Good Metrics Research Articles

Abilities and Disabilities Among Children With Developmental Language Disorder.

The aim of this study was to situate developmental language disorder (DLD) within the impairment and disability framework of the International Classification of Functioning, Disability, and Health (ICF); describe the functional strengths and weaknesses of a cohort of first-grade children with DLD and their peers; and explore the ways that language-related disabilities relate to language impairment, developmental risk, and receipt of language services. We queried the caregivers of 35 children with DLD and 44 peers with typical language development about their children's language-related functions, developmental risks, and language services using mixed quantitative and qualitative methods. The children with DLD presented with weaknesses in domains that are highly dependent upon language skill, including communication, community function, interpersonal relationships, and academics. They presented with strengths in domestic and personal aspects of daily living, play and coping aspects of socialization, and gross motor function. Caregivers of children with DLD expressed pride in their children's agentive and prosocial qualities. Consistent with the ICF, what distinguished children with DLD who had functional weaknesses and disabilities from those who had healthy function was not the severity of language impairment as measured by decontextualized tests of language skill, but the presence of cumulative developmental risks. Compared to those with healthy function, a larger portion of children with weaknesses and disabilities were receiving language services; however, two girls who had disabilities despite mild levels of impairment were without services. Children with DLD present with predictable strengths and weaknesses in everyday language-related functioning. For some children, the weaknesses are mild, but for others, they limit function to a greater extent and should be considered disabilities. The severity of language impairment is not a strong indicator of language-related function and, therefore, is not a good metric for determining service qualification.

Effect of Interleukin-2 Inducible Kinase (ITK) on the Temporal Dynamics of TCR Signaling in CD8 T cells

Abstract Antigen affinity has been linked to the rate of downstream TCR activation and the activation of transcription factors such as NFAT and NF-kb. Interleukin-2 inducible T cell kinase (ITK) plays a key role in the differential rate and magnitude of TCR-dependent transcription factor activation, especially when T cells are stimulated with lower affinity ligands. The Nr4a3-Tocky Timer mouse model carries a construct for a fluorescent protein controlled by the Nr4a3 promoter. Nr4a3 is highly expressed following TCR activation and thus is a good metric for TCR signaling. When first produced, the Tocky protein fluoresces “blue,” at peak wavelength 466nm and then, after ~7hrs, degrades to a “red” fluorescence at peak wavelength 583nm. Splenocytes were harvested from OT-I RAG1KO mice crossed to Nr4a3-Tocky mice and stimulated using varying affinities and doses of ova peptide over 7 days in vitro. Cells were exposed to the ITK inhibitor, PRN694, to observe the effect of ITK on the rate of expression of Tocky fluorescence along with activation surface markers CD69, CD25, CD44, and CD122. With ITK inhibition, CD8 T cells expressed the “Blue” protein slower and in fewer cells compared to OT-I cells stimulated without PRN694 at 4hrs. At 48hrs after stimulation, sustained signaling was observed in control cells, whereas a subset of ITK-inhibited cells showed evidence of halted TCR-induced Tocky transcription. This decrease in the longevity of TCR signaling was more evident when lower peptide doses or lower affinity peptides were used for stimulation. These results suggest that ITK plays a role in amplifying the initial TCR signal, and is also required for sustained signaling after TCR activation in CD8 T cells, resulting in the tuning of early gene expression. Supported by grants from the NIH (Dissecting the pathways controlling tunable responses to TCR signaling: 5R01AI132419-06)

Systematic Study of Solid-State Fluorescence and Molecular Packing of Methoxy-trans-Stilbene Derivatives, Exploration of Weak Intermolecular Interactions Based on Hirshfeld Surface Analysis

