Ryan Silva - Academia.edu (original) (raw)
Papers by Ryan Silva
AGRARIAN ACADEMY
RESUMO Uma das estratégias elementares para o sucesso de empreendimentos silviculturais é a defin... more RESUMO Uma das estratégias elementares para o sucesso de empreendimentos silviculturais é a definição dos espaçamentos de plantio. Assim, este trabalho teve como objetivo a análise do desenvolvimento dendrométrico sobre o crescimento de plantas de mogno africano (Khaya grandifoliola C. Dc.) cultivadas em cinco diferentes espaçamentos (3x3 m, 4x4 m, 5x5 m, 6x6 m e 6x7 m). O plantio foi estabelecido em 2015 em Bonfinópolis/GO. A avaliação dendrométrica deu-se pela medição do diâmetro à altura do peito (DAP), por meio de suta e medição da altura total através do clinômetro digital. Foram medidas 78 árvores em cada espaçamento e determinada a média para o povoamento de cada espaçamento. O volume médio individual para cada espaçamento foi determinado por meio das variáveis diâmetro e altura, estimando por meio de modelo matemático ajustado para a região. Como o esperado, maiores médias de DAP foram observadas nos espaçamentos mais amplos (5x5 m, 6x6 m e 6x7 m). Para as variáveis altura e volume individual, foi constatada superioridade nos valores do espaçamento 5x5 m. Em relação à análise de desenvolvimento dendrométrico, mesmo com o espaçamento 5x5 m apresentando melhores médias de DAP, altura e volume individual, não é possível indicá-lo como melhor escolha de espaçamento, já que os mais adensados (3x3 m e 4x4 m) não estavam sob condições ideais de manejo, pois não sofreram desbaste.
The American Journal of Surgery
2022 52nd Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W)
Defending computer networks from cyber attack requires timely responses to alerts and threat inte... more Defending computer networks from cyber attack requires timely responses to alerts and threat intelligence. Decisions about how to respond involve coordinating actions across multiple nodes based on imperfect indicators of compromise while minimizing disruptions to network operations. Currently, playbooks are used to automate portions of a response process, but often leave complex decision-making to a human analyst. In this work, we present a deep reinforcement learning approach to autonomous response and recovery in large industrial control networks. We propose an attention-based neural architecture that is flexible to the size of the network under protection. To train and evaluate the autonomous defender agent, we present an industrial control network simulation environment suitable for reinforcement learning. Experiments show that the learned agent can effectively mitigate advanced attacks that progress with few observable signals over several months before execution. The proposed deep reinforcement learning approach outperforms a fully automated playbook method in simulation, taking less disruptive actions while also defending more nodes on the network. The learned policy is also more robust to changes in attacker behavior than playbook approaches.
2022 ACM/IEEE 13th International Conference on Cyber-Physical Systems (ICCPS)
Granulocyte colony-stimulating factor (G-CSF) is increasingly being used to prevent febrile neutr... more Granulocyte colony-stimulating factor (G-CSF) is increasingly being used to prevent febrile neutropenia associated with chemotherapy. Large-vessel vasculitis (LVV) has been recognized as a rare side effect of G-CSF treatment. We report a case of G-CSF associated LVV in a patient with breast cancer. While clear pathogenic mechanisms remain unknown, G-CSF may cause vasculitis due to inflammatory cytokines production. This adverse reaction should be recognized in patients with suggestive symptoms following the administration of pegfilgrastim. A 56-year-old woman with luminal B breast cancer who had undergone surgery and adjuvant chemotherapy, initially with paclitaxel, was started on a doxorubicin plus cyclophosphamide protocol, followed by supportive use of long-acting G-CSF pegfilgrastim. Following the administration of pegfilgrastim, the patient developed intermittent fever and was given empiric antibiotics in the outpatient setting with no improvement. There were no signs of cancer...
