Explainable Artificial Intelligence for Bias Detection in COVID CT-Scan Classifiers.

PROBLEM: An application of Explainable Artificial Intelligence Methods for COVID CT-Scan classifiers is presented. MOTIVATION: It is possible that classifiers are using spurious artifacts in dataset images to achieve high performances, and such explainable techniques can help identify this issue. AIM: For this purpose, several approaches were used in tandem, in order to create a complete overview of the classificatios. METHODOLOGY: The techniques used included GradCAM, LIME, RISE, Squaregrid, and direct Gradient approaches (Vanilla, Smooth, Integrated). MAIN RESULTS: Among the deep neural networks architectures evaluated for this image classification task, VGG16 was shown to be most affected by biases towards spurious artifacts, while DenseNet was notably more robust against them. Further impacts: Results further show that small differences in validation accuracies can cause drastic changes in explanation heatmaps for DenseNet architectures, indicating that small changes in validation accuracy may have large impacts on the biases learned by the networks. Notably, it is important to notice that the strong performance metrics achieved by all these networks (Accuracy, F1 score, AUC all in the 80 to 90% range) could give users the erroneous impression that there is no bias. However, the analysis of the explanation heatmaps highlights the bias.

Immunology of chronic BVDV infections.

Bovine viral diarrhea virus can maintain prolonged infections within immunoprivileged sites after an otherwise transient infection of a cow, calf, or bull. Various sites provide unique niches for viral replication which are not susceptible to the complete surveillance commonly provided by the bovine immune system. Evidence indicates that pestiviral infections may be significantly prolonged within ovarian tissue, testicular tissue, central nervous system tissue, and circulating white blood cells. Within avascular portions of the ovarian follicle, granulosa cells and oocytes may maintain BVDV infections which cannot be attacked by cell-mediated immunity. When infections occur within seminiferous tubules in testicular tissue, similar protection from the immune system is provided for BVDV by the blood-testes barrier. Likewise, the blood-brain barrier has been hypothesized to provide protection for BVDV in a case involving neuropathology associated with immunohistochemical detection of BVDV. Furthermore, infections of circulating white blood cells may perturb their stimulation of an adaptive immune response and facilitate chronic infection of these cells. Thus, BVDV has demonstrated an ability to maintain prolonged viral infections in immunoprivileged sites within its natural host. The role of chronic infections in maintaining and disseminating BVDV within the cattle population and heterologous host species remains to be fully understood.

Cluster of Legionnaires disease cases caused by Legionella longbeachae serogroup 1, Scotland, August to September 2013.

We report six confirmed cases of Legionnaires' disease in Scotland caused by Legionella longbeachae serogroup 1, identified over a four-week period in August­September 2013. All cases required admission to hospital intensive care facilities. All cases were amateur gardeners with frequent exposure to horticultural growing media throughout their incubation period. L. longbeachae was identified in five samples of growing media linked to five cases. Product tracing did not identify a common product or manufacturer.

Virtual assessment of coronal balance prior to bone resection with the MAKO robotic-assisted system accurately predicts final balance in TKA.

Total knee arthroplasty (TKA) has traditionally relied on the surgeon's judgement and manual instruments to determine balance. The MAKO robotic system (Stryker Ltd, Kalamazoo, MI, USA) allows assessment of virtual compartmental gaps from CT-derived bone models intra-operatively as a predictor of soft tissue balance that will be achieved, prior to any bony resection. This study aims to assess the accuracy of this pre-resection balancing technique in determining the resultant final soft tissue balance of the TKA. A consecutive prospective cohort of 2027 Robotic-Assisted TKAs (RATKA) were performed between January'17 and March'22. Osteophytes were removed; initial virtual gaps on the virtual bone model were measured at 10° and 90° of flexion prior to bone resections. Optimization of the virtual component positions was then made and final pre-resection gaps were measured. The gaps were then re-assessed post-implantation and compared to the final pre-resection values. Virtual balancing in extension within 1 mm was achieved in 95% of cases. Of those, 98% maintained coronal balance within 1 mm after implantation, with 1.5% requiring a coronal plane soft tissue release. Inability to virtually balance a TKA within 2 mm prior to bone resection resulted in a soft tissue release in 44.4% of cases. The absolute values of the final gaps achieved were a mean of 1.3 mm greater than virtual gaps. The ability to balance a knee on the virtual bone model prior to bone resection, in conjunction with robotic-assisted execution of TKA, consistently achieves a balanced knee after component implantation.

