search

earlng / academic-pdf-scrap

Code that scraps the contents of the PDF papers submitted for NeurIPS 2020
MIT License
4 stars 2 forks source link

Included acknowledgement #18

Closed earlng closed 3 years ago

earlng commented 3 years ago

For some papers, the new code pulls in the acknowledgement section (which usually follows the BIS) although we don’t want that

  1. A89b71bb5227c75d463dd82a03115738
  2. D3b1fb02964aa64e257f9f26a31f72cf
  3. D40d35b3063c11244fbf38e9b55074be
  4. F52a7b2610fb4d3f74b4106fb80b233d
  5. fea16e782bc1b1240e4b3c797012e289
earlng commented 3 years ago

A89b71bb5227c75d463dd82a03115738

The reason this happened is because "Acknowledgements" is in the same region class

      <section class="DoCO:Section">
        <h1 class="DoCO:SectionTitle" id="103" page="9" column="1">Broader Impact</h1>
        <region class="DoCO:TextChunk" id="104" page="9" column="1">This paper is a theoretical study that brings together two seemingly disjoint but equally impact- ful fields of sparse recovery and mixture models: the first having numerous applications in signal processing while the second being the main statistical model for clustering. Given that, this work belongs to the foundational area of data science and enhances our understanding of some basic the- oretical questions. We feel the methodology developed in this paper is instructive, and exemplifies the use of several combinatorial objects and techniques in signal recovery and classification, that are hitherto underused. Therefore we foresee the technical content of this paper to form good teach- ing material in foundational data science and signal processing courses. The content of this paper can raise interest of students or young researchers in discrete mathematics to applications areas and problems of signal processing and machine learning. While primarily of theoretical interest, the results of the paper can be immediately applicable to some real-life scenarios and be useful in recommendation systems, one of the major drivers of data science research. In particular, if in any case of feedback/rating from users of a service there is ambiguity about the source of the feedback, our framework can be used. This is also applicable to crowdsourcing applications.</region>
        <region class="DoCO:TextChunk" id="105" confidence="possible" page="9" column="1">Acknowledgements: This research is supported in part by NSF CCF 1909046 and NSF 1934846.</region>
      </section>
paulsedille commented 3 years ago

There is a "quick fix" for me to identify where this mistake might be happening in other impact statements (searching for the word "acknowledgement" in my sheet's BIS) so I will mark this closed for now. Low priority.