Refocusing on Relevance: Personalization in NLG.

Many NLG tasks such as summarization, dialogue response, or open domain question answering focus primarily on a source text in order to generate a target response. This standard approach falls short, however, when a user's intent or context of work is not easily recoverable based solely on that source text-a scenario that we argue is more of the rule than the exception. In this work, we argue that NLG systems in general should place a much higher level of emphasis on making use of additional context, and suggest that relevance (as used in Information Retrieval) be thought of as a crucial tool for designing user-oriented text-generating tasks. We further discuss possible harms and hazards around such personalization, and argue that value-sensitive design represents a crucial path forward through these challenges.

COBrA: a bio-ontology editor.

COBrA is a Java-based ontology editor for bio-ontologies that distinguishes itself from other editors by supporting the linking of concepts between two ontologies, and providing sophisticated analysis and verification functions. In addition to the Gene Ontology and Open Biology Ontologies formats, COBrA can import and export ontologies in the Semantic Web formats RDF, RDFS and OWL.

Computer-generated trauma management plans: comparison with actual care.

TraumAID is a computer-based decision aid that uses decision rules and logical deduction to generate management plans for the initial definitive management of injured patients; its use is currently confined to assessing penetrating thoracoabdominal injuries in nonpregnant adults. These management plans were compared to trauma center care in an American urban medical school hospital. The TraumAID program was available to trauma chief residents for patient care during the "golden hour." Resulting changes in plans were documented by the residents. For 15 months the management plans for all applicable patients were documented serially, as were computer plans and outcomes. The sequential care and computer-generated care plans were then blinded and judged by three other trauma surgeons as to acceptability and preference. A consecutive series of 97 patients was evaluated. The residents used the computer for 40 cases during patient care. In 5 of the 40 cases, they altered their evaluation, diagnosis, or treatment; in none of these 5 was the alteration judged an error. Of the 97 patients, 10 had adverse outcomes, 2 of which were judged potentially avoidable, with unacceptable errors in management. TraumAID's plans were acceptable for both. In 31 cases previously managed by the judges themselves, the TraumAID plans were preferred by the judge to the care he or she had provided, by a ratio of 25:6 (p < 0.01). In a preliminary assessment, computer-generated, patient-specific protocols for the acute management of injuries were preferred to actual care previously given by the judges themselves and were associated with improved care and potential improvement in outcome.

Combining geometric and probabilistic reasoning for computer-based penetrating-trauma assessment.

OBJECTIVE: To ascertain whether three-dimensional geometric and probabilistic reasoning methods can be successfully combined for computer-based assessment of conditions arising from ballistic penetrating trauma to the chest and abdomen. DESIGN: The authors created a computer system (TraumaSCAN) that integrates three-dimensional geometric reasoning about anatomic likelihood of injury with probabilistic reasoning about injury consequences using Bayesian networks. Preliminary evaluation of TraumaSCAN was performed via a retrospective study testing performance of the system on data from 26 cases of actual gunshot wounds. MEASUREMENTS: Areas under the receiver operating characteristics (ROC) curve were calculated for each condition modeled in TraumaSCAN that was present in the 26 cases. The comprehensiveness and relevance of the TraumaSCAN diagnosis for the 26 cases were used to assess the overall performance of the system. To test the ability of TraumaSCAN to handle limited findings, these measurements were calculated both with and without input of observed findings into the Bayesian network. RESULTS: For the 11 conditions assessed, the worst area under the ROC curve with no observed findings input into the Bayesian network was 0.542 (95% CI, 0.146-0.937), the median was 0.883 (95% CI, 0.713-1.000), and the best was 1.00 (95% CI, 1.000-1.000). The worst area under the ROC curve with all observed findings input into the Bayesian network was 0.835 (95% CI, 0.602-1.000), the median was 0.941 (95% CI, 0.827-1.000), and the best was 0.992 (95% CI, 0.965-1.000). A comparison of the areas under the curve obtained with and without input of observed findings into the Bayesian network showed that there were significant differences for 2 of the 11 conditions assessed. CONCLUSION: A computer-based method that combines geometric and probabilistic reasoning shows promise as a tool for assessing ballistic penetrating trauma to the chest and abdomen.