| Introduction | First, the model’s generative process encourages coherence in the local features and placement of relation instances. |
| Model | This section describes the generative process , while Sections 4 and 5 discuss declarative constraints. |
| Model | 3.2 Generative Process |
| Model | There are three steps to the generative process . |
| Constraints Shape Topics | In LDA, a document’s token is produced in the generative process by choosing a topic 2 and sampling a word from the multinomial distribution gbz of topic 2. |
| Constraints Shape Topics | If that is an unconstrained word, the word is emitted and the generative process for that token is done. |
| Constraints Shape Topics | Then the generative process for constrained LDA is: |
| Putting Knowledge in Topic Models | introduce ambiguity over the path associated with an observed token in the generative process . |
| Enriched Two-Tiered Topic Model | Thus; we present enriched TTM (ETTM) generative process (Fig.3), which samples words not only from low-level topics but also from high-level topics as well. |
| Enriched Two-Tiered Topic Model | Similar to TTM’s generative process , if wid is related to a given query, then cc 2 1 is deterministic, otherwise cc 6 {0,1} is stochastically determined if wid should be sampled as a background word (2123) or through hierarchical path, i.e., HT pairs. |
| Topic Coherence for Summarization | We identify sentences as meta-variables of document clusters, which the generative process models both sentences and documents using tiered topics. |