International Journal of Colorectal Disease, 2021
Preoperative neoadjuvant chemotherapy (NAC) for colon cancer in advanced stages is associated wit... more Preoperative neoadjuvant chemotherapy (NAC) for colon cancer in advanced stages is associated with improved outcomes and tumor regression. The aim of our study was to identify outcomes in patients with colon cancer who received preoperative NAC. A 4-year analysis of the American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) was performed. We included patients with locally advanced colon cancer (non-metastatic T3, T4 with or without nodal involvement) who underwent colon cancer resection. Patients were stratified into two groups (NAC and No-NAC). Our outcome measures were anastomotic leaks, hospital length of stay (LOS), 30-day complications, 30-day mortality, and 30-day readmissions. We performed a multi-variable logistic regression analysis. We analyzed 90,055 patients of which 7694 (NAC: 7.8 % (n = 599) and No-NAC: 92.2% (n = 7095)) met the inclusion criteria and included in the analysis. Mean age was 67 ± 13 years, 51% were males, and 72% were whites. Patients in the NAC group were more likely to be younger (60 ± 12 years vs. 68 ± 13 years, p < 0.01) and males (62% vs. 50%, p < 0.01) compared to No-NAC. On regression analysis, preoperative NAC was independently associated with higher odds of anastomotic leak (OR 1.35 [1.05–1.97], p = 0.03) and 30-day readmission (OR 1.54 [1.24–2.05], p < 0.01) in reference to No-NAC. However, no association was found between NAC and 30-day complications and 30-day mortality. Preoperative NAC might be associated with adverse outcomes of anastomotic leaks and 30-day readmissions, however does not appear to impact 30-day complications nor 30-day mortality.
Recent advancements in multilayer, multicellular, genetic logic circuits often rely on manual int... more Recent advancements in multilayer, multicellular, genetic logic circuits often rely on manual intervention throughout the computation cycle and orthogonal signals for each chemical "wire". These constraints can prevent genetic circuits from scaling. Microfluidic devices can be used to mitigate these constraints. However, continuousflow microfluidics are largely designed through artisanal processes involving handdrawing features and accomplishing design rule checks visually: processes that are also inextensible. Additionally, continuous-flow microfluidic routing is only a consideration during chip design and, once built, the routing structure becomes "frozen in silicon," or for many microfluidic chips "frozen in polydimethylsiloxane (PDMS)"; any changes to fluid routing often require an entirely new device and control infrastructure. The cost of fabricating and controlling a new device is high in terms of time and money; attempts to reduce one cost measure are, generally, paid through increases in the vi other. This work has three main thrusts: to create a microfluidic fabrication framework, called MakerFluidics, that lowers the barrier to entry for designing and fabricating microfluidics in a manner amenable to automation (Chapter 3); to prove this methodology can design, fabricate, and control complex and novel microfluidic devices (Chapter 4); and to demonstrate the methodology can be used to solve biologically-relevant problems (Chapter 5). Utilizing accessible technologies, rapid prototyping, and scalable design practices, the MakerFluidics framework has demonstrated its ability to design, fabricate and control novel, complex and scalable microfludic devices. This was proven through the development of a reconfigurable, continuous-flow routing fabric driven by a modular, scalable primitive called a transposer. In addition to creating complex microfluidic networks, MakerFluidics was deployed in support of cutting-edge, application-focused research at the Charles Stark Draper Laboratory. Informed by a design of experiments approach using the parametric rapid prototyping capabilities made possible by Mak-erFluidics, a plastic blood-bacteria separation device was optimized, demonstrating that the new device geometry can separate bacteria from blood while operating at 275% greater flow rate as well as reduce the power requirement by 82% for equivalent separation performance when compared to the state of the art. Ultimately, MakerFluidics demonstrated the ability to design, fabricate, and control complex and practical microfluidic devices while lowering the barrier to entry to continuous-flow microfluidics, thus democratizing cutting edge technology beyond a handful of well-resourced and specialized labs. vii A Othermill Standard Operating Procedure 92 References 98
Digital electronic circuits have inspired synthetic biologists to program living cells with synth... more Digital electronic circuits have inspired synthetic biologists to program living cells with synthetic decision-mak- ing circuits by creating multilevel genetic logic gates. In both genetic and electronic circuits, logic synthesis translates an abstract functional description into a representation that can be physically implemented. However, inherent differences in genetic logic devices present new synthesis challenges. Here we provide a starting point for a growing set of biodesign au- tomation tools to tackle this challenge in a unified way. This framework will enable direct comparison of new approaches, results that are transferable and standardized, and research into independent areas of the problem formulation with a clear path toward future integration.
