“…Much discussion revolves around how to choose the number of topics k , with some adopting what has been called a hierarchical Dirichlet process (HDP) (Teh et al, 2005). However, this process has been criticized for finding topics that group together terms that appear Figure 2. Finding ideal number of topics according to four different methodsunrelated (Deveaud, SanJuan & Bellot, 2014; Xie & Passonneau, 2012). It is also computationally expensive and adds a layer of complexity that can make interpretation difficult.…”
The UN Security Council (UNSC) has transformed from a body almost exclusively focused on conflict to one that addresses a wide variety of issues. Despite a series of powerful works in recent years showing how international norms have developed over time, we still lack clear understanding of why and when international institutions change their missions. This article argues that while international politics is usually characterized by inertia, shocks to the system, or focal point events, can compel rational actors to adopt new logics of appropriateness. Since 1945, the end of the Cold War and the signing of the Helsinki Accords stand out as such events. Through latent Dirichlet allocation, a machine learning algorithm used to classify text, UNSC resolutions between 1946 and 2017 can be divided into the subjects of War, Punitive, and Humanitarian. The topic Humanitarian exploded in frequency after the Cold War, and more refined models show that words related to human rights and elections similarly increased after Helsinki. These changes are rapid and occur in almost the immediate aftermath of focal point events, showing their importance for norm diffusion. The analysis also reveals another shift towards humanitarian topics in the mid-2000s, demonstrating the ability of topic modeling to uncover changes that have been missed by earlier kinds of analysis.
“…Much discussion revolves around how to choose the number of topics k , with some adopting what has been called a hierarchical Dirichlet process (HDP) (Teh et al, 2005). However, this process has been criticized for finding topics that group together terms that appear Figure 2. Finding ideal number of topics according to four different methodsunrelated (Deveaud, SanJuan & Bellot, 2014; Xie & Passonneau, 2012). It is also computationally expensive and adds a layer of complexity that can make interpretation difficult.…”
The UN Security Council (UNSC) has transformed from a body almost exclusively focused on conflict to one that addresses a wide variety of issues. Despite a series of powerful works in recent years showing how international norms have developed over time, we still lack clear understanding of why and when international institutions change their missions. This article argues that while international politics is usually characterized by inertia, shocks to the system, or focal point events, can compel rational actors to adopt new logics of appropriateness. Since 1945, the end of the Cold War and the signing of the Helsinki Accords stand out as such events. Through latent Dirichlet allocation, a machine learning algorithm used to classify text, UNSC resolutions between 1946 and 2017 can be divided into the subjects of War, Punitive, and Humanitarian. The topic Humanitarian exploded in frequency after the Cold War, and more refined models show that words related to human rights and elections similarly increased after Helsinki. These changes are rapid and occur in almost the immediate aftermath of focal point events, showing their importance for norm diffusion. The analysis also reveals another shift towards humanitarian topics in the mid-2000s, demonstrating the ability of topic modeling to uncover changes that have been missed by earlier kinds of analysis.
“…Moreover, in Bartlett et al (2010), it has been stated that nonparametric models with Pitman-Yor process priors cannot scale to large scale datasets. There are other proposed supervised nonparametric topic modeling approaches such as (Perotte et al 2011;Storkey and Dai 2014;Lakshminarayanan and Raich 2011;Xie and Passonneau 2012;Liao et al 2014;Acharya et al 2013). These models too cannot perform document retrieval learning task.…”
One limitation of most existing probabilistic latent topic models for document classification is that the topic model itself does not consider useful side-information, namely, class labels of documents. Topic models, which in turn consider the side-information, popularly known as supervised topic models, do not consider the word order structure in documents. One of the motivations behind considering the word order structure is to capture the semantic fabric of the document. We investigate a low-dimensional latent topic model for document classification. Class label information and word order structure are integrated into a supervised topic model enabling a more effective interaction among such information for solving document classification. We derive a collapsed Gibbs sampler for our model. Likewise, supervised topic models with word order structure have not been explored in document retrieval learning. We propose a novel supervised topic model for document retrieval learning which can be regarded as a pointwise model for tackling the learning-to-rank task. Available relevance assessments and word order structure are integrated into the topic model itself. We conduct extensive experiments on several publicly available benchmark datasets, and show that our model improves upon the state-of-the-art models.
“…An HDP is a non-parametric approach to topic modelling which automatically learns the number of topics from data. Applied to natural language processing, Xie and Rassoneau [17] proposed a semisupervised HDP model, where the "label" is the distribution of topics of the words -effectively a word-level labeling. Thus, this model is not directly applicable to document classification tasks of the type common in computer vision.…”
We present an extension to the Hierarchical Dirichlet Process (HDP), which allows for the inclusion of supervision. Our model marries the non-parametric benefits of HDP with those of Supervised Latent Dirichlet Allocation (SLDA) to enable learning the topic space directly from data while simultaneously including the labels within the model. The proposed model is learned using variational inference which allows for the efficient use of a large training dataset. We also present the online version of variational inference, which makes the method scalable to very large datasets. We show results comparing our model to a traditional supervised parametric topic model, SLDA, and show that it outperforms SLDA on a number of benchmark datasets.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
