Novel AI to avert the mental health crisis in COVID-19: Novel application of GPT2 in Cognitive Behaviour Therapy

The effect of the COVID-19 pandemic on mental health is substantial. The World Health Organization has called for action to avert an impending mental health crisis. To respond to this call, this paper contributes a novel application of Deep Learning in Natural Language Generation (NLG) to seed healthy thoughts for mental health therapy. For the 1st time in literature, a transfer learning capable large neural network with more than 100 million parameters for a NLG based mental health therapy application is proposed & demonstrated. This AI is designed to address scalable impact for millions of families with a timely health intervention in a privacy-safe approach. To the best of our knowledge, this is the first research paper to apply GPT2 (Generative Pretrained Transformer) for Cognitive Behavior therapy (CBT). Further, the paper demonstrates the proposed neural network architecture with a lab prototype implementation with reproducible results. This paper demonstrates this AI’s ability to generate conditional synthetic human-like text intended to seed a healthy mental outlook. This is accomplished by fine tuning a pre-trained GPT2 language model. The source code and video demonstration is contributed at https://sites.google.com/view/ai-in-mental-health.Also, for the 1st time in literature, a novel idea of NLU (Natural Language Understanding) activated NLG therapy is demonstrated with reproducible results using a BERT based classifier to activate the GPT2 based therapy. Performance of GPT2 models of three different sizes (124, 355, 774 million parameters) was the same for a very small dataset, thus a small GPT2 model is suggested for on-device AI inference. This AI is a step forward in responding to WHO’s call for action to avert the crisis. Towards addressing all the three dimensions of the monumental challenge, the paper designed a novel AI architecture by taking advantage of both BERT & GPT2. It also demonstrated the feasibility of Transformers-based AI for developing a mental health therapy solution. Further, this paper contributed an open-source AI prototype to support research communities to transform global mental wellness.

Endobronchial hamartoma in a young COVID-19 symptomatic woman. Radical endoscopic treatment with a disposable bronchoscope. Case Report

Pulmonary hamartomas represent the most frequent family of benign lung tumors that typically involve the lung parenchyma and only rarely grow as endobronchial tumors. The elective treatment of endobronchial hamartoma is the bronchoscopic resection, and in those cases in which tumor extension and localization makes it not possible, surgical treatment must be evaluated. Patients with symptomatic COVID-19, hospitalized, frequently undergo a chest CT scan and in some cases, occasional findings may emerge, requiring diagnostic investigations such as bronchoscopy and interventional pulmonology procedures. Therefore, in such a delicate pathological condition, such as COVID-19, the need to perform bronchoscopy and interventional pulmonology procedures, minimizing the risk of viral transmission and ensuring necessary assistance, represents a great challenge for pulmonologists. In this article authors describe, for the first time in literature, a rare case of endobronchial hamartoma, radically resected using a single use bronchoscope, in a young female patient hospitalized for symptomatic COVID-19.

Development of generalized tool life model for constant and variable speed turning

In this research, a generalized tool life modelling for considering non-stationary cutting conditions was developed . In particular, for the first time in literature, the model was conceived for predicting the life of the tool when spindle speed variation SSV, one of the most effective techniques for suppressing regenerative chatter vibrations, is used. The proposed formulation takes into account the main cutting parameters and the parameters associated to the SSV. A dedicated experimental campaign of turning tests was executed and the data were used for modelling purposes. The model validation was carried out performing additional tool life tests. According to the analyzed technological scenario, it was found that the generalized formulation can be used for predicting the tool life both at constant spindle machining CSM and adopting SSV with the maximum estimating error of 6%.

COVID-19 triggered encephalopathic crisis in a patient with glutaric aciduria type 1

Objectives The impact of coronavirus disease-19 (COVID-19) on metabolic outcome in patients with inborn errors of metabolism has rarely been discussed. Herein, we report a case with an acute encephalopathic crisis at the course of COVID-19 disease as the first sign of glutaric aciduria type 1 (GA-1). Case presentation A 9-month-old patient was admitted with encephalopathy and acute loss of acquired motor skills during the course of COVID-19 disease. She had lethargy, hypotonia, and choreoathetoid movements. In terms of COVID-19 encephalopathy, the reverse transcription-polymerase chain reaction assay test for COVID-19 was negative in cerebral spinal fluid. Brain imaging showed frontotemporal atrophy, bilateral subcortical and periventricular white matter, basal ganglia, and thalamic involvement. Elevated glutarylcarnitine in plasma and urinary excretion of glutaric and 3-OH-glutaric acids was noted. A homozygote mutation in the glutaryl-CoA dehydrogenase gene led to the diagnosis of GA-1. Conclusions With this report, neurological damage associated with COVID-19 has been reported in GA-1 patients for the first time in literature.