ArXiv, 2021
Defending computer networks from cyber attack requires coordinating actions across multiple nodes... more Defending computer networks from cyber attack requires coordinating actions across multiple nodes based on imperfect indicators of compromise while minimizing disruptions to network operations. Advanced attacks can progress with few observable signals over several months before execution. The resulting sequential decision problem has large observation and action spaces and a long time-horizon, making it difficult to solve with existing methods. In this work, we present techniques to scale deep reinforcement learning to solve the cyber security orchestration problem for large industrial control networks. We propose a novel attention-based neural architecture with size complexity that is invariant to the size of the network under protection. A pre-training curriculum is presented to overcome early exploration difficulty. Experiments show in that the proposed approaches greatly improve both the learning sample complexity and converged policy performance over baseline methods in simulat...
Angewandte Chemie International Edition, 2021
Detection of pathogenic bacteria in complex biological matrices remains a major challenge. Herein... more Detection of pathogenic bacteria in complex biological matrices remains a major challenge. Herein, we report the selection and optimization of a new DNAzyme for Staphylococcus aureus (SA) and the use of the DNAzyme to develop a simple lateral flow device (LFD) for detection of SA in nasal mucus. The DNAzyme was generated by in vitro selection using a crude extra/intracellular mixture derived from SA, which could be used directly for simple solution or paper‐based fluorescence assays for SA. The DNAzyme was further modified to produce a DNA cleavage fragment that acted as a bridging element to bind DNA‐modified gold nanoparticles to the test line of a LFD, producing a simple colorimetric dipstick test. The LFD was evaluated with nasal mucus samples spiked with SA, and demonstrated that SA detection was possible in minutes with minimal sample processing.
The NIST-hosted Synthetic Biology Standards Consortium (SBSC) will collectively build the infrast... more The NIST-hosted Synthetic Biology Standards Consortium (SBSC) will collectively build the infrastructure to support a fully integrated global synthetic biology enterprise. We aim to accomplish this by developing metrology products – standards, including reference materials, reference data, reference methods, and documentary standards – that will enable coordination of labor and reuse of materials. We will present the results of the kick-off workshop for the SBSC, held on March 31, 2015 at Stanford University. A summary of the plans developed by each working group will be shared, and mechanisms of future consortium operations will be discussed.
Giant-Cell Arteritis [Working Title], 2021
Giant cell arteritis (GCA), also known as temporal arteritis or Horton disease, is categorized as... more Giant cell arteritis (GCA), also known as temporal arteritis or Horton disease, is categorized as a large- and medium-sized vessels vasculitis. Systemic symptoms are common in GCA and although vascular involvement may be widespread, the cranial branches of the aortic arch are responsible for the hallmark symptoms of GCA: headache, jaw claudication and ocular symptoms, particularly visual loss. The large vessel (LV)-GCA phenotype may differ or overlap from cranial arteritis. Clinical consequences of LV-GCA comprise aneurysms and dissections of the aorta, as well as stenosis, occlusion and ectasia of large arteries. Symptoms of polymyalgia rheumatica occurring in a patient with GCA include characteristic proximal polyarthralgias and myalgias, sometimes accompanied by remitting seronegative symmetrical synovitis with pitting edema (RS3PE), Less common manifestations reported include central nervous system involvement, audiovestibular and upper respiratory symptoms, pericarditis, mesent...
International Journal of Colorectal Disease, 2021
Purpose Brain metastasis (BM) in colorectal cancer patients is rare and is associated with dismal... more Purpose Brain metastasis (BM) in colorectal cancer patients is rare and is associated with dismal outcomes. Our study aims to evaluate the incidence and predictors of BM in patients with colorectal cancer. Methods We performed a retrospective analysis (2010–2017) of patients with a primary diagnosis of colorectal cancer (CRC). Patients were stratified into two groups (BM vs. No-BM). Outcome measures were the incidence and predictors of BM. Multivariate logistic regression analysis was performed. Results A total of 230,806 patients were analyzed. A total of 0.30% ( n = 691) of the patients were found to have BM. On multivariate logistics regression, bone (OR: 5.39 [3.36–8.65], p < 0.001), lung (OR: 3.75 [2.67–5.28], < 0.001), and distant node metastasis (OR: 32.75 [20.47–52.41], p < 0.001) were independent predictors of BM. Conclusion Our study supports the low incidence of brain metastasis in patients with colorectal cancer. A unique set of characteristics is identified to confer an increased risk of brain metastases.
