Polyphenolic substances obtained from the leaves of Vitex negundo have gained attention for its therapeutic benefits. To obtain the maximized antioxidant-rich polyphenolic substances from the Vitex negundo leaves via microwave-assisted extraction (MAE), Box–Behnken design (BBD) was employed to evaluate the influence of several variables of MAE on the total phenolic content (TPC) and total flavonoid content (TFC) as determined by Folin–Ciocalteu assay method and Aluminum Chloride assay method, and the antioxidant capacity as measured by DPPH and ABTS assay methods, respectively. Response surface methodology (RSM) was used for finding optimal extraction conditions. GC-MS analysis was conducted on the extract. The tocopherol content and anticancer potential were estimated using HPLC and MTT assay respectively. The ideal extraction parameters were found to be 14 min, 48 °C, 65 % (v/v) methanol concentration, and 14 ml of extraction solvent. The optimum experimental conditions produced the TPC and TFC values of 1.46 mg GAE/g of dried extract and 1.06 mg QE/g of dried extract respectively. Furthermore, the DPPH and ABTS assays results showed the optimum values of 53 % and 77 % respectively. 12 bioactive compounds were identified using GC-MS. The amount of tocopherol was found to be 414.87 µg/g. Lastly, the obtained leaf extract demonstrated its anticancer potential on PC3 cell lines. The findings demonstrated leaf extract's potential as a useful source of polyphenols with strong antioxidant qualities that can be used in a variety of pharmaceutical applications.

Deep learning’s preponderance across scientific domains has reshaped high-stakes decision-making, making it essential to follow rigorous operational frameworks that include both Right-to-Privacy (RTP) and Right-to-Explanation (RTE). This paper examines the complexities of combining these two requirements. For RTP, we focus on ‘Differential privacy’ (DP), which is considered the current gold standard for privacy-preserving machine learning due to its strong quantitative guarantee of privacy. For RTE, we focus on post-hoc explainers: they are the go-to option for model auditing as they operate independently of model training. We formally investigate DP models and various commonly-used post-hoc explainers: how to evaluate these explainers subject to RTP, and analyze the intrinsic interactions between DP models and these explainers. Furthermore, our work throws light on how RTP and RTE can be effectively combined in high-stakes applications. Our study concludes by outlining an industrial software pipeline, with the example of a widely used use case, that respects both RTP and RTE requirements.

Modern Education is not Modern without AI. However, AI's complex nature makes understanding and fixing problems challenging. Research worldwide shows that a parent's income greatly influences a child's education. This led us to explore how AI, especially complex models, makes important decisions using Explainable AI tools. Our research uncovered many complexities linked to parental income and offered reasonable explanations for these decisions. However, we also found biases in AI that go against what we want from AI in education: clear transparency and equal access for everyone. These biases can impact families and children's schooling, highlighting the need for better AI solutions that offer fair opportunities to all. This chapter tries to shed light on the complex ways AI operates, especially concerning biases. These are the foundational steps towards better educational policies, which include using AI in ways that are more reliable, accountable, and beneficial for everyone involved.

Link prediction (LP) is an important problem in network science and machine learning research. The state-of-the-art LP methods are usually evaluated in a uniform setup, ignoring several factors associated with the data- and application-specific needs. We identify a number of such factors, such as network type, problem type, geodesic distance between the end nodes and its distribution over the classes, nature and applicability of LP methods, class imbalance and its impact on early retrieval, evaluation metric, etc., and present an experimental setup which allows us to evaluate LP methods in a rigorous and controlled manner. We perform extensive experiments with 58 LP methods of diverse categories over 24 real network datasets in this controlled setup, and gather valuable insights on the interactions of these factors with the performance of LP through an array of carefully designed hypothesis tests. Following the insights, we provide recommendations on the experimental setup and the evaluation measures to be employed considering application scenarios, to be followed as best practice for evaluating LP methods.