Evolved explainable classifications for lymph node metastases.

A novel evolutionary approach for Explainable Artificial Intelligence is presented: the "Evolved Explanations" model (EvEx). This methodology combines Local Interpretable Model Agnostic Explanations (LIME) with Multi-Objective Genetic Algorithms to allow for automated segmentation parameter tuning in image classification tasks. In this case, the dataset studied is Patch-Camelyon, comprised of patches from pathology whole slide images. A publicly available Convolutional Neural Network (CNN) was trained on this dataset to provide a binary classification for presence/absence of lymph node metastatic tissue. In turn, the classifications are explained by means of evolving segmentations, seeking to optimize three evaluation goals simultaneously. The final explanation is computed as the mean of all explanations generated by Pareto front individuals, evolved by the developed genetic algorithm. To enhance reproducibility and traceability of the explanations, each of them was generated from several different seeds, randomly chosen. The observed results show remarkable agreement between different seeds. Despite the stochastic nature of LIME explanations, regions of high explanation weights proved to have good agreement in the heat maps, as computed by pixel-wise relative standard deviations. The found heat maps coincide with expert medical segmentations, which demonstrates that this methodology can find high quality explanations (according to the evaluation metrics), with the novel advantage of automated parameter fine tuning. These results give additional insight into the inner workings of neural network black box decision making for medical data.

[Current status of HIV self-testing application].

HIV self-testing is a key measure and strategy to expand the coverage of HIV testing in key populations, which has great value in application and high acceptance in key populations. The World Health Organization (WHO) issued the Guidelines on HIV self-testing and partner notification in 2016 to support the application of HIV self-testing and provide specific guidance for countries and organizations to develop corresponding public health policies for the purpose of improving the accessibility and utilization of HIV diagnostics. HIV self-testing has been promoted in many countries and has achieved great results. This article provides an overview of the current domestic and international HIV self-testing policies, intervention models, the use of self-testing kits, the accuracy of self-testing kits, and the application of pre- and post-testing counseling services to provide reference for the further improvement of HIV self-testing in China.

The dusty atmosphere of the brown dwarf Gliese 229B.

The brown dwarf Gliese 229B has an observable atmosphere too warm to contain ice clouds like those on Jupiter and too cool to contain silicate clouds like those on low-mass stars. These unique conditions permit visibility to higher pressures than possible in cool stars or planets. Gliese 229B's 0.85- to 1.0-micrometer spectrum indicates particulates deep in the atmosphere (10 to 50 bars) having optical properties of neither ice nor silicates. Their reddish color suggests an organic composition characteristic of aerosols in planetary stratospheres. The particles' mass fraction (10(-7)) agrees with a photochemical origin caused by incident radiation from the primary star and suggests the occurrence of processes native to planetary stratospheres.

Atmospheric, evolutionary, and spectral models of the brown dwarf Gliese 229 B.

Theoretical spectra and evolutionary models that span the giant planet-brown dwarf continuum have been computed based on the recent discovery of the brown dwarf Gliese 229 B. A flux enhancement in the 4- to 5-micrometer wavelength window is a universal feature from jovian planets to brown dwarfs. Model results confirm the existence of methane and water in the spectrum of Gliese 229 B and indicate that its mass is 30 to 55 jovian masses. Although these calculations focus on Gliese 229 B, they are also meant to guide future searches for extrasolar giant planets and brown dwarfs.

Bovine viral diarrhea virus is inactivated when whole milk from persistently infected cows is heated to prepare semen extender.