Journal of the American College of Surgeons, 2020
2012 ASEE Annual Conference & Exposition Proceedings
In June 2010, we reported to the United States Air Force Academy to teach in the Department of El... more In June 2010, we reported to the United States Air Force Academy to teach in the Department of Electrical and Computer Engineering. Each of us brought different skill sets, different experiences, and different interests from different jobs in different locations. Yet, throughout our first year we simultaneously observed common stumbling blocks leading to some dilemmas and curiosities. This paper presents some of these common observations from three dissimilar instructors within a contextual framework promoting a learning-centered paradigm and balancing technical syllabus content with real-world "soft" skills. It is not our intention to define these observations as "Tricks of the Trade" or "Best Practices," although some recommendations are made. Rather, our intent is that these 10 observations will open dialogue in other institutions and departments to collectively address these issues.
2014 ASEE Annual Conference & Exposition Proceedings
As technology continues to advance and competition within the global economy becomes fierce, it i... more As technology continues to advance and competition within the global economy becomes fierce, it is increasingly important that engineering students can not only select the proper equations, perform the correct computations/simulations and build circuits correctly, but also possess an appreciation for the variety of knowledge areas within their field of study. In an effort to better prepare electrical and computer engineering students, the Electrical and Computer Engineering (ECE) Department at the United States Air Force Academy created a breadth-first introductory course to give students this view as a starting point in their education. A thorough review of the curriculum revealed primary knowledge areas that the students need early in their education in order to better prepare them for the depth of a rigorous ECE curriculum. This knowledge includes, but is not limited to, Radio Frequency (RF) communications, RADAR and electronic warfare, analog circuits including power generation and distribution and digital circuits and systems. These topics were selected due to their extensive use in senior capstone projects and needs the industry of the program constituents? The solution proposed here is to create a breadth-first introductory course to motivate and inspire the students to dig deeper into topics they will see later in the curriculum. Through early exposure to a broad set of knowledge and simulation/laboratory techniques, students can begin to develop intellectual curiosity and intuition about how electrical and computer systems work and, in the process, see the fun and excitement in electrical and computer engineering. This paper delves into the development of the course, from the determination of the goals through the implementation of the course structure and teaching philosophy. The paper concludes with an analysis of student feedback.
Colloids and Surfaces B: Biointerfaces, 2020
This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Journal of Vascular and Interventional Radiology, 2020
GE - Portuguese Journal of Gastroenterology, 2019
Introduction: Brucellosis is a zoonotic disease that can involve different organs and tissues. Fe... more Introduction: Brucellosis is a zoonotic disease that can involve different organs and tissues. Fever, fatigue, lymphadenopathy, hepatosplenomegaly, cytopenia, and arthritis are the usual modes of presentation. Gastrointestinal manifestations of human brucellosis are common but documented ileal involvement is extremely rare. Case Description: A 68-year-old female presented with a history of 10 days of intense temporal migraine, photophobia, and phonophobia with partial response to paracetamol. The patient referred night sweats, anorexia, and colicky abdominal pain after her meals for the past 4 months followed by diarrhea. She denied nausea, vomiting, hypersensitivity of the scalp, blurry vision, melena, or rectal bleeding. She denied travelling or contact with animals. Physical examination revealed fever (38.3°C) and splenomegaly. Laboratory workup revealed Hb 7.8 g/dL, leukopenia (3.47 × 109/L), C-reactive protein 5.94 mg/dL, and erythrocyte sedimentation rate 23 mm/h. Abdominal ul...
