Assessing the influence of personal preferences on the choice of vocabulary for natural language generation

Referring expression generation is the part of natural language generation that decides how to refer to the entities appearing in an automatically generated text. Lexicalization is the part of this process which involves the choice of appropriate vocabulary or expressions to transform the conceptual content of a referring expression into the corresponding text in natural language. This problem presents an important challenge when we have enough knowledge to allow more than one alternative. In those cases, we need some heuristics to decide which alternatives are more appropriate in a given situation. Whereas most work on natural language generation has focused on a generic way of generating language, in this paper we explore personal preferences as a type of heuristic that has not been properly addressed. We empirically analyze the TUNA corpus, a corpus of referring expression lexicalizations, to investigate the influence of language preferences in how people lexicalize new referring expressions in different situations. We then present two corpus-based approaches to solve the problem of referring expression lexicalization, one that takes preferences into account and one that does not. The results show a decrease of 50% in the similarity error against the reference corpus when personal preferences are used to generate the final referring expression.     

Interest,inferences, and learning from texts

Topic interest and learning from texts have been found to be positively associated with each other. However, the reason for this positive association is not well understood. The purpose of this study is to examine a cognitive process, inference generation, that could explain the positive association between interest and learning from texts. In Study 1, sixty undergraduate students participated by reading two science texts, which differed in coherence levels, silently. The results replicated previous findings that topic interest is positively associated with recall and accurate answers to comprehension questions for both texts. In Study 2, sixty-nine undergraduate students participated by reading the same two science texts while thinking aloud. The results indicated that topic interest was positively associated with inference generation while reading for the more coherently-written text. Subsequent analyses indicated inference generation partly explained the positive association between topic interest and accurate answers to comprehension questions for the more coherently-written text. The findings from Study 2 were independent of the effects of reading comprehension skill. Theoretical implications of the findings, in regard to standards of coherence and depth of processing while reading, are discussed.     

The mind behind the message: advancing theory-of-mind scales for typically developing children, and those with deafness, autism, or Asperger syndrome

Children aged 3-12 years (n = 184) with typical development, deafness, autism, or Asperger syndrome took a series of theory-of-mind (ToM) tasks to confirm and extend previous developmental scaling evidence. A new sarcasm task, in the format of H. M. Wellman and D. Liu's (2004) 5-step ToM Scale, added a statistically reliable 6th step to the scale for all diagnostic groups. A key previous finding, divergence in task sequencing for children with autism, was confirmed. Comparisons among diagnostic groups, controlling age, and language ability, showed that typical developers mastered the 6 ToM steps ahead of each of the 3 disabled groups, with implications for ToM theories. The final (sarcasm) task challenged even nondisabled 9-year-olds, demonstrating the new scale's sensitivity to post-preschool ToM growth.     

Food‐Based Science Curriculum Increases 4th Graders Multidisciplinary Science Knowledge

Health professionals and policymakers are asking educators to place more emphasis on food and nutrition education. Integrating these topics into science curricula using hand‐on, food‐based activities may strengthen students’ understanding of science concepts. The Food, Math, and Science Teaching Enhancement Resource (FoodMASTER) Initiative is a compilation of programs aimed at using food as a tool to teach mathematics and science. Previous studies have shown that students experiencing the FoodMASTER curriculum were very excited about the activities, became increasingly interested in the subject matter of food, and were able to conduct scientific observations. The purpose of this study was to: (1) assess 4th graders food‐related multidisciplinary science knowledge, and (2) compare gains in food‐related science knowledge after implementation of an integrated, food‐based curriculum. During the 2009–2010 school year, FoodMASTER researchers implemented a hands‐on, food‐based intermediate curriculum in eighteen 4th grade classrooms in Ohio (n = 9) and North Carolina (n = 9). Sixteen classrooms in Ohio (n = 8) and North Carolina (n = 8), following their standard science curricula, served as comparison classrooms. Students completed a researcher‐developed science knowledge exam, consisting of 13 multiple‐choice questions administered pre‐ and post‐test. Only subjects with pre‐ and post‐test scores were entered into the sample (Intervention n = 343; Control n = 237). No significant differences were observed between groups at pre‐test. At post‐test, the intervention group scored (9.95 ± 2.00) significantly higher (p = 0.000) than the control group (8.84 ± 2.37) on a 13‐point scale. These findings suggest the FoodMASTER intermediate curriculum is more effective than a standard science curriculum in increasing students’ multidisciplinary science knowledge related to food.