Applying semantic knowledge to the automatic processing of temporal expressions and events in natural language

This paper addresses the problem of the automatic recognition and classification of temporal expressions and events in human language. Efficacy in these tasks is crucial if the broader task of temporal information processing is to be successfully performed. We analyze whether the application of semantic knowledge to these tasks improves the performance of current approaches. We therefore present and evaluate a data-driven approach as part of a system: TIPSem. Our approach uses lexical semantics and semantic roles as additional information to extend classical approaches which are principally based on morphosyntax. The results obtained for English show that semantic knowledge aids in temporal expression and event recognition, achieving an error reduction of 59% and 21%, while in classification the contribution is limited. From the analysis of the results it may be concluded that the application of semantic knowledge leads to more general models and aids in the recognition of temporal entities that are ambiguous at shallower language analysis levels. We also discovered that lexical semantics and semantic roles have complementary advantages, and that it is useful to combine them. Finally, we carried out the same analysis for Spanish. The results obtained show comparable advantages. This supports the hypothesis that applying the proposed semantic knowledge may be useful for different languages.

Seismically-induced landslides by a low-magnitude earthquake: The Mw 4.7 Ossa De Montiel event (central Spain)

The Ossa de Montiel (2015/02/23, Mw 4.7) earthquake struck the central part of Spain and was felt far from the epicenter (> 300 km). Even though ground shaking was slight (Imax = V, EMS-98 scale), the earthquake triggered many small rock falls, most at distances of 20–30 km from the epicenter, greater than previously recorded in S Spain (16 km) for earthquakes of similar magnitudes. The comparative analysis of available data for this event with records from other quakes of the Betic cordillera (S and SE Spain) seems to indicate a slower pattern of ground-motion attenuation in central Spain. This could explain why slope instabilities occurred at larger distances. Instability was more frequent, and occurred at larger distances, in road cuts than in natural slopes, implying that such slope types are highly susceptible to seismically induced landslides.

Global assessment of coral bleaching and required rates of adaptation under climate change

Elevated ocean temperatures can cause coral bleaching, the loss of colour from reef-building corals because of a breakdown of the symbiosis with the dinoflagellate Symbiodinium. Recent studies have warned that global climate change could increase the frequency of coral bleaching and threaten the long-term viability of coral reefs. These assertions are based on projecting the coarse output from atmosphere-ocean general circulation models (GCMs) to the local conditions around representative coral reefs. Here, we conduct the first comprehensive global assessment of coral bleaching under climate change by adapting the NOAA Coral Reef Watch bleaching prediction method to the output of a low- and high-climate sensitivity GCM. First, we develop and test algorithms for predicting mass coral bleaching with GCM-resolution sea surface temperatures for thousands of coral reefs, using a global coral reef map and 1985-2002 bleaching prediction data. We then use the algorithms to determine the frequency of coral bleaching and required thermal adaptation by corals and their endosymbionts under two different emissions scenarios. The results indicate that bleaching could become an annual or biannual event for the vast majority of the world's coral reefs in the next 30-50 years without an increase in thermal tolerance of 0.2-1.0 degrees C per decade. The geographic variability in required thermal adaptation found in each model and emissions scenario suggests that coral reefs in some regions, like Micronesia and western Polynesia, may be particularly vulnerable to climate change. Advances in modelling and monitoring will refine the forecast for individual reefs, but this assessment concludes that the global prognosis is unlikely to change without an accelerated effort to stabilize atmospheric greenhouse gas concentrations.

Using intervention time series analyses to assess the effects of imperfectly identifiable natural events: A general method and example

BACKGROUND: Intervention time series analysis (ITSA) is an important method for analysing the effect of sudden events on time series data. ITSA methods are quasi-experimental in nature and the validity of modelling with these methods depends upon assumptions about the timing of the intervention and the response of the process to it. METHOD: This paper describes how to apply ITSA to analyse the impact of unplanned events on time series when the timing of the event is not accurately known, and so the problems of ITSA methods are magnified by uncertainty in the point of onset of the unplanned intervention. RESULTS: The methods are illustrated using the example of the Australian Heroin Shortage of 2001, which provided an opportunity to study the health and social consequences of an abrupt change in heroin availability in an environment of widespread harm reduction measures. CONCLUSION: Application of these methods enables valuable insights about the consequences of unplanned and poorly identified interventions while minimising the risk of spurious results.