Blucher Biophysics Proceedings, 2019
AGRARIAN ACADEMY
RESUMO Uma das estratégias elementares para o sucesso de empreendimentos silviculturais é a defin... more RESUMO Uma das estratégias elementares para o sucesso de empreendimentos silviculturais é a definição dos espaçamentos de plantio. Assim, este trabalho teve como objetivo a análise do desenvolvimento dendrométrico sobre o crescimento de plantas de mogno africano (Khaya grandifoliola C. Dc.) cultivadas em cinco diferentes espaçamentos (3x3 m, 4x4 m, 5x5 m, 6x6 m e 6x7 m). O plantio foi estabelecido em 2015 em Bonfinópolis/GO. A avaliação dendrométrica deu-se pela medição do diâmetro à altura do peito (DAP), por meio de suta e medição da altura total através do clinômetro digital. Foram medidas 78 árvores em cada espaçamento e determinada a média para o povoamento de cada espaçamento. O volume médio individual para cada espaçamento foi determinado por meio das variáveis diâmetro e altura, estimando por meio de modelo matemático ajustado para a região. Como o esperado, maiores médias de DAP foram observadas nos espaçamentos mais amplos (5x5 m, 6x6 m e 6x7 m). Para as variáveis altura e volume individual, foi constatada superioridade nos valores do espaçamento 5x5 m. Em relação à análise de desenvolvimento dendrométrico, mesmo com o espaçamento 5x5 m apresentando melhores médias de DAP, altura e volume individual, não é possível indicá-lo como melhor escolha de espaçamento, já que os mais adensados (3x3 m e 4x4 m) não estavam sob condições ideais de manejo, pois não sofreram desbaste.
The American Journal of Surgery
2022 52nd Annual IEEE/IFIP International Conference on Dependable Systems and Networks Workshops (DSN-W)
Defending computer networks from cyber attack requires timely responses to alerts and threat inte... more Defending computer networks from cyber attack requires timely responses to alerts and threat intelligence. Decisions about how to respond involve coordinating actions across multiple nodes based on imperfect indicators of compromise while minimizing disruptions to network operations. Currently, playbooks are used to automate portions of a response process, but often leave complex decision-making to a human analyst. In this work, we present a deep reinforcement learning approach to autonomous response and recovery in large industrial control networks. We propose an attention-based neural architecture that is flexible to the size of the network under protection. To train and evaluate the autonomous defender agent, we present an industrial control network simulation environment suitable for reinforcement learning. Experiments show that the learned agent can effectively mitigate advanced attacks that progress with few observable signals over several months before execution. The proposed deep reinforcement learning approach outperforms a fully automated playbook method in simulation, taking less disruptive actions while also defending more nodes on the network. The learned policy is also more robust to changes in attacker behavior than playbook approaches.
2022 ACM/IEEE 13th International Conference on Cyber-Physical Systems (ICCPS)
Granulocyte colony-stimulating factor (G-CSF) is increasingly being used to prevent febrile neutr... more Granulocyte colony-stimulating factor (G-CSF) is increasingly being used to prevent febrile neutropenia associated with chemotherapy. Large-vessel vasculitis (LVV) has been recognized as a rare side effect of G-CSF treatment. We report a case of G-CSF associated LVV in a patient with breast cancer. While clear pathogenic mechanisms remain unknown, G-CSF may cause vasculitis due to inflammatory cytokines production. This adverse reaction should be recognized in patients with suggestive symptoms following the administration of pegfilgrastim. A 56-year-old woman with luminal B breast cancer who had undergone surgery and adjuvant chemotherapy, initially with paclitaxel, was started on a doxorubicin plus cyclophosphamide protocol, followed by supportive use of long-acting G-CSF pegfilgrastim. Following the administration of pegfilgrastim, the patient developed intermittent fever and was given empiric antibiotics in the outpatient setting with no improvement. There were no signs of cancer...
