Search

Climate change is increasing the frequency and severity of short-term (~1 y) drought events—the most common duration of drought—globally. Yet the impact of this intensification of drought on ecosystem functioning remains poorly resolved. This is due in part to the widely disparate approaches ecologists have employed to study drought, variation in the severity and duration of drought studied, and differences among ecosystems in vegetation, edaphic and climatic attributes that can mediate drought impacts. To overcome these problems and better identify the factors that modulate drought responses, we used a coordinated distributed experiment to quantify the impact of short-term drought on grassland and shrubland ecosystems. With a standardized approach, we imposed ~a single year of drought at 100 sites on six continents. Here we show that loss of a foundational ecosystem function—aboveground net primary production (ANPP)—was 60% greater at sites that experienced statistically extreme drought (1-in-100-y event) vs. those sites where drought was nominal (historically more common) in magnitude (35% vs. 21%, respectively). This reduction in a key carbon cycle process with a single year of extreme drought greatly exceeds previously reported losses for grasslands and shrublands. Our global experiment also revealed high variability in drought response but that relative reductions in ANPP were greater in drier ecosystems and those with fewer plant species. Overall, our results demonstrate with unprecedented rigor that the global impacts of projected increases in drought severity have been significantly underestimated and that drier and less diverse sites are likely to be most vulnerable to extreme drought.

Climate change is increasing the frequency and severity of short-term (~1 y) drought events—the most common duration of drought—globally. Yet the impact of this intensification of drought on ecosystem functioning remains poorly resolved. This is due in part to the widely disparate approaches ecologists have employed to study drought, variation in the severity and duration of drought studied, and differences among ecosystems in vegetation, edaphic and climatic attributes that can mediate drought impacts. To overcome these problems and better identify the factors that modulate drought responses, we used a coordinated distributed experiment to quantify the impact of short-term drought on grassland and shrubland ecosystems. With a standardized approach, we imposed ~a single year of drought at 100 sites on six continents. Here we show that loss of a foundational ecosystem function—aboveground net primary production (ANPP)—was 60% greater at sites that experienced statistically extreme drought (1-in-100-y event) vs. those sites where drought was nominal (historically more common) in magnitude (35% vs. 21%, respectively). This reduction in a key carbon cycle process with a single year of extreme drought greatly exceeds previously reported losses for grasslands and shrublands. Our global experiment also revealed high variability in drought response but that relative reductions in ANPP were greater in drier ecosystems and those with fewer plant species. Overall, our results demonstrate with unprecedented rigor that the global impacts of projected increases in drought severity have been significantly underestimated and that drier and less diverse sites are likely to be most vulnerable to extreme drought., Climate change is increasing the frequency and severity of short-term (~1 y) drought events—the most common duration of drought—globally. Yet the impact of this intensification of drought on ecosystem functioning remains poorly resolved. This is due in part to the widely disparate approaches ecologists have employed to study drought, variation in the severity and duration of drought studied, and differences among ecosystems in vegetation, edaphic and climatic attributes that can mediate drought impacts. To overcome these problems and better identify the factors that modulate drought responses, we used a coordinated distributed experiment to quantify the impact of short-term drought on grassland and shrubland ecosystems. With a standardized approach, we imposed ~a single year of drought at 100 sites on six continents. Here we show that loss of a foundational ecosystem function—aboveground net primary production (ANPP)—was 60% greater at sites that experienced statistically extreme drought (1-in-100-y event) vs. those sites where drought was nominal (historically more common) in magnitude (35% vs. 21%, respectively). This reduction in a key carbon cycle process with a single year of extreme drought greatly exceeds previously reported losses for grasslands and shrublands. Our global experiment also revealed high variability in drought response but that relative reductions in ANPP were greater in drier ecosystems and those with fewer plant species. Overall, our results demonstrate with unprecedented rigor that the global impacts of projected increases in drought severity have been significantly underestimated and that drier and less diverse sites are likely to be most vulnerable to extreme drought.

Voice faking, driven primarily by recent advances in text-to-speech (TTS) synthesis technology, poses significant societal challenges. Currently, the prevailing assumption is that unaltered human speech can be considered genuine, while fake speech comes from TTS synthesis. We argue that this binary distinction is oversimplified. For instance, altered playback speeds can be used for malicious purposes, like in the 'Drunken Nancy Pelosi' incident. Similarly, editing of audio clips can be done ethically, e.g., for brevity or summarization in news reporting or podcasts, but editing can also create misleading narratives. In this paper, we propose a conceptual shift away from the binary paradigm of audio being either 'fake' or 'real'. Instead, our focus is on pinpointing 'voice edits', which encompass traditional modifications like filters and cuts, as well as TTS synthesis and VC systems. We delineate 6 categories and curate a new challenge dataset rooted in the M-AILABS corpus, for which we present baseline detection systems. And most importantly, we argue that merely categorizing audio as fake or real is a dangerous over-simplification that will fail to move the field of speech technology forward.

