ABSTRACT
This work addresses the vital need of keeping people informed with relevant, correct and essential information during the pandemic. Advanced NLP and machine learning mechanisms have been leveraged to generate responses to user queries through contextual conversation. In order to help people be discerning about what information they receive, a conversational system is proposed that identifies the correct intent of the query and a reinforcement Learning based generation model is used to proceed with conversation. We propose an end-to-end real-time text generation model that can respond to users queries on covid19. We created a new dataset with 1200+ covid-related questions from various sources and pre-processed them for a brief and direct answer. The dataset has also been manually observed to identify depressed questions and the responses are converted to be more empathetic. The dataset has been used to fine-tune DailoGPT, a GPT2-based transformer model to generate the responses related to COVID. COVID-related queries are bucketed into 15 categories to identify the exact intent of people. Our model generated both contextual and empathetic responses and achieved a human evaluation score of 3.48 (on a scale of 5) in terms of contextual relevance and a score of 2.12 (on a scale of 3). © 2022 IEEE.
ABSTRACT
Purpose: National guidelines recommend universal germline genetic testing (GT) for patients with Pancreatic Ductal Adenocarcinoma (PDAC), but rates of testing remain low. Given the aggressiveness of PDAC, the window of opportunity for GT is short and often overshadowed by treatment initiation and other clinical milestones. Thus, there is an unmet need for a model that streamlines GT and makes it available to a wider audience in a rapid fashion. Moreover, in pandemic times, video-based alternatives for medical care are increasingly relevant. Methods: We implemented a novel care delivery model in which a seven-minute educational video describing the benefits, risks, and implications of GT was shown to PDAC patients. The video was shown in lieu of an initial consult with a genetic counselor. Only patients who had not undergone GT or previously met with a genetic counselor were included. After watching the video, patients could elect to pursue GT and get tested on-side or remotely (at home). Genetic counselors disclosed results and provided post-test counseling by phone. Clinical and germline data were collected through medical records on a cohort of PDAC patients seen at the Gastrointestinal Center-MD Anderson during a 2-year enrollment period (May 2019-July 2021), which included the COVID-19 pandemic period. Results: A total of 286 PDAC patients watched the educational video. From 175 patients that watched the video pre-pandemic, 12 declined testing, whereas in the post-pandemic period, none of the 111 patients declined testing (6.9% vs 0%;p<0.004). We excluded data from 29 patients who elected to undergo GT but declined to participate in the registry. From the 241 patients with successfully collected samples, 21 patients (8.7%) had a pathogenic variant (PV), 38 patients (15.8%) had a Variant of Uncertain Significance (VUS), and 182 patients (75.5%) tested negative. The pathogenic variants detected included: BRCA2 (most frequent), ATM, BRCA1, CDKN2A, PALB2 and APC. Conclusions: GT can have tremendously beneficial effects, such as qualifying for targeted treatment options and facilitating cancer prevention in probands' at-risk family members. Comparing uptake of GT pre- versus post-pandemic suggests that patients were more willing to trust information from a video platform, likely due to the global effect of living in a virtual society as a result of the pandemic. We suggest an approach in which every PDAC patient is shown a genetics educational video and given the choice to undergo GT and post-result counseling, greatly reducing the burden on genetic counselors. We report here the feasibility of implementing video-based germline testing in PDAC patients which resulted in unexpectedly high uptake levels, particularly post-pandemic. Further investigations are needed to explore the feasibility of a fully remote GT model in diverse populations to assess additional barriers to universal GT.
ABSTRACT
Exchange-Traded Funds (ETFs) have revolutionised the asset-management industry with high liquidity and low bid-asks allowing investors to access a diversified portfolio cheaply. The desirable liquidity characteristics of ETFs are, however, in contrast with their behavior in crisis periods. This paper studies the breakdown in the fixed-income ETFs (FIETFs) market, during the peak of the COVID-19-driven liquidity shock of March 2020. We argue that FIETFs provide an illusion of liquidity and the liquidity mismatch between the ETF and the underlying manifests itself in terms of very significant differences between the price and the Net Asset Value (NAV). We run a further analysis on the dislocation by comparing it with equity ETFs. Comparisons suggest that the cost of liquidity is very high for fixed-income ETFs with significant tracking errors during volatile periods in contrast with the better-behaved equity ETFs. Further analysis on the performance patterns of FIETFs indicates that both the price and NAV might have deviated from fundamentals with an overreaction in the ETF price accompanying an underreaction in the NAV. We also study the phenomena of ‘dealer inventory management’. FIETFs are uniquely different from equity ETFs in that the authorised participant (arbitrageurs) tend to be the large banks that are also market-makers in the underlying securities. Therefore, we show that the incentives of the arbitrageurs may not always be aligned towards arbitraging the price-NAV differential. In a novel empirical study using trading volumes and position changes for the largest corporate bond ETF (LQD), we suggest that inventory management by the larger broker-dealers may have exacerbated the dislocation. We believe that the conflicting objectives of dealers have further increased due to high balance-sheet costs imposed upon them post the Global Financial Crisis. Finally, we propose the use of derivatives, in particular credit derivatives, for the risk management of liquidity shocks. We show that the drawdown from rapid liquidity shocks can be reduced significantly through exposure to credit convexity. © Henry Stewart Publications.
