Physical Sciences and Mathematics Commons^™

Automatically Evaluating Content Selection In Summarization Without Human Models, Annie Louis, Ani Nenkova

Ani Nenkova

We present a fully automatic method for content selection evaluation in summarization that does not require the creation of human model summaries. Our work capitalizes on the assumption that the distribution of words in the input and an informative summary of that input should be similar to each other. Results on a large scale evaluation from the Text Analysis Conference show that input-summary comparisons are very effective for the evaluation of content selection. Our automatic methods rank participating systems similarly to manual model-based pyramid evaluation and to manual human judgments of responsiveness. The best feature, Jensen- Shannon divergence, leads to …

Entity-Driven Rewrite For Multi-Document Summarization, Ani Nenkova

Ani Nenkova

In this paper we explore the benefits from and shortcomings of entity-driven noun phrase rewriting for multi-document summarization of news. The approach leads to 20% to 50% different content in the summary in comparison to an extractive summary produced using the same underlying approach, showing the promise the technique has to offer. In addition, summaries produced using entity-driven rewrite have higher linguistic quality than a comparison non-extractive system. Some improvement is also seen in content selection over extractive summarization as measured by pyramid method evaluation.

Detecting Prominence In Conversational Speech: Pitch Accent, Givenness And Focus, Vivek Kumar Rangarajan Sridhar, Ani Nenkova, Shrikanth Narayanan, Dan Jurafsky

Ani Nenkova

The variability and reduction that are characteristic of talking in natural interaction make it very difficult to detect prominence in conversational speech. In this paper, we present analytic studies and automatic detection results for pitch accent, as well as on the realization of information structure phenomena like givenness and focus. For pitch accent, our conditional random field model combining acoustic and textual features has an accuracy of 78%, substantially better than chance performance of 58%. For givenness and focus, our analysis demonstrates that even in conversational speech there are measurable differences in acoustic properties and that an automatic detector for …

Modelling Prominence And Emphasis Improves Unit-Selection Synthesis, Volker Strom, Ani Nenkova, Robert Clark, Yolanda Vazquez-Alvarez, Jason Brenier, Simon King, Dan Jurafsky

Ani Nenkova

We describe the results of large scale perception experiments showing improvements in synthesising two distinct kinds of prominence: standard pitch-accent and strong emphatic accents. Previously prominence assignment has been mainly evaluated by computing accuracy on a prominence-labelled test set. By contrast we integrated an automatic pitch-accent classifier into the unit selection target cost and showed that listeners preferred these synthesised sentences. We also describe an improved recording script for collecting emphatic accents, and show that generating emphatic accents leads to further improvements in the fiction genre over incorporating pitch accent only. Finally, we show differences in the effects of prominence …

Predicting The Fluency Of Text With Shallow Structural Features: Case Studies Of Machine Tanslation And Human-Written Text, Jieun Chae, Ani Nenkova

Ani Nenkova

Sentence fluency is an important component of overall text readability but few studies in natural language processing have sought to understand the factors that define it. We report the results of an initial study into the predictive power of surface syntactic statistics for the task; we use fluency assessments done for the purpose of evaluating machine translation. We find that these features are weakly but significantly correlated with fluency. Machine and human translations can be distinguished with accuracy over 80%. The performance of pairwise comparison of fluency is also very high—over 90% for a multi-layer perceptron classifier. We also test …

Class-Level Spectral Features For Emotion Recognition, Dmitri Bitouk, Ragini Verma, Ani Nenkova

Ani Nenkova

The most common approaches to automatic emotion recognition rely on utterance-level prosodic features. Recent studies have shown that utterance-level statistics of segmental spectral features also contain rich information about expressivity and emotion. In our work we introduce a more fine-grained yet robust set of spectral features: statistics of Mel-Frequency Cepstral Coefficients computed over three phoneme type classes of interest – stressed vowels, unstressed vowels and consonants in the utterance. We investigate performance of our features in the task of speaker-independent emotion recognition using two publicly available datasets. Our experimental results clearly indicate that indeed both the richer set of spectral …

Performance Confidence Estimation For Automatic Summarization, Annie Louis, Ani Nenkova

Ani Nenkova

We address the task of automatically predicting if summarization system performance will be good or bad based on features derived directly from either single- or multi-document inputs. Our labelled corpus for the task is composed of data from large scale evaluations completed over the span of several years. The variation of data between years allows for a comprehensive analysis of the robustness of features, but poses a challenge for building a combined corpus which can be used for training and testing. Still, we find that the problem can be mitigated by appropriately normalizing for differences within each year. We examine …

'At-Risk' Places: Inequaties In The Distribution Of Environmental Stressors And Prescription Rates Of Mental Health Medications In Glasgow, Scotland, Juliana A. Maantay, Andrew Maroko

Publications and Research

Using geospatial analytical methods, this study examines the association between one aspect of the built environment, namely, the concentration of vacant and derelict land (VDL), and the prevalence of mental health disorders (using the proxy variable of mental health medication prescription rates) in Glasgow, Scotland. This study builds on our previous research, which demonstrated the spatial correspondence between the locations of VDL in Glasgow and several physical health outcomes. Numerous studies of other locales have found similar correspondence between different elements of the built environment and various health outcomes. This is the first study of its kind to look at …

Force-Enhanced Atomic Refinement: Structural Modeling With Interatomic Forces In A Reverse Monte Carlo Approach Applied To Amorphous Si And Sio2, A. Pandey, Parthapratim Biswas, D. A. Drabold

Faculty Publications

We introduce a structural modeling technique, called force-enhanced atomic refinement (FEAR). The technique incorporates interatomic forces in reverse Monte Carlo (RMC) simulations for structural refinement by fitting experimental diffraction data using the conventional RMC algorithm, and minimizes the total energy and forces from an interatomic potential. We illustrate the usefulness of the approach by studying a−SiO₂ and a−Si. The structural and electronic properties of the FEAR models agree well with experimental neutron and x-ray diffraction data and the results obtained from previous molecular dynamics simulations of a−SiO₂ and a−Si. We have shown that the method is more efficient …

Installation Instructions, Ian M. Anderson, Charles G. Torre

Downloads

DifferentialGeometry and DGApplications are contained in Maple library archives (mla files). The documentation is contained in Maple help files. Unzip the downloaded file (if necessary) to reveal the mla and help files.