Most recent speech privacy efforts have focused on anonymizing acoustic speaker attributes but there has not been as much research into protecting information from speech content. We introduce a toy problem that explores an emerging type of privacy called "content masking" which conceals selected words and phrases in speech. In our efforts to define this problem space, we evaluate an introductory baseline masking technique based on modifying sequences of discrete phone representations (phone codes) produced from a pre-trained vector-quantized variational autoencoder (VQ-VAE) and re-synthesized using WaveRNN. We investigate three different masking locations and three types of masking strategies: noise substitution, word deletion, and phone sequence reversal. Our work attempts to characterize how masking affects two downstream tasks: automatic speech recognition (ASR) and automatic speaker verification (ASV). We observe how the different masks types and locations impact these downstream tasks and discuss how these issues may influence privacy goals., Comment: Accepted to ITG Speech Conference 2023

Dispersal is a central life history trait that affects the ecological and evolutionary dynamics of populations and communities. The recent use of experimental evolution for the study of dispersal is a promising avenue for demonstrating valuable proofs of concept, bringing insight into alternative dispersal strategies and trade-offs, and testing the repeatability of evolutionary outcomes. Practical constraints restrict experimental evolution studies of dispersal to a set of typically small, short-lived organisms reared in artificial laboratory conditions. Here, we argue that despite these restrictions, inferences from these studies can reinforce links between theoretical predictions and empirical observations and advance our understanding of the eco-evolutionary consequences of dispersal. We illustrate how applying an integrative framework of theory, experimental evolution and natural systems can improve our understanding of dispersal evolution under more complex and realistic biological scenarios, such as the role of biotic interactions and complex dispersal syndromes.

This paper summarizes the structure and findings from the first Workshop on Troubles and Failures in Conversations between Humans and Robots. The workshop was organized to bring together a small, interdisciplinary group of researchers working on miscommunication from two complementary perspectives. One group of technology-oriented researchers was made up of roboticists, Human-Robot Interaction (HRI) researchers and dialogue system experts. The second group involved experts from conversation analysis, cognitive science, and linguistics. Uniting both groups of researchers is the belief that communication failures between humans and machines need to be taken seriously and that a systematic analysis of such failures may open fruitful avenues in research beyond current practices to improve such systems, including both speech-centric and multimodal interfaces. This workshop represents a starting point for this endeavour. The aim of the workshop was threefold: Firstly, to establish an interdisciplinary network of researchers that share a common interest in investigating communicative failures with a particular view towards robotic speech interfaces; secondly, to gain a partial overview of the “failure landscape” as experienced by roboticists and HRI researchers; and thirdly, to determine the potential for creating a robotic benchmark scenario for testing future speech interfaces with respect to the identified failures. The present article summarizes both the “failure landscape” surveyed during the workshop as well as the outcomes of the attempt to define a benchmark scenario., QC 20231220

The validity of brain monitoring using electroencephalography (EEG), particularly to guide care in patients with acute or critical illness, requires that experts can reliably identify seizures and other potentially harmful rhythmic and periodic brain activity, collectively referred to as "ictal-interictal-injury continuum" (IIIC). Previous interrater reliability (IRR) studies are limited by small samples and selection bias. This study was conducted to assess the reliability of experts in identifying IIIC., SCOPUS: ar.j, info:eu-repo/semantics/published

Privacy in speech and audio has many facets. A particularly under-developed area of privacy in this domain involves consideration for information related to content and context. Speech content can include words and their meaning or even stylistic markers, pathological speech, intonation patterns, or emotion. More generally, audio captured in-the-wild may contain background speech or reveal contextual information such as markers of location, room characteristics, paralinguistic sounds, or other audible events. Audio recording devices and speech technologies are becoming increasingly commonplace in everyday life. At the same time, commercialised speech and audio technologies do not provide consumers with a range of privacy choices. Even where privacy is regulated or protected by law, technical solutions to privacy assurance and enforcement fall short. This position paper introduces three important and timely research challenges for content privacy in speech and audio. We highlight current gaps and opportunities, and identify focus areas, that could have significant implications for developing ethical and safer speech technologies., Comment: Accepted for publication in ISCA SPSC Symposium 2022

Dispersal is a central life history trait that affects the ecological and evolutionary dynamics of populations and communities. The recent use of experimental evolution for the study of dispersal is a promising avenue for demonstrating valuable proofs of concept, bringing insight into alternative dispersal strategies and trade‐offs, and testing the repeatability of evolutionary outcomes. Practical constraints restrict experimental evolution studies of dispersal to a set of typically small, short‐lived organisms reared in artificial laboratory conditions. Here, we argue that despite these restrictions, inferences from these studies can reinforce links between theoretical predictions and empirical observations and advance our understanding of the eco‐evolutionary consequences of dispersal. We illustrate how applying an integrative framework of theory, experimental evolution and natural systems can improve our understanding of dispersal evolution under more complex and realistic biological scenarios, such as the role of biotic interactions and complex dispersal syndromes.

