Intrapreneurship and technological innovation in optimizing qualitative research as evidenced at Infectious Diseases Institute, Uganda

Background Discrepancies between what is transcribed and the actual interview recordings were noticed in qualitative research reports. This study aimed at the development of a new transcription software (Jiegnote), and the evaluation of its effectiveness in the optimization of the transcription process, to minimize transcription completion time, and errors in qualitative research. Methods The study was a mixed methods project implemented from September to November 2020. The qualitative aspect of the study was phenomenological in perspective whereas the quantitative consisted of a randomized controlled trial (RCT) with a parallel design. Results At the time of the study, the Jiegnote software was a working prototype. We enrolled a total of 26 participants; 14 participants had their data analyzed in the RCT part of the study, 13 participated in the in-depth interviews, and 22 in the answering of Semi Structured Questionnaires. Upon the execution of an independent t test, results showed that, there was no statistical significance between the intervention and control means. On considering the total average transcription completion time and the type of language in which an audio case was recorded, the effect size evaluation implied that the Jiegnote software had a small impact (Hedges' g = 0.413438) in reducing the total average time taken to translate and transcribe audio cases that were recorded in a local language (Luganda), and a large impact (Hedges' g = 1.190919) in reducing the total average time taken to transcribe audio cases that were recorded in a foreign language (English). On considering the total average number of transcription errors and the type of language in which an audio case is recorded, the effect size evaluation implied that the Jiegnote software had a small impact (Hedges' g = 0.213258) in reducing the total average time taken to translate and transcribe audio cases that were recorded in a local language (Luganda). This was further observed (Hedges' g = 0.039928) in the transcription of cases that were recorded in a foreign language (English). On considering the in-depth interview data outcomes, participants responded that the Jiegnote software media looping functions (algorithm) enabled them to accomplish their transcription tasks in a shorter time and with fewer errors compared to the traditional methods. Conclusion The study demonstrates utilities associated with intrapreneurship and technological innovation in an organization setting whereby, the Jiegnote technology that was developed by the researchers, had some impact on the optimization of the qualitative research value chain. This was observed through the effect size (impact) evaluations that were conducted to investigate the superiority of the Jiegnote software against the traditional transcription methods, in minimizing the average number of errors committed, and time taken to complete a transcription process.

Towards Open-Set Text Recognition via Label-to-Prototype Learning

<div><br></div><div>Scene text recognition is a popular topic and can benefit various tasks. Although many methods have been proposed for the close-set text recognition challenges, they cannot be directly applied to open-set scenarios, where the evaluation set contains novel characters not appearing in the training set. Conventional methods require collecting new data and retraining the model to handle these novel characters, which is an expensive and tedious process. In this paper, we propose a label-to-prototype learning framework to handle novel characters without retraining the model. In the proposed framework, novel characters are effectively mapped to their corresponding prototypes with a label-to-prototype learning module. This module is trained on characters with seen labels and can be easily generalized to novel characters. Additionally, feature-level rectification is conducted via topology-preserving transformation, resulting in better alignments between visual features and constructed prototypes while having a reasonably small impact on model speed. A lot of experiments show that our method achieves promising performance on a variety of zero-shot, close-set, and open-set text recognition datasets.</div>

Towards Open-Set Text Recognition via Label-to-Prototype Learning

<div><br></div><div>Scene text recognition is a popular topic and can benefit various tasks. Although many methods have been proposed for the close-set text recognition challenges, they cannot be directly applied to open-set scenarios, where the evaluation set contains novel characters not appearing in the training set. Conventional methods require collecting new data and retraining the model to handle these novel characters, which is an expensive and tedious process. In this paper, we propose a label-to-prototype learning framework to handle novel characters without retraining the model. In the proposed framework, novel characters are effectively mapped to their corresponding prototypes with a label-to-prototype learning module. This module is trained on characters with seen labels and can be easily generalized to novel characters. Additionally, feature-level rectification is conducted via topology-preserving transformation, resulting in better alignments between visual features and constructed prototypes while having a reasonably small impact on model speed. A lot of experiments show that our method achieves promising performance on a variety of zero-shot, close-set, and open-set text recognition datasets.</div>

The Effect of Covid-19 Pandemic on Large-cap Stocks in Malaysia

Covid-19 that emerged in Wuhan, China and spread to Malaysia starting from 25th January 2020 has changed people’s lives and impacted the world’s economy, including the stock markets. The study investigates the impact of the Covid-19 pandemic on the stock returns in Malaysia by using a sample of thirty (30) constituents of FBM KLCI. The study utilises Malaysia’s daily Covid-19 new cases, death cases, cumulative cases, and cumulative death cases, as well as Singapore new cases and death cases. The impact is observed from 31st December 2019 until 9th June 2020 using the panel regression model. The results show a significant positive but small impact of Covid-19 variables on the stocks’ returns except for Singapore daily cases and death cases, which were negative. The study also identifies that the Malaysian stock market is more sensitive to Covid-19 local death cases during the pandemic.

