(ii) Significant hardship exception. CMS may, on a case-by-case basis, exempt an eligible professional (or in the case of a group practice under paragraph (e) of this section, a group practice) from the application of the payment adjustment under paragraph (c)(2) of this section if, CMS determines, subject to annual renewal, that compliance with the requirement for being a successful electronic...

Existing deep learning-based models for image captioning typically consist of an encoder to extract visual features and a language model decoder, architecture that has shown promising results in single high-level sentence generation. However, only the word-level guiding signal is available when optimized features. The inconsistency between parallel extraction sequential text supervision limits ...

Nowadays, graphical forms are used to capture input data and feed traditional databases. They are associated with rigid schema and constraints to ensure data validity and to control entry sequencce. Over the Internet, electronic documents are becoming widely exchanged and their data reused in a large range of tasks. As a direct consequence, a capture interface that relies on documents can serve...

We present a taxonomy and classification system for distinguishing between different types of paragraphs in movie reviews: formal vs. functional paragraphs and, within the latter, between description and comment. The classification is used for sentiment extraction, achieving improvement over a baseline without paragraph classification.

This paper describes TextTiling, an algorithm for partitioning expository texts into coherent multi-paragraph discourse units which reeect the subtopic structure of the texts. The algorithm uses domain-independent lexical frequency and distribution information to recognize the interactions of multiple simultaneous themes. Two fully-implemented versions of the algorithm are described and shown t...