Self harm detection and support on Twitter

A preliminary analysis shows how contents related tonon-suicidal self-injury (NSSI) proliferate on Twitter . We use Twitter to collect relevant data, analyse, and proffer ways of supporting users prone to NSSI behaviour . The study is based on the premise that social media can be used as a tool to support proactive measures to ease the negative impact of self-harm .…

A Multi Stage Stochastic Programming Approach to Epidemic Resource Allocation with Equity Considerations

Existing compartmental models in epidemiology are limited in terms ofoptimizing the resource allocation to control an epidemic outbreak underdisease growth uncertainty . The proposed multi-stage stochastic program involves various disease growth scenarios and optimizes the distribution oftreatment centers and resources while minimizing the total expected number of infections and funerals .…

So you want to be a Super Researcher

The ‘Super Researcher’ app is built on R shiny which provides an interactive interface to users . It uses the BigData framework Apache Spark to mine relevant information from a huge authorinformation database . The author’s information is stored and manipulated using both SQL(SQLite) and NoSQL(HBase) databases .…

Infrastructure Resilience Curves Performance Measures and Summary Metrics

Resilience curves communicate system behavior and resilience properties to critical infrastructure stakeholders . This manuscript,through a review of 274 publications, defines a taxonomy of resilience curveperformance measures and discusses associated performance targets and normalization techniques . It also synthesizes recommendations for selecting measures and metrics for common applications and stakeholdersacross critical infrastructure resilience domains .…

Generate Natural Language Explanations for Recommendation

Current explainable recommendation models mostlygenerate textual explanations based on pre-defined sentence templates . We propose a hierarchical sequence-to-sequence model (HSS) for personalized explanation generation . We further propose an auto-denoisingmechanism based on topical item feature words for sentence generation . This research is one of the initial steps to grant intelligent agents with the ability toexplain itself based on natural language sentences, say the authors of the study .…

Reconfigurable Intelligent Surface Assisted Mobile Edge Computing with Heterogeneous Learning Tasks

Mobile edgecomputing (MEC) provides a natural platform for AI applications since it is with rich computation resources to train machine learning (ML) models, as well as low-latency access to the data generated by mobile and internet of things(IoT) devices . An alternating optimization (AO)-based framework is proposed to optimizethe three terms iteratively, where a successive convex approximation(SCA)-based algorithm is developed to solve the power allocation problem .…

Prediction of High Performance Computing Input Output Variability and Its Application to Optimization for System Configurations

Performance variability is an important measure for a reliable highperformance computing (HPC) system . The prediction of HPC variability is a challenging problem in HPC systems and there is little statistical work on this problem to date . We evaluate performance of themethods by measuring prediction accuracy at previously unseen system configurations .…