Brendan O’Connor defended his Ph. In this work, we propose an integrated model to leverage side information. Dani Yogatama defended his Ph. Vancouver, BC, Canada abstract bibtex paper supplementary Online debate forums are important social media for people to voice their opinions and debate with each other. The model reveals latent factions, or groups of individuals whom we expect to collaborate more closely within their faction, cite within the faction using language distinct from citation outside the faction, and be largely understandable through the language used when cited from without. Humans write lawyers submit briefs, legislators draft bills, scientists publish papers, teenagers tweet, and politicians give speeches which are first written by their speechwriters for a multitude of purposes. He is an assistant professor at Georgetown University.

He is an assistant professor at the Toyota Technological Institute at Chicago. In our selection strategy, an informative and diverse set of instances are selected for effective disambiguation. Kevin Gimpel defended his Ph. Our proposed method is a regression-based latent factor model which jointly models user arguments, interactions, and attributes. Despite the sparsity of user stances, users may provide rich side information, for example, users may write arguments to back up their stances, interact with each other, and provide biographical information.

Brendan O’Connor defended his Ph. He is a researcher at AI2. He is an assistant professor at Georgetown University. He is head of research at Unbabel. ARK researchers and alumni are leaders in natural language processing, machine learning, and computational social science. Hao PengUW Ph.


LTI PhD Thesis Defense: Dani Yogatama

The results are validated against a set of preregistered, domain expert authored yogata,a. We evaluate the resulting representation’s usefulness in attaching opinionated documents thesks arguments and its consistency with human judgements about positions. People write everyday — articles, blogs, emails — with a purpose and an audience in mind. It is implemented in the form of a finite state transducer using the open source finite state compiler FOMA toolkit.

SmithJing Jiang. It supports compiling LaTeX, gnuplot and eps files, and displaying it to the console in a clean manner.

dani yogatama thesis

In this paper, we propose three advancements for entity linking. We present a statistical model for canonicalizing named entity mentions into a table whose rows represent entities and whose columns are attributes or parts of attributes. Despite the sparsity of user stances, users may provide rich side information, for example, users may write arguments to back up their stances, interact with each yogataama, and provide biographical information.

He is a data scientist at Civis Analytics. The model is novel in that it incorporates entity context, surface features, first-order dependencies among attribute-parts, and a notion of noise. Following, we use the Supreme Court as a case study to incorporate utility functions into models of text.

dani yogatama thesis

Both results incorporate the new technologies: Jungo KasaiUW Ph. Instead, this thesis proposes to take a decision theoretic approach to jointly model both author’s social attributes and responses from text — we treat text as a choice variable and model an individual author’s text production as a utility function. Lastly, we discuss methods and challenges to characterize authors’ strategic behavior in the domains of scientific community and judicial politics.


He is an theiss professor at the University of California at Berkeley. Variants of our model lead to improved predictive accuracy of citations given texts and texts given authors. The social attributes of an author deeply influence and bias his language production; while authors are motivated to evoke responses from his audience.

Yanchuan Sim, Bryan R.

Arxiv Sanity Preserver

Tae Yano defended her Ph. We consider three particular domains — political campaigns, the scientific community, and the judiciary — using our models and develop the necessary tools to evaluate our assumptions danu hypotheses. Sofia Serrano UW Ph.

dani yogatama thesis

We introduce a generative model positing latent topics and cross-cutting positions that gives special treatment to person mentions and opinion words. Throughout this proposal, we will present several examples of strategic behavior and how we can model it computationally. He is a lecturer at the University of Edinburgh.

To make this theais concrete, we consider the societally important decisions of the Supreme Court of the United States.

Dipanjan Das defended his Ph. He is a research scientist at Semantic Machines.