Hormonal changes around ovulation divide the menstrual cycle (MC) into the follicular and luteal phases. In addition, oral contraceptives (OCs) have active (higher hormone) and placebo phases. Although there are some MC-based effects on various physiological outcomes, we found these differences relatively subtle and difficult to attribute to specific hormones, as estrogen and progesterone fluctuate rather than operating in a complete on/off pattern as observed in cellular or preclinical models often used to substantiate human data. A broad review reveals that the differences between the follicular and luteal phases and between OC active and placebo phases are not associated with marked differences in exercise performance and appear unlikely to influence muscular hypertrophy in response to resistance exercise training. A systematic review and meta-analysis of substrate oxidation between MC phases revealed no difference between phases in the relative carbohydrate and fat oxidation at rest and during acute aerobic exercise. Vascular differences between MC phases are also relatively small or nonexistent. Although OCs can vary in composition and androgenicity, we acknowledge that much more work remains to be done in this area; however, based on what little evidence is currently available, we do not find compelling support for the notion that OC use significantly influences exercise performance, substrate oxidation, or hypertrophy. It is important to note that the study of females requires better methodological control in many areas. Previous studies lacking such rigor have contributed to premature or incorrect conclusions regarding the effects of the MC and systemic hormones on outcomes. While we acknowledge that the evidence in certain research areas is limited, the consensus view is that the impact of the MC and OC use on various aspects of physiology is small or nonexistent.

Machine-learning (ML) shortcuts or spurious correlations are artifacts in datasets that lead to very good training and test performance but severely limit the model's generalization capability. Such shortcuts are insidious because they go unnoticed due to good in-domain test performance. In this paper, we explore the influence of different shortcuts and show that even simple shortcuts are difficult to detect by explainable AI methods. We then exploit this fact and design an approach to defend online databases against crawlers: providers such as dating platforms, clothing manufacturers, or used car dealers have to deal with a professionalized crawling industry that grabs and resells data points on a large scale. We show that a deterrent can be created by deliberately adding ML shortcuts. Such augmented datasets are then unusable for ML use cases, which deters crawlers and the unauthorized use of data from the internet. Using real-world data from three use cases, we show that the proposed approach renders such collected data unusable, while the shortcut is at the same time difficult to notice in human perception. Thus, our proposed approach can serve as a proactive protection against illegitimate data crawling., Comment: Published at BMVC 2023

Personal relative deprivation (PRD) is the feeling of resentment one gets from perceived inequality or unfairness. This feeling is argued to be a precursor to risk-taking as well as a causal factor in the development of intensive gambling and problem gambling (PG). However, the evidence supporting the relationship between PG and PRD has been cross-sectional or laboratory based. The present research reinvestigated this relationship using a representative online sample of Canadian gamblers (n = 4,400, which included several hundred problem gamblers) followed over a one-year period (Baseline and Follow-Up). A series of multiple regressions endeavoured to determine PRD’s causal relationship to PG, even though PRD was only administered at Follow-Up. The first multiple regression confirmed PRD to be one of the strongest cross-sectional predictors of problem gambling. However, the second cross-sectional multiple regression found no significant relationship between PRD and measures of gambling intensity (i.e., # formats engaged in; gambling frequency), which is theoretically problematic considering that intensive gambling involvement is the immediate antecedent to PG. The final multiple regression found that PG at Baseline was one of the strongest predictors of PRD at Follow-Up. Taken together, the present results suggest that the robust cross-sectional association between PG and PRD is in large part due to PG leading to PRD, rather than PRD leading to PG.

Impact of increasingly prevalent smart technologies We think that these changes will result in better or worse futures will be determined in large part by what is prioritised in technological innovation, and to what extent (if at all), previously marginalised, silenced, and exploited voices become part of the conversation. Our major concern is that innovation for profit is incentivised over societal benefits. We recommend that the government incentivises innovation for societal benefit. Impact of smart technologies on different groups Increased risks, especially for the most vulnerable and marginalised in society Smart technologies can improve the lives of vulnerable people but at the same time pose more significant risks if not appropriately managed. Therefore, we advocate for more inclusive development of smart devices at every stage of the product lifecycle (e.g. design, testing, implementation and maintenance). For example, datasets used for training, testing, and validating smart technologies are developed for common use cases; taking into account underprivileged and diverse parts of the society is key to their successful and fair adoption. While some technological solutions can be developed especially for vulnerable groups, a recent example with eye implants showed there is no protection once a company decides to end its production. We also advocate for more research with diverse groups. Smart technologies are mainly developed for commercial purposes, meaning that they do not consider different people’s needs and appetite for adoption. Digital literacies and skills The government needs to take responsibility for equipping society with the skills needed to use smart technologies in a safe and efficient way (eg to manage their privacy settings or set up secure passwords). Essential Digital Skills Report 2021 showed that c.10 million (19%) of UK adults do not have fundamental digital skills (for example, be able to use a device, connect to a Wi-Fi network and cre