International Journal of Colorectal Disease, 2021
Preoperative neoadjuvant chemotherapy (NAC) for colon cancer in advanced stages is associated wit... more Preoperative neoadjuvant chemotherapy (NAC) for colon cancer in advanced stages is associated with improved outcomes and tumor regression. The aim of our study was to identify outcomes in patients with colon cancer who received preoperative NAC. A 4-year analysis of the American College of Surgeons-National Surgical Quality Improvement Program (ACS-NSQIP) was performed. We included patients with locally advanced colon cancer (non-metastatic T3, T4 with or without nodal involvement) who underwent colon cancer resection. Patients were stratified into two groups (NAC and No-NAC). Our outcome measures were anastomotic leaks, hospital length of stay (LOS), 30-day complications, 30-day mortality, and 30-day readmissions. We performed a multi-variable logistic regression analysis. We analyzed 90,055 patients of which 7694 (NAC: 7.8 % (n = 599) and No-NAC: 92.2% (n = 7095)) met the inclusion criteria and included in the analysis. Mean age was 67 ± 13 years, 51% were males, and 72% were whites. Patients in the NAC group were more likely to be younger (60 ± 12 years vs. 68 ± 13 years, p < 0.01) and males (62% vs. 50%, p < 0.01) compared to No-NAC. On regression analysis, preoperative NAC was independently associated with higher odds of anastomotic leak (OR 1.35 [1.05–1.97], p = 0.03) and 30-day readmission (OR 1.54 [1.24–2.05], p < 0.01) in reference to No-NAC. However, no association was found between NAC and 30-day complications and 30-day mortality. Preoperative NAC might be associated with adverse outcomes of anastomotic leaks and 30-day readmissions, however does not appear to impact 30-day complications nor 30-day mortality.
Recent advancements in multilayer, multicellular, genetic logic circuits often rely on manual int... more Recent advancements in multilayer, multicellular, genetic logic circuits often rely on manual intervention throughout the computation cycle and orthogonal signals for each chemical "wire". These constraints can prevent genetic circuits from scaling. Microfluidic devices can be used to mitigate these constraints. However, continuousflow microfluidics are largely designed through artisanal processes involving handdrawing features and accomplishing design rule checks visually: processes that are also inextensible. Additionally, continuous-flow microfluidic routing is only a consideration during chip design and, once built, the routing structure becomes "frozen in silicon," or for many microfluidic chips "frozen in polydimethylsiloxane (PDMS)"; any changes to fluid routing often require an entirely new device and control infrastructure. The cost of fabricating and controlling a new device is high in terms of time and money; attempts to reduce one cost measure are, generally, paid through increases in the vi other. This work has three main thrusts: to create a microfluidic fabrication framework, called MakerFluidics, that lowers the barrier to entry for designing and fabricating microfluidics in a manner amenable to automation (Chapter 3); to prove this methodology can design, fabricate, and control complex and novel microfluidic devices (Chapter 4); and to demonstrate the methodology can be used to solve biologically-relevant problems (Chapter 5). Utilizing accessible technologies, rapid prototyping, and scalable design practices, the MakerFluidics framework has demonstrated its ability to design, fabricate and control novel, complex and scalable microfludic devices. This was proven through the development of a reconfigurable, continuous-flow routing fabric driven by a modular, scalable primitive called a transposer. In addition to creating complex microfluidic networks, MakerFluidics was deployed in support of cutting-edge, application-focused research at the Charles Stark Draper Laboratory. Informed by a design of experiments approach using the parametric rapid prototyping capabilities made possible by Mak-erFluidics, a plastic blood-bacteria separation device was optimized, demonstrating that the new device geometry can separate bacteria from blood while operating at 275% greater flow rate as well as reduce the power requirement by 82% for equivalent separation performance when compared to the state of the art. Ultimately, MakerFluidics demonstrated the ability to design, fabricate, and control complex and practical microfluidic devices while lowering the barrier to entry to continuous-flow microfluidics, thus democratizing cutting edge technology beyond a handful of well-resourced and specialized labs. vii A Othermill Standard Operating Procedure 92 References 98
Digital electronic circuits have inspired synthetic biologists to program living cells with synth... more Digital electronic circuits have inspired synthetic biologists to program living cells with synthetic decision-mak- ing circuits by creating multilevel genetic logic gates. In both genetic and electronic circuits, logic synthesis translates an abstract functional description into a representation that can be physically implemented. However, inherent differences in genetic logic devices present new synthesis challenges. Here we provide a starting point for a growing set of biodesign au- tomation tools to tackle this challenge in a unified way. This framework will enable direct comparison of new approaches, results that are transferable and standardized, and research into independent areas of the problem formulation with a clear path toward future integration.