To install a package, proceed as follows.

(1) Create or identify a directory (folder) where you will keep the mla file(s) and the help file. This should be a directory of your own, not belonging to Maple (e.g., do not use the Maple "lib" directory). Copy the mla and help files there.

(2) Open Maple, at the prompt type libname. This will display where maple currently looks for …

Policy And Planning Mechanisms For Coastal Relocation: Barriers And Opportunities, Anamaria Bukvic

October 30, 2015: Beyond Toolkits: Adaptation Strategies and Lessons

No abstract provided.

Panel Presentation: Norfolk: Thriving With Water, Norfolk Working Group

October 30, 2015: Beyond Toolkits: Adaptation Strategies and Lessons

No abstract provided.

Lessons From Cities On The Front Line Of Sea Level Rise, Diane Horn

October 30, 2015: Beyond Toolkits: Adaptation Strategies and Lessons

No abstract provided.

Corps Efforts On Sea Level Rise In Hampton Roads, Greg Steele

October 30, 2015: Beyond Toolkits: Adaptation Strategies and Lessons

No abstract provided.

Panel Presentation: Regional Working Group, Regional Working Group

October 30, 2015: Beyond Toolkits: Adaptation Strategies and Lessons

No abstract provided.

Agenda, Hr Adaptation Forum

October 30, 2015: Beyond Toolkits: Adaptation Strategies and Lessons

No abstract provided.

Coast-To-Interior Gradient In Recent Northwest Greenland Precipitation Trends (1952–2012), G J. Wong, E C. Osterberg, R L. Hawley, Z R. Courville, D G. Ferris, J A. Howley

Dartmouth Scholarship

The spatial and temporal variability of precipitation on the Greenland ice sheet is an essential component of surface mass balance, which has been declining in recent years with rising temperatures. We present an analysis of precipitation trends in northwest (NW) Greenland (1952–2012) using instrumental (coastal meteorological station) and proxy records (snow pits and ice cores) to characterize the precipitation gradient from the coast to the ice sheet interior. Snow-pit-derived precipitation near the coast (1950–2000) has increased (~7% decade⁻¹, p < 0.01) whereas there is no significant change observed in interior snow pits. This trend holds for 1981–2012, where calculated precipitation changes decrease in magnitude with increasing distance from the coast: 13% decade⁻¹ (2.4 mm water equivalent (w.e.) decade⁻²) at coastal Thule air base (AB), 8.6% decade⁻¹ (4.7 …

Estuarine Sediment Dynamics, C. T. Friedrichs

Presentations

No abstract provided.

The Firece Green Fire: Vol. 6 Issue 8, Wofford College Environmental Studies Program

The Fierce Green Fire

No abstract provided.

Analysis Of A Complex Mesovertex Evolution During The Second 30 June-1 July 2014 Derecho Event, Adam Clayton, Anthony Lyza

Von Braun Symposium Student Posters

No abstract provided.

Alto Orinoco Health & Air Quality Utilizing Nasa Earth Observations To Locate Remote Yanomami Villages, Sara Amirazodi

Von Braun Symposium Student Posters

No abstract provided.

Analysis Of The 14 July 2015 Nocturnal Convective Initiation Event During The Pecan Field Campaign, Ashley Ravenscraft

Von Braun Symposium Student Posters

No abstract provided.

Analysis Of The 5 July 2015 Nocturnal Convective Initiation Event During The Pecan Field Campaign, Alexander Staarmann

Von Braun Symposium Student Posters

No abstract provided.

Combining Diffusive Shock Acceleration And Stochastic Particle Energization By Multiple Plasmoids, Parisa Mostafavi

Von Braun Symposium Student Posters

No abstract provided.

Project Low Horizon - Horizon Detection Through The Use Of Infrared Sensors, Adam Bower, Evan Unruh, Justin Smith

Von Braun Symposium Student Posters

No abstract provided.

Searching For Terrestrial Gamma-Ray Flashes Via Lightning And The Fermi Space Telescope, Kareem Omar

Von Braun Symposium Student Posters

No abstract provided.

Tornadogenesis Within Hurricanes: A Study Of All Tropical Cyclone Tornadoes Between 2008 & 2015, Montana Etten-Bohm

Von Braun Symposium Student Posters

No abstract provided.

Study Solar Activity In The Backside Of The Sun, Catherine A. Gibbs

Von Braun Symposium Student Posters

No abstract provided.

Sunspot Proxy To Predict The Number Of Spes Per Year Based On Yearly Sunspot Number, Zachary Robinson

Von Braun Symposium Student Posters

No abstract provided.

Texas And Arizona Ecological Forecasting Utilizing Nasa Earth Observations To Monitor And Manage Ocelot Habitat Loss, Daryl Ann Winstead, Kaushik Narasimhan, Christina Fischer, Amberle Keith

Von Braun Symposium Student Posters

No abstract provided.