An evaluation of the efficacy of a four-grade fat infiltration classification method, presented for the first time in literature

BackgroundThe aim of this study was to evaluate the efficacy of a semi-quantitative simplified 4-grade fat infiltration measurement system, described for the first time in literature, through comparison with the existing simplified 3-grade fat infiltration system in the prediction of lumbar disc herniation.Material and MethodThe study included 39 lumbar disc herniation patients (LDH) and 38 healthy subjects (control), comprising 33 (42.9%) males and 44 (57.1%) females with a mean age of 37 ± 11.3 years (range, 20–64 years). The patients were evaluated in respect of fat infiltration of the right and left lumbar multifidus and erector spina muscles on axial magnetic resonance imaging slices passing through the centre of the disc at L3-S1 level using the 3 and 4-grade fat infiltration measurement systems. The results were compared and the correlations of the results of the two systems with lumbar disc herniation were examined.ResultsThe 3-grade fat infiltration system was found to be insufficient in the prediction of lumbar disc herniation (p > 0.05) and the 4-grade fat infiltration system was determined to be effective in the prediction of lumbar disc herniation (p = 0.003).ConclusionThe 4-grade fat infiltration system was seen to be more effective than the 3-grade fat infiltration system in the determination of the level of fat infiltration in the paraspinal muscles and the prediction of lumbar disc herniation. The 4-grade fat infiltration system is an effective semi-quantitative grading system which can be used instead of the simplified 3-grade system.

DL-LA: Deep Learning Leakage Assessment

In recent years, deep learning has become an attractive ingredient to side-channel analysis (SCA) due to its potential to improve the success probability or enhance the performance of certain frequently executed tasks. One task that is commonly assisted by machine learning techniques is the profiling of a device’s leakage behavior in order to carry out a template attack. At CHES 2019, deep learning has also been applied to non-profiled scenarios for the first time, extending its reach within SCA beyond template attacks. The proposed method, called DDLA, has some tempting advantages over traditional SCA due to merits inherited from (convolutional) neural networks. Most notably, it greatly reduces the need for pre-processing steps< when the SCA traces are misaligned or when the leakage is of a multivariate nature. However, similar to traditional attack scenarios the success of this approach highly depends on the correct choice of a leakage model and the intermediate value to target. In this work we explore, for the first time in literature, whether deep learning can similarly be used as an instrument to advance another crucial (non-profiled) discipline of SCA which is inherently independent of leakage models and targeted intermediates, namely leakage assessment. In fact, given the simple classification-based nature of common leakage assessment techniques, in particular distinguishing two groups fixed-vs-random or fixed-vs-fixed, it comes as a surprise that machine learning has not been brought into this context, yet. Our contribution is the development of the first full leakage assessment methodology based on deep learning. It gives the evaluator the freedom to not worry about location, alignment and statistical order of the leakages and easily covers multivariate and horizontal patterns as well. We test our approach against a number of case studies based on FPGA, ASIC and μC implementations of the PRESENT block cipher, equipped with state-of-the-art SCA countermeasures. Our results clearly show that the proposed methodology and network structures are robust across all case studies and outperform the classical detection approaches (t-test and X2-test) in all considered scenarios.

Systematic Design, Control and Parametric Testing of an Automated Resuscitator Bag Mechanical Ventilator

The COVID-19 crisis has revealed and exacerbated a shortage of mechanical ventilators in hospitals around the world, regardless of their government's resources. Where some countries can respond to the situation by ordering more high-end ventilators, the price is often too high for Low and Middle Income Countries (LMICs) and securing them can be difficult. The goal of this work is to design, prototype, and test a low-cost ventilator based on the automated compression of a resuscitator bag. A holistic and systematic design approach is taken to create a compact and adaptable device that can safely meet the current requirements. This is achieved by using 72% standard parts and prioritizing compactness in the mechanical design. The control system is developed to provide both continuous mandatory ventilation (CMV) and spontaneous breathing support, or Assist Control (AC), which significantly extends the potential use cases beyond patient sedation. The prototype is tested for accuracy, modularity, and oxygen response using a full physiological artificial lung. The results show for the first time in literature that the design operates within the defined requirements, based on emergency government regulations, and can be used with different sizes of resuscitator bags and different positions of the flow sensor. This provides a sound basis for further development of a low-cost, portable mechanical ventilator for potential use in LMICs.