Ventspils apkārtnes apdzīvojums 17.–18. gadsimtā

The article deals with questions relating to the settlement in the area of Ventspils. Attention is gi-ven to environmental factors that could have had a greater or lesser influence on the settlement structure. The stream bank erosion along the River Venta had a relatively small impact on populated areas, while the wind erosion (sand deposition) caused the individual farms as well as villages to be abandoned. The Great Plague epidemic of 1710 was particularly devastating in the Ventspils area, during which about 40% of farms disappeared and most of them were not restored until the end of the 18th century. Final abandon-ment of farms, their renovation or addition of the former farmland to the land of the manor caused chan-ges in the population structure.

Vaccine and COVID-19 Trajectories: Equal vaccine rates do not reduce inequality in COVID-19 rates

Researchers and public health officials posit that vaccine equity holds the key to ending the pandemic. Yet, most prior work on vaccine equity focuses on vaccine hesitancy and seldom compares the vaccine trajectories of neighborhoods with varying COVID-19 levels. Notably scarce are also studies that examine the extent to which vaccination helps reduce inequalities in the prevalence of COVID-19. Using administrative data from the City of Toronto, we compare the vaccine trajectories of neighborhoods with low, moderate, and high COVID-19 rates. We also examine whether disparities in COVID-19 rates by a neighborhood’s COVID-19 rates as vaccinations have increased. By mid-June 2021, differences in vaccination rates by the neighborhoods’ COVID-19 levels are small. The vaccination rollout has only had a small impact on disparities in COVID-19 rates across neighborhoods. Equality in vaccination rates is by no means a silver bullet to reduce inequalities in COVID-19 infections across neighborhoods with varying socio-demographic characteristics.

Detection of Small Impact Craters via Semantic Segmenting Lunar Point Clouds Using Deep Learning Network

Impact craters refer to the most salient features on the moon surface. They are of huge significance for analyzing the moon topography, selecting the lunar landing site and other lunar exploration missions, etc. However, existing methods of impact crater detection have been largely implemented on the optical image data, thereby causing them to be sensitive to the sunlight. Thus, these methods can easily achieve unsatisfactory detection results. In this study, an original two-stage small crater detection method is proposed, which is sufficiently effective in addressing the sunlight effects. At the first stage of the proposed method, a semantic segmentation is conducted to detect small impact craters by fully exploiting the elevation information in the digital elevation map (DEM) data. Subsequently, at the second stage, the detection accuracy is improved under the special post-processing. As opposed to other methods based on DEM images, the proposed method, respectively, increases the new crusher percentage, recall and crusher level F1 by 4.89%, 5.42% and 0.67%.

An Evolving Understanding of Enigmatic Large Ripples on Mars

&lt;p&gt;Unlike terrestrial sandy deserts, Mars hosts two scales of ripples in fine sand. Larger, meter-scale ripples are morphologically distinct from small, decimeter-scale ripples, and their size, in particular, decreases with increasing atmospheric density. As a result, it was recently proposed that the equilibrium size of the larger ripples is set by an aerodynamic process, which makes them larger under thinner atmospheres. Under this hypothesis, large martian ripples would be distinct from smaller, decimeter-scale impact ripples in a mechanistic sense. Several workers have followed up on these initial observations to either corroborate, counter, or expand upon that hypothesis. Notably, a mechanistic model that not only corroborates the hypothesis that the size of large martian ripples is set by an aerodynamic process but also suggests that they arise from an aerodynamic instability, distinct from the grain-impact instability thought to be responsible for the formation of impact ripples, was developed. Conversely, other workers proposed that large ripples can develop from small impact ripples in a numerical model due to Mars&amp;#8217; low atmospheric pressure. In the latter model, the ripples&amp;#8217; growth-limiting mechanism is consistent with an aerodynamic process, but the large ripples would not be a separate class of ripples &amp;#8211; they would simply be a larger version of the small impact ripples. Here, we explore this debate by synthesizing recent advances in large-ripple formation and offer potential avenues to address outstanding questions. Although significant knowledge gaps remain, it is clear that large martian ripples are larger where the atmosphere is less dense. The size of large martian ripples thus remain a powerful paleoclimate indicator.&lt;/p&gt;

The Competitiveness Issue of the Japanese Economy Under Carbon Pricing: A Computable General Equilibrium Analysis of 2050

Using a computable general equilibrium (CGE) model, this paper investigates the impact of carbon regulations on the Japanese economy. We use an 11-sector, 15-region global dynamic CGE model with a time span from 2011 to 2050. We assume that Japan (along with other developed regions) reduces CO2 emissions by 80% by 2050 and analyze the impact on the Japanese economy. In particular, we consider multiple scenarios of CO2 reduction rates in less developed regions and analyze how changes in CO2 reduction in these regions affect Japan. In addition, we also consider multiple scenarios of the use of a border adjustment policy and analyze its impact. Our simulation results are summarized as follows. First, an 80% CO2 reduction in Japan generates large negative impacts on the Japanese economy in terms of both the macroeconomy and individual sectors. Second, changes in the reduction rates in less developed regions have only a small impact on Japan. Third, the use of border adjustment in Japan has a small impact on the GDP and welfare of Japan overall but a large impact on output in the energy intensive sectors. When future climate change policies in Japan are discussed, much attention is usually paid to climate policy in less developed regions. However, the second result of our analysis suggests that climate change policy in less developed regions has only a small impact on Japan. In addition, the third result indicates that the effectiveness of border adjustment is limited.