Faithfulness is a critical metric to assess the reliability of explainable AI. In NLP, current methods for faithfulness evaluation are fraught with discrepancies and biases, often failing to capture the true reasoning of models. We introduce Adversarial Sensitivity as a novel approach to faithfulness evaluation, focusing on the explainer’s response when the model is under adversarial attack. Our method accounts for the faithfulness of explainers by capturing sensitivity to adversarial input changes. This work addresses significant limitations in existing evaluation techniques, and furthermore, quantifies faithfulness from a crucial yet under-explored paradigm.

Data derived from wearable activity trackers may provide important clinical insights into disease progression and response to intervention, but only if clinicians can interpret it in a meaningful manner. Longitudinal activity data can be visually presented in multiple ways, but research has failed to explore how clinicians interact with and interpret these visualisations. In response, this study developed a variety of visualisations to understand whether alternative data presentation strategies can provide clinicians with meaningful insights into patient's physical activity patterns. Objective: To explore clinicians' opinions on different visualisations of actigraphy data. Methods: Four visualisations (stacked bar chart, clustered bar chart, linear heatmap and radial heatmap) were created using Matplotlib and Seaborn Python libraries. A focus group was conducted with 14 clinicians across 2 hospitals. Focus groups were audio-recorded, transcribed and analysed using inductive thematic analysis. Results: Three major themes were identified: (1) the importance of context, (2) interpreting the visualisations and (3) applying visualisations to clinical practice. Although clinicians saw the potential value in the visualisations, they expressed a need for further contextual information to gain clinical benefits from them. Allied health professionals preferred more granular, temporal information compared to doctors. Specifically, physiotherapists favoured heatmaps, whereas the remaining members of the team favoured stacked bar charts. Overall, heatmaps were considered more difficult to interpret. Conclusion: The current lack of contextual data provided by wearables hampers their use in clinical practice. Clinicians favour data presented in a familiar format and yet desire multi-faceted filtering. Future research should implement user-centred design processes to identify ways in which all clinical needs can be met, potentially using an interactive system that caters for multiple levels of granularity. Irrespective of how data is displayed, unless clinicians can apply it in a manner that best supports their role, the potential of this data cannot be fully realised.

Background: Wearable sensors allow researchers to remotely capture digital health data, including physical activity, which may identify digital biomarkers to differentiate healthy and clinical cohorts. To date, research has focused on high-level data (e.g., overall step counts) which may limit our insights to whether people move differently, rather than how they move differently. Objective: This study therefore aimed to use actigraphy data to thoroughly examine activity patterns during the first hours following waking in arthritis patients (n = 45) and healthy controls (n = 30). Methods: Participants wore an Actigraph GT9X Link for 28 days. Activity counts were analysed and compared over varying epochs, ranging from 15 min to 4 h, starting with waking in the morning. The sum, and a measure of rate of change of cumulative activity in the period immediately after waking (area under the curve [AUC]) for each time period, was calculated for each participant, each day, and individual and group means were calculated. Two-tailed independent t tests determined differences between the groups. Results: No differences were seen for summed activity counts across any time period studied. However, differences were noted in the AUC analysis for the discrete measures of relative activity. Specifically, within the first 15, 30, 45, and 60 min following waking, the AUC for activity counts was significantly higher in arthritis patients compared to controls, particularly at the 30 min period (t = -4.24, p = 0.0002). Thus, while both cohorts moved the same amount, the way in which they moved was different. Conclusion: This study is the first to show that a detailed analysis of actigraphy variables could identify activity pattern changes associated with arthritis, where the high-level daily summaries did not. Results suggest discrete variables derived from raw data may be useful to help identify clinical cohorts and should be explored further to determine if they may be effective clinical biomarkers.

In social-based delay tolerant network (DTN) applications, hand-held mobile devices exchange information. The inherent social property of DTN has encouraged contemporary researchers in exploiting social metrics to devise forwarding techniques for efficient routing. This work observes evidence of seasonal behavior in contacts between node-pairs in real mobility traces, and exploits it to devise a novel seasonality aware similarity measure. We incorporate seasonality information into tie-strength, and then use it as link weight in a weighted similarity measure which we extend from Katz similarity index. We propose a Seasonality Aware Social-based (SAS) DTN forwarding technique based on the proposed similarity measure and ego-betweenness centrality. Finally we perform real trace driven simulations to show that SAS outperforms baseline social-based DTN forwarding methods significantly.