In recent years, fluorescent compounds that emit efficiently in the solid state have become particularly interesting, especially those that are easily prepared and inexpensive. Hence, exploring the photophysical properties of stilbene derivatives, supported by a detailed analysis of molecular packing obtained from single-crystal X-ray diffraction data, is a relevant area of research. A complete understanding of the interactions to determine the molecular packing in the crystal lattice and their effect on the material's physicochemical properties is essential to tune various properties effectively. In the present study, we examined a series of methoxy-trans-stilbene analogs with substitution pattern-dependent fluorescence lifetimes between 0.82 and 3.46 ns and a moderate-to-high fluorescence quantum yield of 0.07-0.69. The relationships between the solid-state fluorescence properties and the structure of studied compounds based on X-ray analysis were investigated. As a result, the QSPR model was developed using PLSR (Partial Least Squares Regression). Decomposition of the Hirshfeld surfaces (calculated based on the arrangement of molecules in the crystal lattice) revealed the various types of weak intermolecular interactions that occurred in the crystal lattice. The obtained data, in combination with global reactivity descriptors calculated using HOMO and LUMO energy values, were used as explanatory variables. The developed model was characterized by good validation metrics (RMSECAL = 0.017, RMSECV = 0.029, R2CAL = 0.989, and R2CV = 0.968) and indicated that the solid-state fluorescence quantum yield of methoxy-trans-stilbene derivatives was mainly dependent on weak intermolecular C…C contacts corresponding to π-π stacking and C…O/O…C interactions. To a lesser extent and inversely proportional, the fluorescence quantum yield was affected by the interactions of the type O…H/H…O and H…H and the electrophilicity of the molecule.

Dual polarization and mutual coupling improvement of UWB MIMO antenna with cross shape decoupling structure

A simple, compact MIMO antenna that operates in the UWB frequency band is presented in this paper. The proposed antenna consists of four elements where the shape of the radiators derived from series modifications of a conventional rectangular patch antenna. Radiators are placed orthogonally to each other to achieve the dual polarization operation and cross shape Defected Ground Structure (DGS) is inserted in the ground plane to improve decoupling between the antenna elements. The designed model was fabricated on FR-4 substrate with dimensions of 48 × 44 × 1.6 mm3 and measurements were performed to validate the simulated results. The fractional bandwidth reaches 112% from 3.1 GHz to 11 GHz with reflection coefficient less than -10 dB and mutual coupling between antenna elements is less than -20 dB. The antenna has a stable radiation pattern, and low gain variation. In addition, the time domain characteristics and diversity performance indicators were evaluated and compared with the standard values. The miniaturized dimensioning of the antenna along with the good performance metrics indicates that the proposed antenna can be used in various UWB applications such as portable handheld devices.

Viewpoint

Viewpoint

Microcrystalline Cellulose Decorated with Fe3O4 Nanoparticle Catalysts for the Microwave-Assisted Synthesis of Thioglyoxamides

Ferrite nanoparticles (Fe3O4 NPs) supported on microcrystalline cellulose (MCC) are introduced as an active nanocatalyst for the preparation of the α-keto thioglyoxamides via a microwave (MW)-assisted protocol. This prepared catalyst shows consistent particle size of Fe3O4 NPs as characterized by PXRD, XPS, SEM, EDS, EDX, TEM, HR-TEM, and FT-IR analysis. A wide variety of thioglyoxamindes are synthesized under optimized reaction conditions to afford a good yield in a short time. Importantly, the presented catalyst can be recycled up to five cycles with no major loss in its activity and is equally relevant for the gram-scale synthesis of thioglyoxamides. This environmentally benign approach offers several advantages and displays good green chemistry metrics (GCM), namely, process mass intensity (PMI), E-factor and atom economy (AE), and reaction mass efficiency (RME).