Machine learning is a data-driven field, and the quality of the underlying datasets plays a crucial role in learning success. However, high performance on held-out test data does not necessarily indicate that a model generalizes or learns anything meaningful. This is often due to the existence of machine learning shortcuts - features in the data that are predictive but unrelated to the problem at hand. To address this issue for datasets where the shortcuts are smaller and more localized than true features, we propose a novel approach to detect and remove them. We use an adversarially trained lens to detect and eliminate highly predictive but semantically unconnected clues in images. In our experiments on both synthetic and real-world data, we show that our proposed approach reliably identifies and neutralizes such shortcuts without causing degradation of model performance on clean data. We believe that our approach can lead to more meaningful and generalizable machine learning models, especially in scenarios where the quality of the underlying datasets is crucial., Comment: Accepted at XAI4CV @ CVPR2023

Deepfakes are synthetically generated media often devised with malicious intent. They have become increasingly more convincing with large training datasets advanced neural networks. These fakes are readily being misused for slander, misinformation and fraud. For this reason, intensive research for developing countermeasures is also expanding. However, recent work is almost exclusively limited to deepfake detection - predicting if audio is real or fake. This is despite the fact that attribution (who created which fake?) is an essential building block of a larger defense strategy, as practiced in the field of cybersecurity for a long time. This paper considers the problem of deepfake attacker attribution in the domain of audio. We present several methods for creating attacker signatures using low-level acoustic descriptors and machine learning embeddings. We show that speech signal features are inadequate for characterizing attacker signatures. However, we also demonstrate that embeddings from a recurrent neural network can successfully characterize attacks from both known and unknown attackers. Our attack signature embeddings result in distinct clusters, both for seen and unseen audio deepfakes. We show that these embeddings can be used in downstream-tasks to high-effect, scoring 97.10% accuracy in attacker-id classification., Comment: Submitted to Insterspeech 2022

To study information processing in the brain, neuroscientists manipulate experimental stimuli while recording participant brain activity. They can then use encoding models to find out which brain "zone" (e.g. which region of interest, volume pixel or electrophysiology sensor) is predicted from the stimulus properties. Given the assumptions underlying this setup, when stimulus properties are predictive of the activity in a zone, these properties are understood to cause activity in that zone. In recent years, researchers have used neural networks to construct representations that capture the diverse properties of complex stimuli, such as natural language or natural images. Encoding models built using these high-dimensional representations are often able to significantly predict the activity in large swathes of cortex, suggesting that the activity in all these brain zones is caused by stimulus properties captured in the representation. It is then natural to ask: "Is the activity in these different brain zones caused by the stimulus properties in the same way?" In neuroscientific terms, this corresponds to asking if these different zones process the stimulus properties in the same way. Here, we propose a new framework that enables researchers to ask if the properties of a stimulus affect two brain zones in the same way. We use simulated data and two real fMRI datasets with complex naturalistic stimuli to show that our framework enables us to make such inferences. Our inferences are strikingly consistent between the two datasets, indicating that the proposed framework is a promising new tool for neuroscientists to understand how information is processed in the brain., Comment: Accepted to CLeaR 2022

Describes student exploration of a variety of approaches to learning how to operate with rational numbers. (YDS)

Shares the results of some of the algebraic-thinking pattern problems presented in "Principles and Standards for School Mathematics." (KHR)

There is limited research that explores Indigenous gambling and gaming, and the research that does exist focuses primarily on problem gambling and addiction. Western understandings of gambling are not an appropriate lens to provide a full understanding of the cultural depth and meaning behind traditional Indigenous games. Indigenous gaming has existed since pre colonization and occurs in many Indigenous cultures throughout North America. The following research project explores traditional Blackfoot forms of gambling and games. The current study used a mixed methods approach to collect data on Blackfoot traditional games and forms of gambling. The first method was an anonymous, online survey that collected data on games played today and demographics. The second method was key informant interviews with Elders from the local Blackfoot community and collected data regarding the traditional context of Blackfoot games. Survey results indicate that the most common game still played today is handgame. Interview results corroborate this finding. Results of the interviews, drawn from a thematic analysis performed in NVivo 12, show that gambling, types of games and their rules, evident change, and relationships are the most prominent themes surrounding traditional games. Traditional games are multifaceted at their core and serve various meaningful purposes in the lives of players, with the majority of players viewing games positively. Through a combination of scientific methods and Indigenous ways of knowing comes a history of Blackfoot gaming, from tradition, through colonization, to the post-colonial present day – in which these games continue to provide a connection to culture, spirituality, and community for Blackfoot Peoples. By considering traditional games and forms of gambling from an Indigenous perspective, further research can be better informed when engaging with Indigenous populations and exploring Indigenous-related topics

Similar to how differences in the proficiency of the cardiovascular and musculoskeletal system predict an individual's athletic ability, differences in how the same brain region encodes information across individuals may explain their behavior. However, when studying how the brain encodes information, researchers choose different neuroimaging tasks (e.g., language or motor tasks), which can rely on processing different types of information and can modulate different brain regions. We hypothesize that individual differences in how information is encoded in the brain are task-specific and predict different behavior measures. We propose a framework using encoding-models to identify individual differences in brain encoding and test if these differences can predict behavior. We evaluate our framework using task functional magnetic resonance imaging data. Our results indicate that individual differences revealed by encoding-models are a powerful tool for predicting behavior, and that researchers should optimize their choice of task and encoding-model for their behavior of interest.