Contradictions inherent in the management of natural and industrial disasters

These days people keep wondering whether the world is more dangerous now than it was before. Do natural disasters really happen more frequently or is it just that the damage they cause that has become greater? The situation is not quite clear. As a result of the globalizing world and advanced communication infrastructure, the number of known / reported catastrophes is relatively high, but that does not necessarily mean there has been an actual increase in frequency. The red mud spill in Hungary was a special combination of industrial and natural disasters. This is one of the reasons why it is very hard to pinpoint who is responsible for the event. Natural disasters tend to raise questions about responsibility that are different from those concerning industrial catastrophes. Interestingly enough, however, nature often plays an important role in industrial disasters. The present article is concerned with how the issues of responsibility are handled in the case of industrial disasters.

Environmental data from coastal sediments along natural CO2 gradients at a volcanic vent in Papua New Guinea

Natural CO2 venting systems can mimic conditions that resemble intermediate to high pCO2 levels as predicted for our future oceans. They represent ideal sites to investigate potential long-term effects of ocean acidification on marine life. To test whether microbes are affected by prolonged exposure to pCO2 levels, we examined the composition and diversity of microbial communities in oxic sandy sediments along a natural CO2 gradient. Increasing pCO2 was accompanied by higher bacterial richness and by a strong increase in rare members in both bacterial and archaeal communities. Microbial communities from sites with CO2 concentrations close to today's conditions had different structures than those of sites with elevated CO2 levels. We also observed increasing sequence abundance of several organic matter degrading types of Flavobacteriaceae and Rhodobacteraceae, which paralleled concurrent shifts in benthic cover and enhanced primary productivity. With increasing pCO2, sequences related to bacterial nitrifying organisms such as Nitrosococcus and Nitrospirales decreased, and sequences affiliated to the archaeal ammonia-oxidizing Thaumarchaeota Nitrosopumilus maritimus increased. Our study suggests that microbial community structure and diversity, and likely key ecosystem functions, may be altered in coastal sediments by long-term CO2 exposure to levels predicted for the end of the century.

Phenotypical responses of Fucus vesiculosus germlings to warming, acidification, nutrient enrichment and local upwelling under natural fluctuations (July - September 2014)

During summer 2014 (mid-July - mid-September 2014), early life-stage Fucus vesiculosus were exposed to combined ocean acidification and warming (OAW) in the presence and absence of enhanced nutrient levels (OAW x N experiment). Subsequently, F. vesiculosus germlings were exposed to a final upwelling disturbance during 3 days (mid-September 2014). Experiments were performed in the near-natural scenario "Kiel Outdoor Benthocosms" including natural fluctuations in the southwestern Baltic Sea, Kiel Fjord, Germany (54°27 'N, 10°11 'W). Genetically different sibling groups and different levels of genetic diversity were employed to test to which extent genetic variation would result in response variation. The data presented here show the phenotypical response (growth and survival) of the different experimental populations of F. vesiculosus under OAW, nutrient enrichment and the upwelling event. Log effect ratios demonstrate the responses to enhanced OAW and nutrient concentrations relative to the ambient conditons. Carbon, nitrogen content (% DW) and C:N ratios were measured after the exposure of ambient and high nutrient levels. Abiotic conditions the OAW x nutrient experiment and the upwelling event, are shown.

Nitrite consumption and associated isotope changes during a river flood event in the Elbe river

In oceans, estuaries, and rivers, nitrification is an important nitrate source, and stable isotopes of nitrate are often used to investigate recycling processes (e.g. remineralisation, nitrification) in the water column. Nitrification is a two-step process, where ammonia is oxidised via nitrite to nitrate. Nitrite usually does not accumulate in natural environments, which makes it difficult to study the single isotope effect of ammonia oxidation or nitrite oxidation in natural systems. However, during an exceptional flood in the Elbe River in June 2013, we found a unique co-occurrence of ammonium, nitrite, and nitrate in the water column, returning towards normal summer conditions within 1 week. Over the course of the flood, we analysed the evolution of d15N-[NH4]+ and d15N-[NO2]- in the Elbe River. In concert with changes in suspended particulate matter (SPM) and d15N SPM, as well as nitrate concentration, d15N-NO3 - and d18O-[NO3] -, we calculated apparent isotope effects during net nitrite and nitrate consumption. During the flood event, > 97 % of total reactive nitrogen was nitrate, which was leached from the catchment area and appeared to be subject to assimilation. Ammonium and nitrite concentrations increased to 3.4 and 4.4 µmol/l, respectively, likely due to remineralisation, nitrification, and denitrification in the water column. d15N-[NH4]+ values increased up to 12 per mil, and d15N-[NO2]- ranged from -8.0 to -14.2 per mil. Based on this, we calculated an apparent isotope effect 15-epsilon of -10.0 ± 0.1 per mil during net nitrite consumption, as well as an isotope effect 15-epsilon of -4.0 ± 0.1 per mil and 18-epsilon of -5.3 ± 0.1 per mil during net nitrate consumption. On the basis of the observed nitrite isotope changes, we evaluated different nitrite uptake processes in a simple box model. We found that a regime of combined riparian denitrification and 22 to 36 % nitrification fits best with measured data for the nitrite concentration decrease and isotope increase.