Design, Hardware Implementation on FPGA and Performance Analysis of Three Chaos-Based Stream Ciphers

In this paper, we come up with three secure chaos-based stream ciphers, implemented on an FPGA board, for data confidentiality and integrity. To do so, first, we performed the statistical security and hardware metrics of certain discrete chaotic map models, such as the Logistic, Skew-Tent, PWLCM, 3D-Chebyshev map, and 32-bit LFSR, which are the main components of the proposed chaotic generators. Based on the performance analysis collected from the discrete chaotic maps, we then designed, implemented, and analyzed the performance of three proposed robust pseudo-random number generators of chaotic sequences (PRNGs-CS) and their corresponding stream ciphers. The proposed PRNGs-CS are based on the predefined coupling matrix M. The latter achieves a weak mixing of the chaotic maps and a chaotic multiplexing technique or XOR operator for the output function. Therefore, the randomness of the sequences generated is expanded as well as their lengths, and divide-and-conquer attacks on chaotic systems are avoided. In addition, the proposed PRNGs-CS contain polynomial mappings of at least degree 2 or 3 to make algebraic attacks very difficult. Various experimental results obtained and analysis of performance in opposition to different kinds of numerical and cryptographic attacks determine the high level of security and good hardware metrics achieved by the proposed chaos system. The proposed system outperformed the state-of-the-art works in terms of high-security level and a high throughput which can be considered an alternative to the standard methods.

On the prediction of premature births in Hispanic labour patients using uterine contractions, heart beat signals and prediction machines.

Preterm birth is a global epidemic affecting millions of mothers across different ethnicities. The cause of the condition remains unknown but has recognised health-based implications, in addition to financial and economic ones. Machine Learning methods have enabled researchers to combine datasetsusinguterine contraction signals with various forms of prediction machines to improve awareness of the likelihood of premature births. This work investigates the feasibility of enhancing these prediction methods using physiological signals including uterine contractions, and foetal and maternal heart rate signals, for a population of south American women in active labour. As part of this work, the use of the Linear Series Decomposition Learner (LSDL) was seen to lead to an improvement in the prediction accuracies of all models, which included supervised and unsupervised learning models. The results from the supervised learning models showed high prediction metrics upon the physiological signals being pre-processed by the LSDL for all variations of the physiological signals. The unsupervised learning models showed good metrics for the partitioning of Preterm/Term labour patients from their uterine contraction signals but produced a comparatively lower set of results for the various kinds of heart rate signals investigated.

Deep evidential learning in diffusion convolutional recurrent neural network

<abstract><p>Graph neural networks (GNNs) is applied successfully in many graph tasks, but there still exists a limitation that many of GNNs model do not consider uncertainty quantification of its output predictions. For uncertainty quantification, there are mainly two types of methods which are frequentist and Bayesian. But both methods need to sampling to gradually approximate the real distribution, in contrast, evidential deep learning formulates learning as an evidence acquisition process, which could get uncertainty quantification by placing evidential priors over the original Gaussian likelihood function and training the NN to infer the hyperparameters of the evidential distribution without sampling. So evidential deep learning (EDL) has its own advantage in measuring uncertainty. We apply it with diffusion convolutional recurrent neural network (DCRNN), and do the experiment in spatiotemporal forecasting task in a real-world traffic dataset. And we choose mean interval scores (MIS), a good metric for uncertainty quantification. We summarized the advantages of each method.</p></abstract>

Predicting rockburst damage scale in seismically active mines using a classifier ensemble approach

Rockburst still remains one of the most important sources of hazards in underground mine industry. This phenomenon results in many accidents and casualties in many underground mine projects around the world. Despite the scientific achievements in rock mechanics and engineering, the accurate prediction of rockburst damage potential is still challenging due to the complexity of seismic event occurrence. Hence, this paper aims to develop a reliable classifier ensemble to rockburst intensity in underground mine excavations subjected to seismicity. High quality rockburst database consisting of 254 case histories was used for the study. The classifier ensemble was developed through aggregation of several commonly used machine learning classifiers using the weighted voting. The performance of the classifier ensemble was evaluated using several indicators, namely: accuracy, recall, precision, and F1-score. The overall results indicate that the proposed classifier ensemble achieved good performance metrics and outperformed some existing empirical methods. It is concluded that the classifier ensemble could assist engineers to properly assess rockburst damage and contribute to selecting adequate ground control measures.