Background: Consumer wearables can provide a practical and accessible method of data collection in actigraphy research. However, as this area continues to grow, it is becoming increasingly important for researchers to be aware of the many challenges facing the capture of quality data using consumer wearables. Objective: This study aimed to (1) present the challenges encountered by a research team in actigraphy data collection using a consumer wearable and (2) present considerations for researchers to apply in the pursuit of robust data using this approach. Methods: The Nokia Go was deployed to 33 elite Gaelic footballers from a single team for a planned period of 14 weeks. A bring-your-own-device model was employed for this study where the Health Mate app was downloaded on participants’ personal mobile phones and connected to the Nokia Go via Bluetooth. Retrospective evaluation of the researcher and participant experience was conducted through transactional data such as study logs and email correspondence. The participant experience of the data collection process was further explored through the design of a 34-question survey utilizing aspects of the Technology Acceptance Model. Results: Researcher challenges included device disconnection, logistics and monitoring, and rectifying of technical issues. Participant challenges included device syncing, loss of the device, and wear issues, particularly during contact sport. Following disconnection issues, the data collection period was defined as 87 days for which there were 18 remaining participants. Average wear time was 79 out of 87 days (90%) and 20.8 hours per day. The participant survey found mainly positive results regarding device comfort, perceived ease of use, and perceived usefulness. Conclusions: Although this study did not encounter some of the common published barriers to wearable data collection, our experience was impacted by technical issues such as disconnection and syncing challenges, practical considerations such as loss of the device, issues with personal mobile phones in the bring-your-own-device model, and the logistics and resources required to ensure a smooth data collection with an active cohort. Recommendations for achieving high-quality data are made for readers to consider in the deployment of consumer wearables in research.

Physical activity (PA) is a key component in the treatment of a range of chronic health conditions. It is therefore important for researchers and clinicians to accurately assess and monitor PA. Although advances in wearable technology have improved this, there is a need to investigate PA in greater depth than the sum of its total parts. Specifically, linking deep PA data to patient outcomes offers a valuable, and unexplored use for wearable devices. As a result, this paper extracts useful features from accelerometer data (Actigraph GT3X Link), and applies machine learning algorithms to predict daily pain and stiffness. This was applied to a population of 30 arthritis patients and 15 healthy volunteers. Participants were provided with an Actigraph and asked to wear it continuously for 28 days. Results demonstrate that it is possible to predict both pain and stiffness of patients using the extracted accelerometer features.

This paper addresses link prediction problem in directed networks by exploiting reciprocative nature of human relationships. It first proposes a null model to present evidence that reciprocal links influence the process of “triad formation”. Motivated by this, reciprocal links are exploited to enhance link prediction performance in three ways: (a) a reciprocity-aware link weighting technique is proposed, and existing weighted link prediction methods are applied over the resultant weighted network; (b) new link prediction methods are proposed, which exploit reciprocity; and (c) existing and proposed methods are combined toward supervised prediction to enhance the prediction performance further. All experiments are carried out on two real directed network datasets.

Link prediction problem in complex networks has received substantial amount of attention in the field of social network analysis. Though initial studies consider only static snapshot of a network, importance of temporal dimension has been observed and cultivated subsequently. In recent times, multi-domain relationships between node-pairs embedded in real networks have been exploited to boost link prediction performance. In this paper, we combine multi-domain topological features as well as temporal dimension, and propose a robust and efficient feature set called TMLP (Time-aware Multi-relational Link Prediction) for link prediction in dynamic heterogeneous networks. It combines dynamics of graph topology and history of interactions at dyadic level, and exploits time-series model in the feature extraction process. Several experiments on two networks prepared from DBLP bibliographic dataset show that the proposed framework outperforms the existing methods significantly, in predicting future links. It also demonstrates the necessity of combining heterogeneous information with temporal dynamics of graph topology and dyadic history in order to predict future links. Empirical results find that the proposed feature set is robust against longitudinal bias.