The recent emergence of deepfakes has brought manipulated and generated content to the forefront of machine learning research. Automatic detection of deepfakes has seen many new machine learning techniques, however, human detection capabilities are far less explored. In this paper, we present results from comparing the abilities of humans and machines for detecting audio deepfakes used to imitate someone's voice. For this, we use a web-based application framework formulated as a game. Participants were asked to distinguish between real and fake audio samples. In our experiment, 472 unique users competed against a state-of-the-art AI deepfake detection algorithm for 14912 total of rounds of the game. We find that humans and deepfake detection algorithms share similar strengths and weaknesses, both struggling to detect certain types of attacks. This is in contrast to the superhuman performance of AI in many application areas such as object detection or face recognition. Concerning human success factors, we find that IT professionals have no advantage over non-professionals but native speakers have an advantage over non-native speakers. Additionally, we find that older participants tend to be more susceptible than younger ones. These insights may be helpful when designing future cybersecurity training for humans as well as developing better detection algorithms., Comment: Published at ACM Multimedia 2022 Workshop DDAM First International Workshop on Deepfake Detection for Audio Multimedia at ACM Multimedia 2022

We present our analysis of a significant data artifact in the official 2019/2021 ASVspoof Challenge Dataset. We identify an uneven distribution of silence duration in the training and test splits, which tends to correlate with the target prediction label. Bonafide instances tend to have significantly longer leading and trailing silences than spoofed instances. In this paper, we explore this phenomenon and its impact in depth. We compare several types of models trained on a) only the duration of the leading silence and b) only on the duration of leading and trailing silence. Results show that models trained on only the duration of the leading silence perform particularly well, and achieve up to 85% percent accuracy and an equal error rate (EER) of 15.1%. At the same time, we observe that trimming silence during pre-processing and then training established antispoofing models using signal-based features leads to comparatively worse performance. In that case, EER increases from 3.6% (with silence) to 15.5% (trimmed silence). Our findings suggest that previous work may, in part, have inadvertently learned thespoof/bonafide distinction by relying on the duration of silence as it appears in the official challenge dataset. We discuss the potential consequences that this has for interpreting system scores in the challenge and discuss how the ASV community may further consider this issue.

This work examines the content and usefulness of disentangled phone and speaker representations from two separately trained VQ-VAE systems: one trained on multilingual data and another trained on monolingual data. We explore the multi- and monolingual models using four small proof-of-concept tasks: copy-synthesis, voice transformation, linguistic code-switching, and content-based privacy masking. From these tasks, we reflect on how disentangled phone and speaker representations can be used to manipulate speech in a meaningful way. Our experiments demonstrate that the VQ representations are suitable for these tasks, including creating new voices by mixing speaker representations together. We also present our novel technique to conceal the content of targeted words within an utterance by manipulating phone VQ codes, while retaining speaker identity and intelligibility of surrounding words. Finally, we discuss recommendations for further increasing the viability of disentangled representations., Comment: Accepted to Speech Synthesis Workshop 2021 (SSW11)

Robert D Murray,1 Kirk W Kerr,2 Cory Brunton,2 Jennifer A Williams,2 Tiffany DeWitt,2 Karyn L Wulf2,3 1The Ohio State University College of Medicine, Columbus, OH, USA; 2Abbott Nutrition, Columbus, OH, USA; 3Nationwide Children’s Hospital Division of Emergency Medicine, Columbus, OH, USACorrespondence: Karyn L WulfAbbott Nutrition, 2900 Easton Square Place, Columbus, OH, 43219, USATel +1 614-370-4491Email Karyn.wulf@abbott.comBackground: Childhood malnutrition remains far too common around the world today. In this paper, we discuss pediatric malnutrition in the context of protein-energy undernutrition and hidden hunger (single or multiple micronutrient deficiencies). Recent growth statistics show that nearly 150 million children under 5 years are stunted, and 50 million are wasted. At the same time, millions more children experience undernutrition of lesser severity but with negative effects on growth nonetheless. Beyond poor growth, such nutrient shortfalls can predispose children to impaired physical and mental development, which may have lifelong consequences. The World Health Organization recently set an aspirational goal “to end all forms of childhood malnutrition by 2030.”Plan of Action: Our paper proposes a stepwise strategy to raise awareness of childhood malnutrition risk and to work toward building a consensus on pediatric malnutrition screening as a pathway to ending childhood malnutrition. In a full plan for action, we propose to convene an expert Working Consensus Group on Pediatric Nutrition Screening (WCG-PNS). We propose that this group will work to (1) identify malnutrition screening tools specific for universal screening of children in hospital and community settings, (2) plan and lead healthcare professional training on how to screen for malnutrition risk, then take appropriate follow-up steps, (3) guide and advise quality improvement programs (QIPs) to facilitate incorporation of nutrition screening and interventions