ArXiv, 2021
Defending computer networks from cyber attack requires coordinating actions across multiple nodes... more Defending computer networks from cyber attack requires coordinating actions across multiple nodes based on imperfect indicators of compromise while minimizing disruptions to network operations. Advanced attacks can progress with few observable signals over several months before execution. The resulting sequential decision problem has large observation and action spaces and a long time-horizon, making it difficult to solve with existing methods. In this work, we present techniques to scale deep reinforcement learning to solve the cyber security orchestration problem for large industrial control networks. We propose a novel attention-based neural architecture with size complexity that is invariant to the size of the network under protection. A pre-training curriculum is presented to overcome early exploration difficulty. Experiments show in that the proposed approaches greatly improve both the learning sample complexity and converged policy performance over baseline methods in simulat...
Angewandte Chemie International Edition, 2021
Detection of pathogenic bacteria in complex biological matrices remains a major challenge. Herein... more Detection of pathogenic bacteria in complex biological matrices remains a major challenge. Herein, we report the selection and optimization of a new DNAzyme for Staphylococcus aureus (SA) and the use of the DNAzyme to develop a simple lateral flow device (LFD) for detection of SA in nasal mucus. The DNAzyme was generated by in vitro selection using a crude extra/intracellular mixture derived from SA, which could be used directly for simple solution or paper‐based fluorescence assays for SA. The DNAzyme was further modified to produce a DNA cleavage fragment that acted as a bridging element to bind DNA‐modified gold nanoparticles to the test line of a LFD, producing a simple colorimetric dipstick test. The LFD was evaluated with nasal mucus samples spiked with SA, and demonstrated that SA detection was possible in minutes with minimal sample processing.
The NIST-hosted Synthetic Biology Standards Consortium (SBSC) will collectively build the infrast... more The NIST-hosted Synthetic Biology Standards Consortium (SBSC) will collectively build the infrastructure to support a fully integrated global synthetic biology enterprise. We aim to accomplish this by developing metrology products – standards, including reference materials, reference data, reference methods, and documentary standards – that will enable coordination of labor and reuse of materials. We will present the results of the kick-off workshop for the SBSC, held on March 31, 2015 at Stanford University. A summary of the plans developed by each working group will be shared, and mechanisms of future consortium operations will be discussed.
Giant-Cell Arteritis [Working Title], 2021
Giant cell arteritis (GCA), also known as temporal arteritis or Horton disease, is categorized as... more Giant cell arteritis (GCA), also known as temporal arteritis or Horton disease, is categorized as a large- and medium-sized vessels vasculitis. Systemic symptoms are common in GCA and although vascular involvement may be widespread, the cranial branches of the aortic arch are responsible for the hallmark symptoms of GCA: headache, jaw claudication and ocular symptoms, particularly visual loss. The large vessel (LV)-GCA phenotype may differ or overlap from cranial arteritis. Clinical consequences of LV-GCA comprise aneurysms and dissections of the aorta, as well as stenosis, occlusion and ectasia of large arteries. Symptoms of polymyalgia rheumatica occurring in a patient with GCA include characteristic proximal polyarthralgias and myalgias, sometimes accompanied by remitting seronegative symmetrical synovitis with pitting edema (RS3PE), Less common manifestations reported include central nervous system involvement, audiovestibular and upper respiratory symptoms, pericarditis, mesent...