Characterizing X-Linked Dystonia Parkinsonism Using Clustering Techniques in Data Science

X-linked dystonia parkinsonism (XDP) is a neurodegenerative rapidly progressive movement disorder seen almost exclusively among male patients of Filipino descent with traces of ancestry to Panay Island in the Philippines. The prevalence of XDP is only 0.31/100,000 for the entire Philippines but it is 23.66/100,000 in Capiz Province in Panay. Using a retrospectively obtained data from a previous study collected from all the 16 towns and 1 city of Capiz province in the island of Panay in the Philippines, the XDP dataset showed superior clustering tendency. Two clustering approaches commonly used in data science namely Partitioning Around Medoids (PAM) and K-modes were applied to the XDP dataset. Both clustering approaches obtained good internal and external validation. Likewise, both clustering techniques generated good performance metrics. Overall, K-modes performed better than PAM with an 81% accuracy, 78% recall (sensitivity), 82% specificity, 68% precision, 72% F1-score and 0.58 Matthews Correlation Coefficient. Both PAM and K-modes were able to highlight the differences between the clusters. Thus, the resulting clusters can be useful as screening tools to differentiate those with and without XDP. Feature importance of the attributes was also performed with feet shuffling generating the highest discriminative power in distinguishing patients with and without XDP. The collaboration of data scientists with neurology experts in movement disorders is a step forward to the deployment and acceptability of machine learning tools in clinical practice in neurology.

On the Spatial-Temporal Behavior, and on the Relationship Between Water Quality and Hydrometeorological Information to Predict Dissolved Oxygen in Tropical Reservoirs. Case Study: La Miel, Hydropower Dam

Hydropower is currently one of the leading renewable energy sources in developing countries. Despite the benefits that it can provide, it also triggers significant environmental impacts, such as changes in the reservoirs’ water quality. In quantifying those changes, dissolved oxygen (DO) is used as one of the water quality indicators and is the most used variable to quantify water quality and analyze water pollution. This paper aims to establish a relationship between water quality and hydrometeorological variables in tropical reservoirs to better estimate dissolved oxygen. Univariate and multivariate techniques were used to analyze temporal and spatial changes in watersheds to better select vital variables for the forecast model, such as Vector Autoregression (VAR). The results show that, for all monitoring stations, the water quality variables associated with the DO process are COD, BOD, and PO₄. Likewise, precipitation and flow discharge were the hydrometeorological parameters that had the most significant impact on DO. Also, the principal component analysis (PCA) allowed us to identify that the strength of the relationships between water quality and hydrometeorology changes depending on the location of the monitoring site. Finally, the implementation of a VAR model showed good performance metrics for dissolved oxygen predictions based on all analyses.

Botnet Detection Using Artificial Intelligence

In the year 2021 more than 80 million data were breached by cyber attackers. Most of these attacks were executed as a type of ransomware attack and cyber-attack. There are different methods performed by attackers to target individual user, but to breach an organization, they use Botnet forces. A botnet (robot network), is malware infected network that is controlled by a single attacker called bot-herder. Cyber - attacker attacks many users with their malware script by different mediums like emails, spams and takes command and control (CC) of the victim device. Using these devices attacker forms a network and performs large attack like distributed denial-of-service (DDoS), attack on an organization to breach data. The complex analysis for cybersecurity analyst is to find bot-herder and the infected network. The structures of the botnets are become very different now days. Botnets can be found with its peer-peer (p2p) structure, signature detection, behavioral analysis, domain names (DNS) and network traffic. To make this different feature analysis easier, the usage of artificial intelligence (AI) is introduced in cyber security. Data from previous attacks are collected, trained using a model which helps in prediction of future attacks. Detection of DNS of core CC servers using AI are widely used nowadays. This research mainly focuses on detection of botnet malware from the net flows of malware packets. The botnet attack data set are collected from resources like Czech - university (CTU-13), Information security and object technology (ISOT). The bi-directional net flow data and the calculation of the network packets are used. Using algorithms like support vector machine (SVM), decision tree and multi-layer perceptron, the data set is trained and tested. After the training and testing, the decision tree model has good accuracy and performance metrics of 92%. This model is considered as a best fit model and helps in detection the of malware packets. The research's objective is to build an alerting system which reports once a malware packet is intruded into a network.