ABSTRACT
The Sundarban is the world's biggest mangrove forest, located on Bangladesh's and India's southern coasts. The forest protects biodiversity and the local population from various natural disasters. This paper has primarily focused on the change in land use land cover (LULC) of a location in Sundarban delta (21°56' N, 89°10' E) using two LANDSAT satellite images for the year 2015 and the lockdown period of 2020, with the help of Geographical Information System (GIS) and supervised machine learning technique. With the reduction of upstream water and sediment supplies, the water body in the Bay of Bengal has increased. The extent of vegetation and agricultural land has decreased due to urbanization and erosion by the Bay of Bengal. The difference between the LULC before and after the Amphan cyclone indicates that land area has been reduced over this area. © 2021 German Member Committee of URSI.
ABSTRACT
This letter reveals the prevailing scenario of raindrop size distribution (DSD) in terms of mass-weighted mean drop diameter ( <formula> <tex>$D_{m}$</tex> </formula> ) over a tropical metropolis, Kolkata (22.57°N, 88.37°E), India, in a contrasting aerosol environment that occurred during the COVID-19 pandemic in the absence of usual human activities. In the premonsoon months (March-May), the probability of <formula> <tex>$D_{m}$</tex> </formula> values exceeding 2 mm has increased in 2020, indicating the dominance of large raindrops, compared to the years 2017-2019. Increased number densities of larger drops have influenced the drop fall velocity spectrum as measured by a laser precipitation monitor in terms of the percentage occurrence of high-velocity small drops (superterminal) and low-velocity large drops (subterminal) for both convective and stratiform precipitations. As measured from a Ka-band microrain Doppler radar, the mean melting layer altitude during stratiform rain has decreased by ~800 m during the premonsoon of 2020 compared to 2017-2019. According to the ERA-5 reanalysis data, changing rain microphysical characteristics are related to decreasing zero-degree isotherm height and reduced wind shear. IEEE
ABSTRACT
The ongoing COVID-19 pandemic necessitated a complete lockdown from 24 March to 31 May 2020 as imposed by the Government of India to prevent the spread of the virus that resulted into halting most of the economic activities during that period. The stringent anthropogenic activities during the lockdown phase resulted in a notable drop in the air pollution level. The prevailing global scenario during the pandemic has provided an opportunity of assessing the air quality in the absence of normal anthropogenic activities. In the present article, the impacts of reduced anthropogenic activities on the black carbon concentration and associated atmospheric parameters have been studied over a tropical urban location, Kolkata (22.57 degrees N, 88.37 degrees E) in eastern India. Investigations have been made on the impacts on the parameters like atmospheric electric field, wind speed, temperature, relative humidity, dew point temperature, rain accumulation and soil moisture during this unusual time in comparison to the same time span of the previous year 2019 when the usual human activities were pursued.
ABSTRACT
Background: New-onset hyperglycemia may suggest presence of occult pancreatic cancer. The aim of this pilot study was to assess feasibility of pancreatic imaging in patients enrolled in the prospective NOD cohort study. Methods:We conducted a prospective pilot study November 2018-April 2020 within Kaiser Permanente Southern California. Patients 50-85 years enrolled in the NOD study (newly elevated glycemic parameter, no history of diabetes) were invited to complete a three-phase contrast-enhanced computed tomography pancreas-protocol scan. Time to imaging, abnormal pancreatic findings, incidental extra-pancreatic findings, including those prompting additional clinical evaluation were identified. Variability in clinical reporting (descriptors of pancreatic duct and parenchyma) between medical centers was assessed. Results: The majority (88.4%, N = 130) of 147 patients invited consented to undergo imaging;91 scans were completed (prior to COVID-19 stay-at-home orders). Median time from radiology order to imaging was 19.7 days (IQR, 15.4-27.6). Median age was 60.8 years (IQR, 56.3-68.8), 37.8% female;race/ethnicity was Hispanic (41.1%), followed by non-Hispanic white (27.8%), blacks (13.3%), and Asians (13.3%). One (1.1%) case of pancreatic cancer was detected between enrollment and study imaging;12/91 (13.1%) of patients had other pancreatic findings: 2 atrophy, 5 fatty infiltration, 1 divisum, 3 cysts, 1 calcification. Among those with findings, 2 (16.7%) underwent further diagnostic evaluation. Therewere 57 extra-pancreatic findings among 52 (57.1%) unique patients, of which 21.1% (12/57) prompted clinical evaluation. Reports from one of the 8 participating medical centers more frequently described both normal pancreatic parenchyma and normal ducts (39/42 (92.9%) compared to 9/49 (18.4%), P < 0.0001). Conclusions: Among participants in this pilot NOD study, pancreatic imaging was found to be acceptable and feasible to be completed in a timely fashion. There was a high rate of incidental findings as well as significant variability in clinical reports. These challenges will need to be addressed in future studies evaluating early detection of pancreatic cancer.