Novel AI to avert the mental health crisis in COVID-19: Novel application of GPT2 in Cognitive Behaviour Therapy

The effect of the COVID-19 pandemic on mental health is substantial. The World Health Organization has called for action to avert an impending mental health crisis. To respond to this call, this paper contributes a novel application of Deep Learning in Natural Language Generation (NLG) to seed healthy thoughts for mental health therapy. For the 1st time in literature, a transfer learning capable large neural network with more than 100 million parameters for a NLG based mental health therapy application is proposed & demonstrated. This AI is designed to address scalable impact for millions of families with a timely health intervention in a privacy-safe approach. To the best of our knowledge, this is the first research paper to apply GPT2 (Generative Pretrained Transformer) for Cognitive Behavior therapy (CBT). Further, the paper demonstrates the proposed neural network architecture with a lab prototype implementation with reproducible results. This paper demonstrates this AI’s ability to generate conditional synthetic human-like text intended to seed a healthy mental outlook. This is accomplished by fine tuning a pre-trained GPT2 language model. The source code and video demonstration is contributed at https://sites.google.com/view/ai-in-mental-health.Also, for the 1st time in literature, a novel idea of NLU (Natural Language Understanding) activated NLG therapy is demonstrated with reproducible results using a BERT based classifier to activate the GPT2 based therapy. Performance of GPT2 models of three different sizes (124, 355, 774 million parameters) was the same for a very small dataset, thus a small GPT2 model is suggested for on-device AI inference. This AI is a step forward in responding to WHO’s call for action to avert the crisis. Towards addressing all the three dimensions of the monumental challenge, the paper designed a novel AI architecture by taking advantage of both BERT & GPT2. It also demonstrated the feasibility of Transformers-based AI for developing a mental health therapy solution. Further, this paper contributed an open-source AI prototype to support research communities to transform global mental wellness.

PEPPSI complexes as potential prodrugs: enzyme inhibition, antioxidant activity, electrochemical characterization, molecular docking analysis

In this study, enzyme inhibition and antioxidant activity analyzes of previously characterized pyridine-enhanced precatalyst preparation stabilization and initiation (PEPPSI)-type Palladium(II) complexes with benzimidazole-type ligands {dichloro[L]pyridine palladium(II), L1: 1-(2-methyl-2-propenyl)-3-[benzylbenzimidazole]-2-ylidene, L2: 1-(2-methyl-2-propenyl)-3-[4-chloro benzylbenzimidazole]-2-ylidene, L3: 1-(2-methyl-2-propenyl)-3-[3-methylbenzylbenzimidazole]-2-ylidene, L4: 1-(2-methyl-2-propenyl)-3-[3,4,5-thrimethoxybenzylbenzimidazole]-2-ylidene, L5: 1-(2-methyl-2-propenyl)-3-[3-naphthylbenzylbenzimidazole]-2-ylidene, L6: 1-(2-methyl-2-propenyl)-3-[anthracen-9-ylmethylbenzimidazole]-2-ylidene} were performed and evaluated as potential drugs for neurodegenerative disorders such as Alzheimer disease and Parkinson disease. Inhibition of tyrosinase enzyme of N-heterocyclic carbenes (NHC) complexes was determined for the first time in literature. Chelating activities of the complexes were determined and compared with EDTA. Electrochemical characterization was performed using cyclic voltammetry method. Moreover, global reactivity descriptors and electronic transitions were evaluated by DFT/TDDFT methods and molecular docking interactions with human acetylcholine esterase, human butyrylcholine esterase and oxidoreductase were studied.

Application of Fuzzy AHP and Fuzzy MARCOS Approach for the Evaluation of E-Service Quality in the Airline Industry

Airlines today use e-services extensively for marketing activities and the distribution of services. Monitoring and evaluating e-service quality are essential for customers’ satisfaction and thus the success of airlines. This study aims to evaluate e-service quality in the airline industry from the point of view of the consumers. To achieve this, an integrated Fuzzy Analytical Hierarchy Process (F-AHP) and Fuzzy Measurement Alternatives and Ranking according to Compromise Solution (F-MARCOS) approach was proposed to handle the uncertain and imprecise nature of e-service evaluation. In the first stage, e-service quality criteria were prioritized using the F-AHP method. Then, a real-world case study was carried out on scheduled airlines to demonstrate the applicability of the proposed approach using the F-MARCOS method, utilizing a total sample of 395 airline passengers in Turkey. As a result, the top three e-service criteria were found as reliability, understandability and security. A three-stage sensitivity analysis was also conducted to examine the credibility and stability of the results. This study is the first study to integrate F-AHP and F-MARCOS methods for the first time in literature.