Search for: All records

Operational networks commonly rely on machine learning models for many tasks, including detecting anomalies, inferring application performance, and forecasting demand. Yet, model accuracy can degrade due to concept drift, whereby the relationship between the features and the target to be predicted changes. Mitigating concept drift is an essential part of operationalizing machine learning models in general, but is of particular importance in networking's highly dynamic deployment environments. In this paper, we first characterize concept drift in a large cellular network for a major metropolitan area in the United States. We find that concept drift occurs across many important key performance indicators (KPIs), independently of the model, training set size, and time interval---thus necessitating practical approaches to detect, explain, and mitigate it. We then show that frequent model retraining with newly available data is not sufficient to mitigate concept drift, and can even degrade model accuracy further. Finally, we develop a new methodology for concept drift mitigation, Local Error Approximation of Features (LEAF). LEAF works by detecting drift; explaining the features and time intervals that contribute the most to drift; and mitigates it using forgetting and over-sampling. We evaluate LEAF against industry-standard mitigation approaches (notably, periodic retraining) with more than four years of cellular KPI data. Our initial tests with a major cellular provider in the US show that LEAF consistently outperforms periodic and triggered retraining on complex, real-world data while reducing costly retraining operations.

 

The COVID-19 pandemic demonstrated the public health benefits of reliable and accessible point-of-care (POC) diagnostic tests for viral infections. Despite the rapid development of gold-standard reverse transcription polymerase chain reaction (RT-PCR) assays for SARS-CoV-2 only weeks into the pandemic, global demand created logistical challenges that delayed access to testing for months and helped fuel the spread of COVID-19. Additionally, the extreme sensitivity of RT-PCR had a costly downside as the tests could not differentiate between patients with active infection and those who were no longer infectious but still shedding viral genomes. To address these issues for the future, we propose a novel membrane-based sensor that only detects intact virions. The sensor combines affinity and size based detection on a membrane-based sensor and does not require external power to operate or read. Specifically, the presence of intact virions, but not viral debris, fouls the membrane and triggers a macroscopically visible hydraulic switch after injection of a 40 μL sample with a pipette. The device, which we call the μSiM-DX (microfluidic device featuring a silicon membrane for diagnostics), features a biotin-coated microslit membrane with pores ∼2–3× larger than the intact virus. Streptavidin-conjugated antibody recognizing viral surface proteins are incubated with the sample for ∼1 hour prior to injection into the device, and positive/negative results are obtained within ten seconds of sample injection. Proof-of-principle tests have been performed using preparations of vaccinia virus. After optimizing slit pore sizes and porous membrane area, the fouling-based sensor exhibits 100% specificity and 97% sensitivity for vaccinia virus ( n = 62). Moreover, the dynamic range of the sensor extends at least from 10 5.9 virions per mL to 10 10.4 virions per mL covering the range of mean viral loads in symptomatic COVID-19 patients (10 5.6 –10 7 RNA copies per mL). Forthcoming work will test the ability of our sensor to perform similarly in biological fluids and with SARS-CoV-2, to fully test the potential of a membrane fouling-based sensor to serve as a PCR-free alternative for POC containment efforts in the spread of infectious disease. 

The SARS-CoV-2 pandemic has revealed the need for rapid and inexpensive diagnostic testing to enable population-based screening for active infection. Neither standard diagnostic testing, the detection and measurement of viral RNA (via polymerase chain reaction), or serological testing (via enzyme-linked immunosorbent assay) has the capability to definitively determine active infection. The former due to a lack of ability to distinguish between replicable and inert viral RNA, and the latter due to varying immune responses (ranging from latent to a complete lack of immune response altogether). Despite many companies producing rapid point-of-care (POC) tests, none will address the global scale of testing needed and few help to combat the ever growing issue of testing resource scarcity. Here we discuss our efforts towards the development of a highly manufacturable, microfluidic device that instantly indicates active viral infection status from ~ 20 μL of nasal mucus or phlegm and requires no external power. The device features a biotin functionalized silicon nanomembrane within an acrylic body containing channels and ports for sample introduction and analysis. Virus capture and target confirmation are done using affinity-based capture and size-based occlusion respectively. Modularity of the device is proven with bead and vaccinia virus capture as we work towards testing with both pure SARS-CoV-2 virus and human samples. With success on all fronts, we could achieve an inexpensive POC diagnostic which can determine an individual’s infection status, aiding containment efforts in the current and future pandemics. In addition to direct viral detection, our method can be used as a rapid POC sample preparation tool that limits the application of PCR reagents to those samples which already display viral size and antigen-based positivity through our device. 

Abstract. The Surface Ocean CO2 Atlas (SOCAT) is a synthesis of quality-controlled fCO2 (fugacity of carbon dioxide) values for the global surface oceans and coastal seas with regular updates. Version 3 of SOCAT has 14.7 million fCO2 values from 3646 data sets covering the years 1957 to 2014. This latest version has an additional 4.6 million fCO2 values relative to version 2 and extends the record from 2011 to 2014. Version 3 also significantly increases the data availability for 2005 to 2013. SOCAT has an average of approximately 1.2 million surface water fCO2 values per year for the years 2006 to 2012. Quality and documentation of the data has improved. A new feature is the data set quality control (QC) flag of E for data from alternative sensors and platforms. The accuracy of surface water fCO2 has been defined for all data set QC flags. Automated range checking has been carried out for all data sets during their upload into SOCAT. The upgrade of the interactive Data Set Viewer (previously known as the Cruise Data Viewer) allows better interrogation of the SOCAT data collection and rapid creation of high-quality figures for scientific presentations. Automated data upload has been launched for version 4 and will enable more frequent SOCAT releases in the future. High-profile scientific applications of SOCAT include quantification of the ocean sink for atmospheric carbon dioxide and its long-term variation, detection of ocean acidification, as well as evaluation of coupled-climate and ocean-only biogeochemical models. Users of SOCAT data products are urged to acknowledge the contribution of data providers, as stated in the SOCAT Fair Data Use Statement. This ESSD (Earth System Science Data) "living data" publication documents the methods and data sets used for the assembly of this new version of the SOCAT data collection and compares these with those used for earlier versions of the data collection (Pfeil et al., 2013; Sabine et al., 2013; Bakker et al., 2014). Individual data set files, included in the synthesis product, can be downloaded here: doi:10.1594/PANGAEA.849770. The gridded products are available here: doi:10.3334/CDIAC/OTG.SOCAT_V3_GRID.