One outstanding threat to the honey bee, Apis mellifera L., is the unintended exposure of these pollinators to agricultural pesticides. Anthranilic diamides, such as chlorantraniliprole, work in target pests through activation of the ryanodine receptor causing rapid feeding cessation, lethargy, paralysis, and eventual death. Chlorantraniliprole exhibits low acute toxicity to honey bees but relatively little information exists regarding effects of chlorantraniliprole exposure beyond mortality. The acute toxicity and metabolic detoxification enzyme activities following oral exposure were determined in the honey bee and the model crop pest, the fall armyworm (Spodoptera frugiperda, J.E. Smith). Additionally, chlorantraniliprole effects on honey bee survival, locomotor activity, and two immunological parameters following immune challenge were observed in laboratory bioassays. After a 72 h exposure, neither the technical grade or three formulated products of chlorantraniliprole were observed to be acutely toxic to honey bees at solubility limits or maximum recommended label concentrations, respectively. However, all treatments with chlorantraniliprole were acutely toxic to fall armyworm. Honey bees receiving a 4 h exposure of all chlorantraniliprole treatments survived similar to untreated controls although honey bees receiving a 72 h exposure had significantly decreased survival over 30 d. The locomotor activity of honey bees receiving a 4 h exposure of chlorantraniliprole had decreased locomotion in one formulation treatment group, but walking activities returned to control levels the following day. In contrast, walking distance was significantly decreased at each time point in honey bees receiving chlorantraniliprole treatment for 72 h. Following determination of cytochrome P450 monooxygenase (P450), general esterase (EST), and glutathione S-transferase (GST) activities in both insect species, P450- and GST-mediated detoxification appears likely in the fall armyworm

This report describes the 2020 “From a Thesis to a Paper” Course hosted by the Department of Epidemiology and Global Health (EpiGH) at Umeå University, Sweden, and run across weeks 26-35 2020 (22nd June -28th August). The Course guides and advises early career researchers in scientific writing for targeted peer reviewed journals. Students develop papers based on their own research – typically a masters’ thesis. They work in small groups reviewing one another’s work and revising and re-drafting their manuscripts with support and input from the teacher. By the end of the Course students are expected to produce formatted draft documents suitable for submission to a peer reviewed journal of choice. Eligibility requires either one- or two-year masters’ qualifications in public health (MPH), health care or health education. Students and alumni from EpiGH are the target group. Applications and enrolments are processed by Umeå University. Students residing outside the European Union (EU) or European Economic Area (EEA) are required to pay University tuition fees (12,367 Swedish Krona). The majority of students enrolled in EpiGH masters’ courses are from low- and middle-income countries. In 2019 the Taylor & Francis Company (T&F) introduced a Scholarship Fund to pay tuition fees for up to ten non-EU/EEA students to enrol in the “From a Thesis to a Paper” Course each year. This report is in accordance with section 4.4 of the contract between T&F and Umeå University which affirms this financial support. See Appendix A.

Background: The use of large quantities of antimicrobial drugs for human health and agriculture is advancing the predominance of drug resistant pathogens in the environment. Antimicrobial resistance is now a major public health threat posing significant challenges for achieving the Sustainable Development Goals. In Bangladesh, where over one third of the population is below the poverty line, the achievement of safe and effective antibiotic medication use for human health is challenging. Objective: To explore factors and practices around access and use of antibiotics and understanding of antimicrobial resistance in rural communities in Bangladesh from a socio-cultural perspective. Methods: This qualitative study comprises the second phase of the multi-country ABACUS (Antibiotic Access and Use) project in Matlab, Bangladesh. Information was collected through six focus group discussions and 16 in-depth interviews. Informants were selected from ten villages in four geographic locations using the Health and Demographic Surveillance System database. The Access to Healthcare Framework guided the interpretation and framing of the findings in terms of individuals' abilities to: perceive, seek, reach, pay and engage with healthcare. Results: Village pharmacies were the preferred and trusted source of antibiotics for self-treatment. Cultural and religious beliefs informed the use of herbal and other complementary medicines. Advice on antibiotic use was also sourced from trusted friends and family members. Access to government-run facilities required travel on poorly maintained roads. Reports of structural corruption, stock-outs and patient safety risks eroded trust in the public sector. Some expressed a willingness to learn about antibiotic resistance. Conclusion: Antimicrobial resistance is both a health and development issue. Social and economic contexts shape medicine seeking, use and behaviours. Multi-sectoral action is needed to confront the underlying social, economic, c, Special issue

Health system responsiveness is an indicator that can be used for evaluating how well healthcare systems respond to people's needs in non-clinical areas such as communication, autonomy and confidentiality. This study analyses health system responsiveness from the perspective of community-dwelling adults aged 50 and over in China, Ghana, India, the Russian Federation and South Africa using cross-sectional data from the World Health Organization Study on global AGEing and adult health. The aim is to assess and compare how individual, health condition and healthcare factors impact differently on outpatient and inpatient responsiveness. Poor responsiveness is measured according to participants' responses to questions on a five-point Likert scale. Five univariate and multiple logistic regression models test associations between individual, health condition and healthcare factors and poor responsiveness. The final model adjusts for country. Key results are that travel time is a major contributor to poor responsiveness across all countries. Similarly there are wealth inequalities in responsiveness. However no clear difference in responsiveness was observed in presentations for chronic versus other types of conditions. This study provides an interesting baseline on older patients' perceived treatment within outpatient and inpatient facilities in five diverse low- and middle-income countries.