Bovine viral diarrhea virus (BVDV) can be present in cryopreserved bovine semen and be transmitted through artificial insemination. Because BVDV can be shed in milk, the virus might also be introduced as a contaminant of milk-based semen extenders. Thus, the purpose of this study was to evaluate the epidemiologic risk of using heated, BVDV-contaminated milk to prepare semen extender. Milk was obtained from cows free of and persistently infected (PI) with BVDV. Six replicates of milk samples were processed by heating (85-92.2 degrees C, 10min). Samples of milk collected before and after heating were assayed for BVDV. Additionally, milk was injected intravenously into eight BVDV seronegative calves to monitor for seroconversion and viral infection. Virus was not detected in any milk samples from negative animals. Virus was consistently isolated from unheated milk samples from PI cows by passage of somatic cells, ultracentrifugation, and animal inoculation. Virus was usually detected in these samples by RT-nPCR (reverse transcription nested polymerase chain reaction). In heated milk samples from PI cows, no infectious BVDV was detected using any technique, but viral RNA was detected using RT-nPCR in four of six replicates. Bovine viral diarrhea virus in milk from PI cows was inactivated by heating. Therefore, properly heated milk used in semen extenders will not result in transmission of infectious BVDV. Although RT-nPCR detected the presence of viral RNA in milk samples after heating, the virus was not infectious as demonstrated by lack of replication despite using multiple sensitive techniques.

Bovine viral diarrhea virus (BVDV) associated with single in vivo-derived and in vitro-produced preimplantation bovine embryos following artificial exposure.

The objective was to determine the average amount of bovine viral diarrhea virus (BVDV) associated with single in vivo-derived and in vitro-produced bovine embryos following recommended processing procedures for embryos. In vivo-derived and in vitro-produced bovine embryos at 7d post-fertilization were exposed (for 2h) to 2 x 10(5-7) cell culture infective dose (CCID(50))/mL of SD-1 (a noncytopathic, Type 1a strain of BVDV), and then washed according to International Embryo Transfer Society (IETS) guidelines prior to testing. Of the 87 in vivo-derived embryos tested, 27% were positive for virus by quantitative polymerase chain reaction (qPCR). The range in amount of virus associated with 99% of the contaminated embryos was

Amplification of bovine viral diarrhoea virus introduced into an in vitro embryo production system via oocytes from persistently infected cattle.

The purpose of this study was to determine whether or not embryos derived from in vitro fertilization of oocytes from persistently infected (PI) cattle would contain infectious virus.Three in vitro embryo production treatment groups were assessed: 1) oocytes and uterine tubal cells (UTC) free of bovine viral diarrhoea virus (BVDV) (negative control), 2)oocytes free of BVDV fertilized and cultured in media containing UTC obtained from PI heifers, and 3) oocytes from PI heifers fertilized and cultured in media containing UTC free of BVDV. The developmental media, UTC and embryos (individual or groups of five) were assayed for virus.Virus was not isolated from any samples in treatment group 1.As shown in previous studies, a proportion of embryo samples were positive for BVDV in treatment group 2. In treatment group 3, the virus associated with the oocytes contaminated the developmental media and infected susceptible co-culture cells used during fertilization and culture. In addition, 65% (11/17) of the degenerated ova from treatment group 3 had infectious virus associated with them. While none of the ova developed into transferable embryos, the study did confirm that use of oocytes from PI cows could lead to amplification of BVDV and cross contamination during in vitro embryo production.

Infectious causes of embryonic and fetal mortality.

The purpose of this review is to summarize bacterial, fungal, protozoan, and viral causes of reproductive dysgenesis in cattle, sheep, goats, pigs, horses, dogs, and cats. The clinical presentations of disease due to reproductive pathogens are emphasized, with a focus on assisting development of complete lists of causes that result in abortion and infertility in these species. Clinicians are encouraged to assess clinical presentation, create complete lists of differential diagnoses, obtain appropriate diagnostic samples, maximize diagnostic laboratory support, and avoid zoonotic infections resulting from reproductive pathogens of animals. The foundation of an accurate diagnosis of reproductive loss due to infectious pathogens facilitates the prudent use of immunization and biosecurity to minimize reproductive losses.

Pathogens that cause infertility of bulls or transmission via semen.