Discrete-event Simulation and Optimization to Improve the Performance of a Healthcare System

Thesis (Ph.D.)--University of Washington, 2016-08

Optimization and Applications of Large-Scale Biomedical Event Networks

The overwhelming amount and unprecedented speed of publication in the biomedical domain make it difficult for life science researchers to acquire and maintain a broad view of the field and gather all information that would be relevant for their research. As a response to this problem, the BioNLP (Biomedical Natural Language Processing) community of researches has emerged and strives to assist life science researchers by developing modern natural language processing (NLP), information extraction (IE) and information retrieval (IR) methods that can be applied at large-scale, to scan the whole publicly available biomedical literature and extract and aggregate the information found within, while automatically normalizing the variability of natural language statements. Among different tasks, biomedical event extraction has received much attention within BioNLP community recently. Biomedical event extraction constitutes the identification of biological processes and interactions described in biomedical literature, and their representation as a set of recursive event structures. The 2009–2013 series of BioNLP Shared Tasks on Event Extraction have given raise to a number of event extraction systems, several of which have been applied at a large scale (the full set of PubMed abstracts and PubMed Central Open Access full text articles), leading to creation of massive biomedical event databases, each of which containing millions of events. Sinece top-ranking event extraction systems are based on machine-learning approach and are trained on the narrow-domain, carefully selected Shared Task training data, their performance drops when being faced with the topically highly varied PubMed and PubMed Central documents. Specifically, false-positive predictions by these systems lead to generation of incorrect biomolecular events which are spotted by the end-users. This thesis proposes a novel post-processing approach, utilizing a combination of supervised and unsupervised learning techniques, that can automatically identify and filter out a considerable proportion of incorrect events from large-scale event databases, thus increasing the general credibility of those databases. The second part of this thesis is dedicated to a system we developed for hypothesis generation from large-scale event databases, which is able to discover novel biomolecular interactions among genes/gene-products. We cast the hypothesis generation problem as a supervised network topology prediction, i.e predicting new edges in the network, as well as types and directions for these edges, utilizing a set of features that can be extracted from large biomedical event networks. Routine machine learning evaluation results, as well as manual evaluation results suggest that the problem is indeed learnable. This work won the Best Paper Award in The 5th International Symposium on Languages in Biology and Medicine (LBM 2013).

"Natural" disasters as (neo-liberal) opportunity? Discussing post-hurricane Katrina urban regeneration in New Orleans

By providing a wide literature review, post-hurricane Katrina uneven urban regeneration in New Orleans is presented here by framing it within a historical perspective in order to underline how environmental threats too often seem to be not so much “natural” but rather man-made as well as to highlight both the reasons and the ways in which, in post-disaster reconstruction, competitive growth has been valued over equity, by directly benefiting those who were already the most advantaged. The aim is to highlight how environmental disasters can be considered as socially constructed phenomena, as they cannot be seen as a single event but rather as a process made by a series of progressive steps occurring within different spheres, which do not necessarily concern the environment only.

Surveillance in Cyberspace: Applying Natural and Place Manager Surveillance to System Trespassing

Research on the criminological side of system trespassing (i.e. unlawfully gaining access to a computer system) is relatively rare and has yet to examine the effect of the presence of other users on the system during the trespassing event (i.e. the time of communication between a trespasser’s system and the infiltrated system). This thesis seeks to analyze this relationship drawing on principles of Situational Crime Prevention, Routine Activities Theory, and restrictive deterrence. Data were collected from a randomized control trial of target computers deployed on the Internet network of a large U.S. university. This study examined whether the number (one or multiple) and type (administrative or non-administrative) of computer users present on a system reduced the seriousness and frequency of trespassing. Results indicated that the type of user (administrative) produced a restrictive deterrent effect and significantly reduced the frequency and duration of trespassing events.