COVID-19 shocked health and economic systems leaving millions of people without employment and safety nets. The pandemic disproportionately affects people with substance use disorders (SUDs) due to the collision between SUDs and COVID-19. Comorbidities and risk environments for SUDs are likely risk factors for COVID-19. The pandemic, in turn, diminishes resources that people with SUD need for their recovery and well-being. This article presents an interdisciplinary and international perspective on how COVID-19 and the related systemic shock impact on individuals with SUDs directly and indirectly. We highlight a need to understand SUDs as biopsychosocial disorders and use evidence-based policies to destigmatize SUDs. We recommend a suite of multi-sectorial actions and strategies to strengthen, modernize and complement addiction care systems which will become resilient and responsive to future systemic shocks similar to the COVID-19 pandemic.

BACKGROUND: As in many Sub-Saharan African countries, the health system in Somalia is not operating at the capacity needed to lift childhood vaccination coverage to ninety percent or above, as recommended by United Nations Children's Fund. Current national estimates of coverage for the six major vaccine preventable childhood diseases range from thirty to sixty percent. Infectious disease outbreaks continue to pose significant challenges for the country's health authorities. OBJECTIVE: This important qualitative study, conducted in Galkayo District, Somalia, investigates limiting factors associated with childhood vaccination uptake from the perspective of both communities and health care workers. METHODS: Qualitative information was collected through six focus group discussions with parents (n = 48) and five one-to-one interviews with health workers (n = 15) between March and May 2017, in three settings in the Galkayo District - Galkayo city, Bayra and Bacadwayn. RESULTS: From a health system perspective, the factors are: awareness raising, hard to reach areas, negative attitudes and perceived knowledge of health workers, inadequate supplies and infrastructure, and missed vaccination opportunities. From the perspective of individuals and communities the factors are: low trust in vaccines, misinterpretation of religious beliefs, vaccine refusals, Somalia's patriarchal system and rumours and misinformation. Parents mostly received immunization information from social mobilizers and health facilities. Fathers, who are typically family decision-makers, were poorly informed. The findings highlight the need for in-service training to enable health workers to improve communication with parents, particularly fathers, peripheral communities and local religious leaders. CONCLUSIONS: Enhancing knowledge and awareness of vaccination among parents is crucial. Fathers' involvement is lacking. This may be boosted by highlighting fathers' obligation to protect their children's healt

BackgroundBrain metastases (BrM) develop in 20-40% of cancer patients and represent an unmet clinical need. Limited access of drugs into the brain because of the blood-brain barrier is at least partially responsible for therapeutic failure, necessitating improved drug delivery systems.MethodsGreen fluorescent protein (GFP)-transduced murine and nontransduced human hematopoietic stem cells (HSCs) were administered into mice (n = 10 and 3). The HSC progeny in mouse BrM and in patient-derived BrM tissue (n = 6) was characterized by flow cytometry and immunofluorescence. Promoters driving gene expression, specifically within the BrM-infiltrating HSC progeny, were identified through differential gene-expression analysis and subsequent validation of a series of promoter-green fluorescent protein-reporter constructs in mice (n = 5). One of the promoters was used to deliver tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) to BrM in mice (n = 17/21 for TRAIL vs control group).ResultsHSC progeny (consisting mostly of macrophages) efficiently homed to macrometastases (mean [SD] = 37.6% [7.2%] of all infiltrating cells for murine HSC progeny; 27.9% mean [SD] = 27.9% [4.9%] of infiltrating CD45+ hematopoietic cells for human HSC progeny) and micrometastases in mice (19.3-53.3% of all macrophages for murine HSCs). Macrophages were also abundant in patient-derived BrM tissue (mean [SD] = 8.8% [7.8%]). Collectively, this provided a rationale to optimize the delivery of gene therapy to BrM within myeloid cells. MMP14 promoter emerged as the strongest promoter construct capable of limiting gene expression to BrM-infiltrating myeloid cells in mice. TRAIL delivered under MMP14 promoter statistically significantly prolonged survival in mice (mean [SD] = 19.0 [3.4] vs mean [SD] = 15.0 [2.0] days for TRAIL vs control group; two-sided P = .006), demonstrating therapeutic and translational potential of our approach.ConclusionsOur study establishes HSC gene therapy using a mye