The purpose of this paper is to review scientific evidence regarding pathogens that cause infertility of bulls or that could be transmitted via bovine semen. Although several pathogens can cause male infertility and potentially be transmitted via semen, adhering to disease control recommendations provided by Certified Semen Services (CSS) and the World Organization for Animal Health (OIE) can prevent infectious male infertility and ensure that the risk of pathogen transmission via semen is negligible. Regarding bulls to be used for natural breeding, quarantine prior to herd introduction and appropriate diagnostic testing during quarantine will commonly prevent introduction of pathogens that adversely affect reproduction.

Efficacy of a recombinant trypsin product against bovine herpesvirus 1 associated with in vivo- and in vitro-derived bovine embryos.

Although porcine-origin trypsin will effectively remove bovine herpesvirus 1 (BHV-1) associated with in vivo-derived embryos, TrypLE, a recombinant trypsin-like protease, has not been evaluated. In Experiment 1, 17 groups of 10 in vivo-derived embryos were exposed to BHV-1, treated with TrypLE Express or TrypLE Select (10x concentration) for varying intervals, and assayed as 2 groups of 5 embryos. TrypLE Select treatment for 5 and 10 min (two and seven groups of five embryos, respectively) effectively inactivated BHV-1. In Experiment 2, 22 groups of 10 IVF embryos were treated and assayed. Treatment with TrypLE Select for 7 and 10 min (six groups of five embryos each) and with TrypLE Select diluted 1:2 for 10 min (seven groups of five embryos) was also effective. In Experiment 3, 17 groups of 10 IVF embryos were further evaluated with TrypLE Select undiluted and diluted 1:2 for 10 min. Treatment with the diluted product was effective (18 groups of five embryos), whereas the undiluted product was not completely effective (virus isolated from 2 of 16 groups). In Experiment 4, IVF embryos were treated as described in Experiment 3 and then cultured individually or as groups of five on uterine tubal cells (UTCs) for 48 h; 60% of UTC samples associated with groups of embryos and 35% of UTC associated with individual embryo samples were positive for BHV-1. Therefore, although TrypLE Select appeared to have promise for the treatment of in vivo-derived embryos, it cannot be recommended for treatment of in vitro-derived embryos.

Detection of bovine viral diarrhea virus (BVDV) in single or small groups of preimplantation bovine embryos.

The objectives of this study were to develop techniques to detect BVDV associated with single or small groups of bovine embryos contained in small aliquots of medium using either virus isolation (VI) or real time quantitative polymerase chain reaction (RT-QPCR) assays. In vivo-derived and in vitro-produced bovine embryos at 7 d post-fertilization were exposed to SD-1, a high affinity strain of BVDV, for 2 h and then processed according to the International Embryo Transfer Society (IETS) guidelines prior to testing. Groups of five or two in vivo-derived embryos, and single in vivo-derived embryos, were VI positive for BVDV 100, 50, and 33% of the time, and were RT-QPCR positive 100, 75, and 42% of the time, respectively. The virus was detected by the VI technique in all of the groups of five or two in vitro-produced embryos and in all of the single in vitro-produced embryos, and it was detected in 100, 80, and 50%, using RT-QPCR. Techniques for RT-QPCR were sufficiently sensitive to detect 10 copies of viral RNA in a sample and to detect BVDV associated with single embryos. Application of this new technology, RT-QPCR, will facilitate additional studies to further assess the risk of transmission of BVDV through embryo transfer.

Evolving Spiking Neural Networks for Recognition of Aged Voices.

The aging of the voice, known as presbyphonia, is a natural process that can cause great change in vocal quality of the individual. This is a relevant problem to those people who use their voices professionally, and its early identification can help determine a suitable treatment to avoid its progress or even to eliminate the problem. This work focuses on the development of a new model for the identification of aging voices (independently of their chronological age), using as input attributes parameters extracted from the voice and glottal signals. The proposed model, named Quantum binary-real evolving Spiking Neural Network (QbrSNN), is based on spiking neural networks (SNNs), with an unsupervised training algorithm, and a Quantum-Inspired Evolutionary Algorithm that automatically determines the most relevant attributes and the optimal parameters that configure the SNN. The QbrSNN model was evaluated in a database composed of 120 records, containing samples from three groups of speakers. The results obtained indicate that the proposed model provides better accuracy than other approaches, with fewer input attributes.