The current developments of software defined networking (SDN) paradigm provide a flexible architecture for network control and management, in the cost of deploying new hardwares by replacing the existing routing infrastructure. Further, the centralized controller architecture of SDN makes the network prone to single point failure and creates performance bottleneck. To avoid these issues and to support network manageability over the existing network infrastructure, we develop Flipper in this paper, that uses only software augmentation to convert existing off-the-shelf routers to network policy design and enforcement points (PDEP). We develop a distributed self-stabilized architecture for dynamic role change of network devices from routers to PDEPs, and make the architecture fault-tolerant. The performance of Flipper has been analyzed from both simulation over synthetic networks, and emulation over real network protocol stacks, and we observe that Flipper is scalable, flexible and fail-safe that can significantly boost up the manageability of existing network infrastructure.

Existing studies on evolution of social network largely focus on addition of new nodes and links in the network. However, as network evolves, existing relationships degrade and break down, and some nodes go to hibernation or decide not to participate in any kind of activities in the network where it belongs. Such nodes and links, which we refer as 'dull', may affect analysis and prediction tasks in networks. This paper formally defines the problem of predicting dull nodes and links at an early stage, and proposes a novel time aware method to solve it. Pruning of such nodes and links is framed as 'network data cleaning' task. As the definitions of dull node and link are non-Trivial and subjective, a novel scheme to label such nodes and links is also proposed here. Experimental results on two real network datasets demonstrate that the proposed method accurately predicts potential dull nodes and links. This paper further experimentally validates the need for data cleaning by investigating its effect on the well-known 'link prediction' problem.

Tie weight plays an important role in maintaining cohesiveness of social networks. However, influence of the tie weight on link prediction has not been clearly understood. In few of the previous studies, conflicting observations have been reported. In this paper, we revisit the study of the effect of tie weight on link prediction. Previous studies have focused on additive weighting model. However, the additive model is not suitable for all node proximity based prediction methods. For understanding the effect of weighting models on different prediction methods, we propose two new weighting models namely, min-flow and multiplicative. The effect of all three weighting models on node proximity based prediction methods over ten datasets of different characteristics is thoroughly investigated. From several experiments, we observe that the response of different weighting models varies, subject to prediction methods and datasets. Empirically, we further show that with the right choice of a weighting model, weighted versions may perform better than their unweighted counterparts.We further extend the study to show that proper tuning of the weighting function also influences the prediction performance. We also present an analysis based on the properties of the underlying graph to justify our result. Finally, we perform an analysis of the weak tie theory, and observe that unweighted models are suitable for inter-community link prediction, and weighted models are suitable for intra-community link prediction.

Majority of the studies on modeling the evolution of a social network using spectral graph kernels do not consider temporal effects while estimating the kernel parameters. As a result, such kernels fail to capture structural properties of the evolution over the time. In this paper, we propose temporal spectral graph kernels of four popular graph kernels namely path counting, triangle closing, exponential and neumann. Their responses in predicting future growth of the network have been investigated in detail, using two large datasets namely Facebook and DBLP. It is evident from various experimental setups that the proposed temporal spectral graph kernels outperform all of their non-temporal counterparts in predicting future growth of the networks. Copyright 2014 ACM.

This paper revisits the spectral based link prediction problem of evolutionary social networks reported in [9] and focuses on addressing two empirically observed issues which affect the prediction performance. First, the assumption that eigenvectors are constant over time is not valid for lower order eigenvectors and eigenvectors evolve over time as network evolves. A regression based method is proposed to predict evolving eigenvectors. Second, the spectral condition that higher order eigenvalues are greater than or equal to lower order eigenvalues may not be guaranteed by traditional curve fitting. Two smoothing methods are proposed to address this issue. From various experiments using two large datasets namely DBLP and Facebook, it is observed that proposed methods enhance prediction performance as compared to that of their counterparts. © 2013 IEEE.