Vector Quantized Variational AutoEncoders (VQ-VAE) are a powerful representation learning framework that can discover discrete groups of features from a speech signal without supervision. Until now, the VQ-VAE architecture has previously modeled individual types of speech features, such as only phones or only F0. This paper introduces an important extension to VQ-VAE for learning F0-related suprasegmental information simultaneously along with traditional phone features.The proposed framework uses two encoders such that the F0 trajectory and speech waveform are both input to the system, therefore two separate codebooks are learned. We used a WaveRNN vocoder as the decoder component of VQ-VAE. Our speaker-independent VQ-VAE was trained with raw speech waveforms from multi-speaker Japanese speech databases. Experimental results show that the proposed extension reduces F0 distortion of reconstructed speech for all unseen test speakers, and results in significantly higher preference scores from a listening test. We additionally conducted experiments using single-speaker Mandarin speech to demonstrate advantages of our architecture in another language which relies heavily on F0., Comment: Submitted to Interspeech 2020

We aim to characterize how different speakers contribute to the perceived output quality of multi-speaker Text-to-Speech (TTS) synthesis. We automatically rate the quality of TTS using a neural network (NN) trained on human mean opinion score (MOS) ratings. First, we train and evaluate our NN model on 13 different TTS and voice conversion (VC) systems from the ASVSpoof 2019 Logical Access (LA) Dataset. Since it is not known how best to represent speech for this task, we compare 8 different representations alongside MOSNet frame-based features. Our representations include image-based spectrogram features and x-vector embeddings that explicitly model different types of noise such as T60 reverberation time. Our NN predicts MOS with a high correlation to human judgments. We report prediction correlation and error. A key finding is the quality achieved for certain speakers seems consistent, regardless of the TTS or VC system. It is widely accepted that some speakers give higher quality than others for building a TTS system: our method provides an automatic way to identify such speakers. Finally, to see if our quality prediction models generalize, we predict quality scores for synthetic speech using a separate multi-speaker TTS system that was trained on LibriTTS data, and conduct our own MOS listening test to compare human ratings with our NN predictions., Comment: accepted at Speaker Odyssey 2020

We present a new approach to disentangle speaker voice and phone content by introducing new components to the VQ-VAE architecture for speech synthesis. The original VQ-VAE does not generalize well to unseen speakers or content. To alleviate this problem, we have incorporated a speaker encoder and speaker VQ codebook that learns global speaker characteristics entirely separate from the existing sub-phone codebooks. We also compare two training methods: self-supervised with global conditions and semi-supervised with speaker labels. Adding a speaker VQ component improves objective measures of speech synthesis quality (estimated MOS, speaker similarity, ASR-based intelligibility) and provides learned representations that are meaningful. Our speaker VQ codebook indices can be used in a simple speaker diarization task and perform slightly better than an x-vector baseline. Additionally, phones can be recognized from sub-phone VQ codebook indices in our semi-supervised VQ-VAE better than self-supervised with global conditions., Comment: Accepted to ICASSP 2021

In 2019 the Department of Epidemiology and Global Health (EpiGH) Umeå University, conducted the “From a Thesis to a Paper” Summer Course for Master of Public Health (MPH) students and alumni from June 18th to August 30th. The aim of this optional Course was to guide and advise early career researchers in the development of material from their masters theses into manuscripts suitable for submission to peer reviewed journals. Scholarships awarded to MPH students from low- and middle-income countries (LMICs) do not cover fees for non-compulsory courses. In 2019 the Taylor & Francis Company introduced a Scholarship Fund which provides course fee waivers for up to ten eligible students in the “From a Thesis to a Paper” Course each year. This report describes the implementation of the Fund and sponsored students’ academic progress. Scheduling the Course in the first year of the Scholarship Fund was not without administrative difficulties because University admissions processes for 2019 summer courses opened prior to the signing of the contract with the Taylor & Francis Company. (See Attachment 1). However we are confident that the “teething issues” outlined here will be resolved in future cycles.

Background: Despite global achievements in reducing early childhood mortality, disparities remain. There have been empirical studies of inequalities conducted in low- and middle-income countries. However, there have been no epidemiological studies on socioeconomic inequalities and early childhood survival in Myanmar. Objective: To estimate associations between two measures of parental socioeconomic status - household wealth and education - and age-specific early childhood mortality in Myanmar. Methods: Using cross-sectional data obtained from the Myanmar Demographic Health Survey (2015-2016), univariate and multiple logistic regressions were performed to investigate associations between household wealth and highest attained parental education, and under-5, neonatal, post-neonatal and child mortality. Data for 10,081 children born to 5,932 married women (aged 15-49 years) 10 years prior to the survey, were analysed. Results: Mortality during the first five years was associated with household wealth. In multiple logistic models, wealth was protective for post-neonatal mortality. After adjusting for individual proximate determinants, the odds of post-neonatal mortality in the richest households were 85% lower (95% CI: 50-96%) than in the poorest households. However, significant association was not found between wealth and neonatal mortality. Parental education was important for early childhood mortality; the highest benefit from parental education was for child mortality in the one- to five-year age bracket. After adjusting for proximate determinants, children with a higher educated parent had 95% (95% CI 77-99%) lower odds of death in this age group compared with children whose parents' highest educational attainment was at primary level. The association between parental education and neonatal mortality was not significant. Conclusions: In Myanmar, household wealth and parental education are important for childhood survival before five years of age. This study identif