization-free nonlinear model inversion-based control algorithm is derived for charging the batteries. We demonstrate the efficacy of our method using a spatially discretized high-fidelity pseudo-two-dimensional (P2D) model with thermal dynamics. Conventional methods require computationally demanding optimization to solve the corresponding fast charging problem for such a high-order system, leading to practical difficulties in achieving low-cost implementation. Results from comparative studies show that the proposed controller can achieve performance very close to nonlinear and linearized model predictive control but with much lower computational costs and minimal parameter tuning efforts
Semiconductor nanowires for biosening
No full textSemiconductor nanowires are known to enhance the signal of fluorophores in their proximity. In our recent work, we optimized GaP nanowires to maximize the enhancement of fluorescence excitation. For that, we used optical microscopy to measure the fluorescence photobleaching rate on the nanowires, proportional to local excitation. From the measurements and modelling, we show for red fluorophores (excitation wavelength 640 nm) that nanowires with diameters of 90–130 nm enable enhancement of over a factor of 5 compared to bulk solution. We also demonstrate that a 10 nm oxide layer on the nanowires enables functionalisation using biotin-streptavidin chemistry without hindering the enhancement. Such nanowires were used in our laboratory to detect fluorescently-labelled proteins on a single molecule level. Now we employ these results for detection of down to nanomolar concentrations of DNA in solution, and in our poster, we discuss these biosensors
High Gain Linearly Polarized 1D Beam Scanning Array Antenna based on Gap Waveguide Technology
No full textThis paper presents a feed network for a high gain array antenna capable of 1D beam scanning up to 60 degrees without grating lobes. The feed network uses a layered design where the use of gap waveguide technology enables the layers to be assembled without good electrical contact. Each layer constitute one section of a vertical corporate feed network ending in a horizontal series feed, allowing for a compact design and integration of electronic components inside the antenna. To enable this, a periodic pin structure has been designed together with two vertical power dividers, one in ridge gap waveguide and one from microstrip to ridge gap waveguide. The structure is designed for Ka-band at 27.5-30 GHz and results show a 10 dB return loss bandwidth of 10 %. The array size is 40x8 elements which fits in a volume of 59.2x242.9x13.1 mm, while the boresight realized gain is 28 dBi
A multiscale deep learning model for elastic properties of woven composites
Get PDFTime-consuming and costly computational analysis expresses the need for new methods for generalizing multiscale analysis of composite materials. Combining neural networks and multiscale modeling is favorable for bypassing expensive lower-scale material modeling, and accelerating coupled multi-scale analyses (FE2). In this work, neural networks are used to replace the time-consuming micromechanical finite element analysis of unidirectional composites, representing the local material properties of yarns in woven fabric composites in a multiscale framework. Leveraging the fast multiscale data generation procedure, we presented a second neural networks model to estimate the elastic engineering coefficients of a particular weave architecture based on a broad range of dry resin and fiber properties and yarn fiber volume fraction. As an outcome, this paper provides the user with a generalized, neural network-based approach to tackle the balance of computational efficiency and accuracy in the multiscale analysis of elastic woven composites
Age of Information in Slotted ALOHA With Energy Harvesting
Get PDFWe examine the age of information (AoI) of a status update system that incorporates energy harvesting and uses the slotted ALOHA protocol. We derive analytically the average AoI and the probability that the AoI exceeds a given threshold. Via numerical results, we investigate two strategies to minimize the age of information (AoI): transmitting a new update whenever possible to exploit every chance to reduce the AoI, and transmitting only when sufficient energy is available to increase the chance of successful delivery. The two strategies are beneficial for low and high update generation rates, respectively. However, an optimized approach that balances the two strategies outperforms them significantly in terms of both AoI and throughput
A literature review on thermodynamic sorption models of radionuclides with some selected granitic minerals
No full textThe literature for radionuclide sorption on four common granitic minerals have been surveyed. Mainly, such studies were modelling using Thermodynamic Sorption Models were investigated. Although the studies give a far from concerted results, they agree on the necessity to model radionuclide uptake by granitic minerals with a combination of ion exchange and surface complexation reactions. For the sheet-silicates biotite and chlorite alkaline and alkaline earth mainly bind by ion exchange but there is also a clear pH effect for this, which shows the importance of protons competing with metal cations for the exchange sites. For multivalent metal cations, surface complexation is the model of choice since the binding to mineral surfaces seems to be strongly dependent on pH and to be little affected by an increase in ion strength. Anion sorption seems to be taking place also by surface complexation, where the sorption mainly takes place at low pH. For the feldspar minerals K-feldspar and plagioclase the sorption is also modelled by the two reaction mechanisms ion exchange and surface complexation. Surface complexation seems to be especially prevalent for the M(III) and M(VI) state, while ion exchange probably dominates M (II) uptake. Although the literature on these minerals is sparse, the studies show that also these minerals have considerable sorption capacity and must be considered if sorption onto granite is to be modelled from single mineral data. What is usually missing from these studies are more systematic variations in pH, ion strength and temperature. Instead, there is a certain overemphasis on the establishment of sorption isotherms
6.7 GHz CH3OH masers polarization in massive star-forming regions: The Flux-Limited Sample
Get PDFThe formation process of high-mass stars (M > 8Msun) is still unclear; this is mainly due to their fast evolution and large distances that make difficult to observe them in details. The observational and theoretical efforts made in the last decades have shown that a common and essential component in the formation of high-mass stars is the presence of molecular outflows during the protostellar phase, similarly to what is observed during the formation of low-mass stars. Theoretically, it has been convincingly demonstrated that the magnetic field plays an important role in launching and shaping molecular outflows in massive young stellar objects (YSOs). Therefore, providing new observational measurements of magnetic fields close (10s-100s au) to massive YSOs is of great importance. More than 10 years ago we started a large EVN campaign to measure the magnetic field orientation and strength toward a sample of 30 massive star-forming regions, called the "Flux-Limited sample", by observing the polarized emission of 6.7 GHz CH3OH masers. Here, we present a summary of the final statistics of the Flux-Limited sample, extensively reported in [1], which are focused on the relative orientation of the outflows with the magnetic fields and on the polarized characteristics of 6.7 GHz CH3OH masers