Recyclable mesalamine-functionalized magnetic nanoparticles (mesalamine/GPTMS@SiO2@Fe3O4) for tandem Knoevenagel-Michael cyclocondensation: grinding technique for the synthesis of biologically active 2-amino-4H-benzo[b]pyran derivatives.

In the present study, mesalamine-functionalized on magnetic nanoparticles (mesalamine/GPTMS@SiO2@Fe3O4) is fabricated as an efficient and magnetically recoverable nanocatalyst. The as-prepared nanocatalyst was successfully synthesized in three steps using a convenient and low-cost method via modification of the surface of Fe3O4 nanoparticles with silica and GPTMS, respectively, to afford GPTMS@SiO2@Fe3O4. Finally, treatment with mesalamine as a powerful antioxidant generates the final nanocatalyst. Then, its structure was characterized by FT-IR, SEM, TEM, EDX, XRD, BET, VSM, and TGA techniques. The average size was found to be approximately 38 nm using TEM analysis and the average crystallite size was found to be approximately 27.02 nm using XRD analysis. In particular, the synthesized nanocatalyst exhibited strong thermal stability up to 400 °C and high magnetization properties. The activity of the synthesized nanocatalyst was evaluated in the tandem Knoevenagel-Michael cyclocondensation of various aromatic aldehydes, dimedone and malononitrile under a dry grinding method at room temperature to provide biologically active 2-amino-4H-benzo[b]pyran derivatives products in a short time with good yields. The presented procedure offers several advantages including gram-scale synthesis, good green chemistry metrics (GCM), easy fabrication of the catalyst, atom economy (AE), no use of column chromatography, and avoiding the generation of toxic materials. Furthermore, the nanocatalyst can be reused for 8 cycles with no loss of performance by using an external magnet.

Enhancing stability in cardiovascular disease risk prediction: A deep learning approach leveraging retinal images

BackgroundThe retina provides valuable insights into vascular health within the body. Prior studies have demonstrated the potential of deep learning in predicting Cardiovascular Disease (CVD) risk using color fundus photographs. PurposeTo use fundus images to more consistently predict the World Health Organization (WHO) CVD score and to address the problem of year-to-year fluctuations associated with the traditional CVD risk score calculation. MethodsUtilizing 55,540 fundus images from 3,765 participants with 6-year follow-up data, we designed a DL model named Reti-WHO based on the Swin Transformer to predict CVD risk regression scores. Multiple regression and classification metrics such as coefficient of determination (R2-score), Mean Absolute Error (MAE), sensitivity and specificity were employed to assess the accuracy of the Reti-WHO. Significance differences between WHO CVD scores and Reti-WHO scores were also assessed. Vessel measurements were employed to interpret the model and evaluate the association between Reti-WHO and vascular conditions. ResultsThe deep learning model achieved good classification and regression metrics on the validation set, with an R2-score of 0.503, MAE of 1.58, sensitivity of 0.81, and specificity of 0.66. There was no statistically significant difference between WHO CVD scores and Reti-WHO scores (P value = 0.842). The model exhibited a stronger correlation with vascular measurements, including mean and variance of arc and chord in arteries and veins. Comparing box plots and Vyshyvanka plots depicting changes in patients' CVD over the years, the Reti-WHO calculated by our model demonstrated greater stability compared to non-deep learning-based WHO CVD risk calculations. ConclusionsOur Reti-WHO scores demonstrated enhanced stability compared to WHO CVD scores calculated solely from the patient's physical indicators, suggesting that the features learned from retinal fundus photographs serve as robust indicators of CVD risk. However, the model may still exhibit false negatives in high-risk predictions, requiring ongoing research for refinement. Future directions involve validating the model across diverse populations and exploring multi-image and multi-modal approaches to enhance prediction accuracy.

