[en] An increasing number of Machine Learning based systems are being designed to filter and visualize the relevant information from social media and web streams for disaster management. Given the dynamic disaster events, the notion of relevant information evolves, and thus, the active learning techniques are often considered to keep updating the predictive models for the relevant information filtering. However, the active relevant feedback provided by the human annotators to update the models are not validated. As a result, they can introduce unconscious biases in the learning process of humans and can result in an inaccurate or inefficient predictive system. Therefore, this paper describes the design and implementation of an open-source technology-based learning analytics system ‘ EMAssistant’ for the emergency volunteers or practitioners - referred as the trainee, to enhance their experiential learning cycle with the cause-effect reasoning on providing relevant feedback to the machine learning model. This continuous integration between the cause (providing feedback) and the effect (observing predictions from the updated model) in a visual form will likely to improve the understanding of the trainees to provide more accurate feedback. We propose to present the system design as well as provide hands-on exercises for the conference session.

[en] During large-scale disasters it is not uncommon for Public Safety Answering Points (e.g., 9-1-1) to encounter service disruptions or become overloaded due to call volume. As observed in the two past United States hurricane seasons, citizens are increasingly turning to social media whether as a consequence of their inability to reach 9-1-1, or as a preferential means of communications. Relying on past research that has examined social media use in disasters, combined with the practical knowledge of the first-hand disaster response experiences, this paper develops a knowledge-driven framework containing parameters useful in identifying patterns of shared information on social media when citizens need help. This effort explores the feasibility of determining differences, similarities, common themes, and time-specific discoveries of social media calls for help associated with hurricane evacuations. At a future date, validation of this framework will be demonstrated using datasets from multiple disasters. The results will lead to recommendations on how the framework can be modified to make it applicable as a generic disaster-type characterization tool.

[en] Health response plays a major role during disasters and information management plays a crucial role in situational awareness to adapt to evolving needs. Health organizations exchange information often through narrative-based documents called situation reports. Although situation reports are widely shared, they are an increasingly challenging information source from which to infer knowledge for situational awareness. This paper analyzed health information from traditional health reports using mixed methods during the aftermath of the 2010 Haiti Earthquake and provides insights into the patterns of what’s being said, how it’s being said, and trends over time. Opportunities lie ahead to analyze narrative documents at scale by combining human knowledge from qualitative coding with machine intelligence. In addition, developing unifying health domain ontologies representing diverse humanitarian health concepts will advance computational techniques to improve the efficiency and accuracy of retrieving knowledge for improved situational awareness and potential decision making during humanitarian health response.

[en] Oil spills generate enormous public concern and highlight the need for cost effective ad environmentally acceptable mitigation technologies. Physico-chemical methods are not completely effective after a spill. Hence, there is a need for improved and alternative technologies. Bioremediation is the most environmentally sound technology for clean up. This report intends to determine the potential of a bacterial consortium for degradation of Gulf and Bombay High crude oil. A four membered consortium was designed that could degrade 70% of the crude oil. A member of consortium produced a biosurfactant, rhamnolipid, that emulsified crude oil efficiently for effective degradation by the other members of consortium. The wide range of hydrocarbonoclastic capabilities of the selected members of bacterial consortium leads to the degradation of both aromatic and aliphatic fractions of crude oil in 72 hours. (Author)