International Journal of Colorectal Disease, 2021
Purpose Brain metastasis (BM) in colorectal cancer patients is rare and is associated with dismal... more Purpose Brain metastasis (BM) in colorectal cancer patients is rare and is associated with dismal outcomes. Our study aims to evaluate the incidence and predictors of BM in patients with colorectal cancer. Methods We performed a retrospective analysis (2010–2017) of patients with a primary diagnosis of colorectal cancer (CRC). Patients were stratified into two groups (BM vs. No-BM). Outcome measures were the incidence and predictors of BM. Multivariate logistic regression analysis was performed. Results A total of 230,806 patients were analyzed. A total of 0.30% ( n = 691) of the patients were found to have BM. On multivariate logistics regression, bone (OR: 5.39 [3.36–8.65], p < 0.001), lung (OR: 3.75 [2.67–5.28], < 0.001), and distant node metastasis (OR: 32.75 [20.47–52.41], p < 0.001) were independent predictors of BM. Conclusion Our study supports the low incidence of brain metastasis in patients with colorectal cancer. A unique set of characteristics is identified to confer an increased risk of brain metastases.
Journal of the American College of Surgeons, 2020
2012 ASEE Annual Conference & Exposition Proceedings
In June 2010, we reported to the United States Air Force Academy to teach in the Department of El... more In June 2010, we reported to the United States Air Force Academy to teach in the Department of Electrical and Computer Engineering. Each of us brought different skill sets, different experiences, and different interests from different jobs in different locations. Yet, throughout our first year we simultaneously observed common stumbling blocks leading to some dilemmas and curiosities. This paper presents some of these common observations from three dissimilar instructors within a contextual framework promoting a learning-centered paradigm and balancing technical syllabus content with real-world "soft" skills. It is not our intention to define these observations as "Tricks of the Trade" or "Best Practices," although some recommendations are made. Rather, our intent is that these 10 observations will open dialogue in other institutions and departments to collectively address these issues.
2014 ASEE Annual Conference & Exposition Proceedings
As technology continues to advance and competition within the global economy becomes fierce, it i... more As technology continues to advance and competition within the global economy becomes fierce, it is increasingly important that engineering students can not only select the proper equations, perform the correct computations/simulations and build circuits correctly, but also possess an appreciation for the variety of knowledge areas within their field of study. In an effort to better prepare electrical and computer engineering students, the Electrical and Computer Engineering (ECE) Department at the United States Air Force Academy created a breadth-first introductory course to give students this view as a starting point in their education. A thorough review of the curriculum revealed primary knowledge areas that the students need early in their education in order to better prepare them for the depth of a rigorous ECE curriculum. This knowledge includes, but is not limited to, Radio Frequency (RF) communications, RADAR and electronic warfare, analog circuits including power generation and distribution and digital circuits and systems. These topics were selected due to their extensive use in senior capstone projects and needs the industry of the program constituents? The solution proposed here is to create a breadth-first introductory course to motivate and inspire the students to dig deeper into topics they will see later in the curriculum. Through early exposure to a broad set of knowledge and simulation/laboratory techniques, students can begin to develop intellectual curiosity and intuition about how electrical and computer systems work and, in the process, see the fun and excitement in electrical and computer engineering. This paper delves into the development of the course, from the determination of the goals through the implementation of the course structure and teaching philosophy. The paper concludes with an analysis of student feedback.
Colloids and Surfaces B: Biointerfaces, 2020
This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Journal of Vascular and Interventional Radiology, 2020
GE - Portuguese Journal of Gastroenterology, 2019
Introduction: Brucellosis is a zoonotic disease that can involve different organs and tissues. Fe... more Introduction: Brucellosis is a zoonotic disease that can involve different organs and tissues. Fever, fatigue, lymphadenopathy, hepatosplenomegaly, cytopenia, and arthritis are the usual modes of presentation. Gastrointestinal manifestations of human brucellosis are common but documented ileal involvement is extremely rare. Case Description: A 68-year-old female presented with a history of 10 days of intense temporal migraine, photophobia, and phonophobia with partial response to paracetamol. The patient referred night sweats, anorexia, and colicky abdominal pain after her meals for the past 4 months followed by diarrhea. She denied nausea, vomiting, hypersensitivity of the scalp, blurry vision, melena, or rectal bleeding. She denied travelling or contact with animals. Physical examination revealed fever (38.3°C) and splenomegaly. Laboratory workup revealed Hb 7.8 g/dL, leukopenia (3.47 × 109/L), C-reactive protein 5.94 mg/dL, and erythrocyte sedimentation rate 23 mm/h. Abdominal ul...
Blucher Biophysics Proceedings, 2019