Zones, spots, and planetary-scale waves beating in brown dwarf atmospheres.

Brown dwarfs are massive analogs of extrasolar giant planets and may host types of atmospheric circulation not seen in the solar system. We analyzed a long-term Spitzer Space Telescope infrared monitoring campaign of brown dwarfs to constrain cloud cover variations over a total of 192 rotations. The infrared brightness evolution is dominated by beat patterns caused by planetary-scale wave pairs and by a small number of bright spots. The beating waves have similar amplitudes but slightly different apparent periods because of differing velocities or directions. The power spectrum of intermediate-temperature brown dwarfs resembles that of Neptune, indicating the presence of zonal temperature and wind speed variations. Our findings explain three previously puzzling behaviors seen in brown dwarf brightness variations.

Use of a modified-live vaccine to prevent persistent testicular infection with bovine viral diarrhea virus.

A commercial vaccine containing modified-live bovine viral diarrhea virus (BVDV; types 1 and 2) was administered to one group of 22 peripubertal bulls 28 days before intranasal inoculation with a type 1 strain of BVDV. A second group of 23 peripubertal bulls did not receive the modified-live BVDV vaccine before intranasal inoculation. Ten of 23 unvaccinated bulls--but none of the vaccinated bulls--developed a persistent testicular infection as determined by immunohistochemistry and polymerase chain reaction. Results of this study indicate that administration of a modified-live vaccine containing BVDV can prevent persistent testicular infection if peripubertal bulls are vaccinated before viral exposure.

ON THE COMPOSITION OF YOUNG, DIRECTLY IMAGED GIANT PLANETS.

The past decade has seen significant progress on the direct detection and characterization of young, self-luminous giant planets at wide orbital separations from their host stars. Some of these planets show evidence for disequilibrium processes like transport-induced quenching in their atmospheres; photochemistry may also be important, despite the large orbital distances. These disequilibrium chemical processes can alter the expected composition, spectral behavior, thermal structure, and cooling history of the planets, and can potentially confuse determinations of bulk elemental ratios, which provide important insights into planet-formation mechanisms. Using a thermo/photochemical kinetics and transport model, we investigate the extent to which disequilibrium chemistry affects the composition and spectra of directly imaged giant exoplanets. Results for specific "young Jupiters" such as HR 8799 b and 51 Eri b are presented, as are general trends as a function of planetary effective temperature, surface gravity, incident ultraviolet flux, and strength of deep atmospheric convection. We find that quenching is very important on young Jupiters, leading to CO/CH4 and N2/NH3 ratios much greater than, and H2O mixing ratios a factor of a few less than, chemical-equilibrium predictions. Photochemistry can also be important on such planets, with CO2 and HCN being key photochemical products. Carbon dioxide becomes a major constituent when stratospheric temperatures are low and recycling of water via the H2 + OH reaction becomes kinetically stifled. Young Jupiters with effective temperatures ≲700 K are in a particularly interesting photochemical regime that differs from both transiting hot Jupiters and our own solar-system giant planets.

Analysis and Classification of Voice Pathologies Using Glottal Signal Parameters.

The classification of voice diseases has many applications in health, in diseases treatment, and in the design of new medical equipment for helping doctors in diagnosing pathologies related to the voice. This work uses the parameters of the glottal signal to help the identification of two types of voice disorders related to the pathologies of the vocal folds: nodule and unilateral paralysis. The parameters of the glottal signal are obtained through a known inverse filtering method, and they are used as inputs to an Artificial Neural Network, a Support Vector Machine, and also to a Hidden Markov Model, to obtain the classification, and to compare the results, of the voice signals into three different groups: speakers with nodule in the vocal folds; speakers with unilateral paralysis of the vocal folds; and speakers with normal voices, that is, without nodule or unilateral paralysis present in the vocal folds. The database is composed of 248 voice recordings (signals of vowels production) containing samples corresponding to the three groups mentioned. In this study, a larger database was used for the classification when compared with similar studies, and its classification rate is superior to other studies, reaching 97.2%.