Enhancing the classification metrics of spectroscopy spectrums using neural network based low dimensional space.

Spectroscopy is a methodology for gaining knowledge of particles, especially biomolecules, by quantifying the interactions between matter and light. By examining the level of light absorbed, reflected or released by a specimen, its constituents, properties, and volume can be determined. Spectra obtained through spectroscopy procedures are quick, harmless and contactless; hence nowadays preferred in chemometrics. Due to the high dimensional nature of the spectra, it is challenging to build a robust classifier with good performance metrics. Many linear and nonlinear dimensionality reduction-based classification models have been previously implemented to overcome this issue. However, they lack in capturing the subtle details of the spectra into the low dimension space or cannot efficiently handle the nonlinearity present in the spectral data. We propose a graph-based neural network embedding approach to extract appropriate features into latent space and circumvent the spectrums' nonlinearity problem. Our approach performs dimensionality reduction into two phases: constructing a nearest neighbor graph and producing almost linear embedding using a fully connected neural network. Further, the low dimensional embedding is subjected to classification using the Random Forest algorithm. In this paper, we have implemented and compared our technique with four nonlinear dimensionality techniques widely used for spectral data analysis. In this study, we have considered five different spectral datasets belonging to specific applications. The various classification performance metrics of all the techniques are evaluated. The proposed approach is able to perform competitively well on six different low-dimensional spaces for each dataset with an accuracy score above 95% and Matthew's correlation coefficient value close to 1. The trustworthiness score of almost 1 show that the presented dimensionality reduction approach preserves the closest neighbor structure of high dimensional spectral inputs into latent space.

One-Sided Unsupervised Image Dehazing Network Based on Feature Fusion and Multi-Scale Skip Connection

Haze and mist caused by air quality, weather, and other factors can reduce the clarity and contrast of images captured by cameras, which limits the applications of automatic driving, satellite remote sensing, traffic monitoring, etc. Therefore, the study of image dehazing is of great significance. Most existing unsupervised image-dehazing algorithms rely on a priori knowledge and simplified atmospheric scattering models, but the physical causes of haze in the real world are complex, resulting in inaccurate atmospheric scattering models that affect the dehazing effect. Unsupervised generative adversarial networks can be used for image-dehazing algorithm research; however, due to the information inequality between haze and haze-free images, the existing bi-directional mapping domain translation model often used in unsupervised generative adversarial networks is not suitable for image-dehazing tasks, and it also does not make good use of extracted features, which results in distortion, loss of image details, and poor retention of image features in the haze-free images. To address these problems, this paper proposes an end-to-end one-sided unsupervised image-dehazing network based on a generative adversarial network that directly learns the mapping between haze and haze-free images. The proposed feature-fusion module and multi-scale skip connection based on residual network consider the loss of feature information caused by convolution operation and the fusion of different scale features, and achieve adaptive fusion between low-level features and high-level features, to better preserve the features of the original image. Meanwhile, multiple loss functions are used to train the network, where the adversarial loss ensures that the network generates more realistic images and the contrastive loss ensures a meaningful one-sided mapping from the haze image to the haze-free image, resulting in haze-free images with good quantitative metrics and visual effects. The experiments demonstrate that, compared with existing dehazing algorithms, our method achieved better quantitative metrics and better visual effects on both synthetic haze image datasets and real-world haze image datasets.

Deep learning applications in coronary anatomy imaging: a systematic review and meta-analysis.

The application of deep learning on medical imaging is growing in prevalence in the recent literature. One of the most studied areas is coronary artery disease (CAD). Imaging of coronary artery anatomy is fundamental, which has led to a high number of publications describing a variety of techniques. The aim of this systematic review is to review the evidence behind the accuracy of deep learning applications in coronary anatomy imaging. The search for the relevant studies, which applied deep learning on coronary anatomy imaging, was performed in a systematic approach on MEDLINE and EMBASE databases, followed by reviewing of abstracts and full texts. The data from the final studies was retrieved using data extraction forms. A meta-analysis was performed on a subgroup of studies, which looked at fractional flow reserve (FFR) prediction. Heterogeneity was tested using tau2, I2 and Q tests. Finally, a risk of bias was performed using Quality Assessment of Diagnostic Accuracy Studies (QUADAS) approach. A total of 81 studies met the inclusion criteria. The most common imaging modality was coronary computed tomography angiography (CCTA) (58%) and the most common deep learning method was convolutional neural network (CNN) (52%). The majority of studies demonstrated good performance metrics. The most common outputs were focused on coronary artery segmentation, clinical outcome prediction, coronary calcium quantification and FFR prediction, and most studies reported area under the curve (AUC) of ≥80%. The pooled diagnostic odds ratio (DOR) derived from 8 studies looking at FFR prediction using CCTA was 12.5 using the Mantel-Haenszel (MH) method. There was no significant heterogeneity amongst studies according to Q test (P=0.2496). Deep learning has been used in many applications on coronary anatomy imaging, most of which are yet to be externally validated and prepared for clinical use. The performance of deep learning, especially CNN models, proved to be powerful and some applications have already translated into medical practice, such as computed tomography (CT)-FFR. These applications have the potential to translate technology into better care of CAD patients.

Indicators of Higher Education and the Public Good in Africa: A Dashboard Approach

Indicators and metrics have gained increasing prominence in international higher education in recent years, and global rankings have become a powerful force in shaping ideas of what the university is and should be. Yet these measures do a poor job of capturing the broad role of the institution, and particularly in recognising its actions in promoting the public good and addressing inequalities. African higher education institutions have struggled to perform well in the conventional rankings, whose indicators rely on extensive resources for high-level research. This article explores the possibilities of alternative metrics for understanding the public good contribution of universities in the context of four African countries: Ghana, Kenya, Nigeria and South Africa. After assessing the shortcomings of the existing indicators and metrics, and the challenges of the availability of data, it puts forward a dashboard approach as a possible new model. Dashboards have the advantage of avoiding the conflation of diverse qualities of importance and allow different profiles of an institution to be compared. The article proposes six main elements for the dashboard: solidarity with society, equitable access and deliberative space (which correspond to the intrinsic notion of public good and graduate destinations, knowledge production and community engagement (which correspond to instrumental notions). Finally, the challenges of implementing public good metrics in practice are discussed.

A taxonomy for similarity metrics between Markov decision processes

Although the notion of task similarity is potentially interesting in a wide range of areas such as curriculum learning or automated planning, it has mostly been tied to transfer learning. Transfer is based on the idea of reusing the knowledge acquired in the learning of a set of source tasks to a new learning process in a target task, assuming that the target and source tasks are close enough. In recent years, transfer learning has succeeded in making reinforcement learning (RL) algorithms more efficient (e.g., by reducing the number of samples needed to achieve (near-)optimal performance). Transfer in RL is based on the core concept of similarity: whenever the tasks are similar, the transferred knowledge can be reused to solve the target task and significantly improve the learning performance. Therefore, the selection of good metrics to measure these similarities is a critical aspect when building transfer RL algorithms, especially when this knowledge is transferred from simulation to the real world. In the literature, there are many metrics to measure the similarity between MDPs, hence, many definitions of similarity or its complement distance have been considered. In this paper, we propose a categorization of these metrics and analyze the definitions of similarity proposed so far, taking into account such categorization. We also follow this taxonomy to survey the existing literature, as well as suggesting future directions for the construction of new metrics.