• theadvisor

    We tried to address the following problems, which we believe almost every PhD student, researcher, even professor has gone through while writing an academic paper:

    • Did I miss any important paper?
      With the current keyword-based search methods of digital libraries, it is very likely to miss important papers because of the use of different terminologies.
    • Where should I submit this paper? What is the best conference/journal for it?
      Even if you decide on one, it is always good to keep the alternatives in mind.
    • Whom should I recommend as a reviewer to this paper?
      It is not common in most computer science-related submissions, but there are certain journals that ask for 2-5 reviewer names. You may also want to follow the work of these people.

    and we answer those questions with serious graph algorithms and probabilistic calculations on hundreds of thousands of papers. Think of theadvisor as an extraordinary person who knows all the papers in a field, and you got a chance to ask his/her opinion about your paper.

    While answering the above question, we provide a clean yet powerful interface to display citation information and direct links to most of the papers.

    As you get new paper recommendations, you will be able to give a feedback to the system by telling:

    • "this is definitely relevant to my topic and I would like to see others related to this one", or
    • "this is an irrelevant paper, you can do better than that!".
  • Members

    • Ümit V. Çatalyürek
    • Professor
    • The Ohio State University
      Biomedical Informatics, Electrical and Computer Engineering
    • http://bmi.osu.edu/~umit
  • Data

    theadvisor currently uses two different datasets:

    1. The first one is computer science centric, although it also has limited number of math and physics papers. A list of papers and the citation information between those papers are obtained through DBLP, arXiv, Inria, and CiteSeer. The linked pdf files are hosted by CiteSeer, while the bibtex entries are hosted by DBLP.
    2. The second dataset comes from PubMed Central Open Access Subset. This is a publicly available subset of PMC and PubMed articles.

    We are only covering a fraction of the existing academic documents. We are looking for more data to improve our service. If you know how we can gain access to more bibliography databases, please contact us!

    Quality

    We are trying to improve the quality of the metadata as we merge information from different sources, but it is extremely hard to do when you are dealing with millions of records. We are aware that some author names appear more than once with different encodings in reviewer recommendation, some papers don't have titles, and the service might even suggest a part of a journal name as a reviewer.

    The quality of the metadata we display (including titles, author names, venue names, etc.) depends on the quality of the data we gather. Although some services (e.g., DBLP) provide very high-quality data with name disambiguation, some of them are extracted from processed PDF files; hence, may include various mistakes.

    We suggest that you report or manually correct the errors in CiteSeer if you come across a record with a weird-looking author list, title, or venue.

  • Acknowledgments

  • Publications

    The followings are the publications related to our service. Please cite one of them when referring to the service in your papers.

    1. O. Kucuktunc, E. Saule, K. Kaya, U.V. Catalyurek, Diversified Recommendation on Graphs: Pitfalls, Measures, and Algorithms, Proc. 22nd International World Wide Web Conference (WWW) (acceptance rate 15%), 2013 (to appear).
    2. O. Kucuktunc, E. Saule, K. Kaya, U.V. Catalyurek, Result diversification in automatic citation recommendation, Proc. iConference Workshop on Computational Scientometrics: Theory and Applications, 2013.
    3. O. Kucuktunc, E. Saule, K. Kaya, U.V. Catalyurek, Direction Awareness in Citation Recommendation, Proc. 6th International Workshop on Ranking in Databases (DBRank'12) in conjunction with VLDB'12, 2012.
    4. O. Kucuktunc, K. Kaya, E. Saule, U.V. Catalyurek, Fast Recommendation on Bibliographic Networks, Proc. IEEE/ACM International Conference on Social Networks Analysis and Mining (ASONAM'12) (acceptance rate 16%), 2012.
    5. O. Kucuktunc, E. Saule, K. Kaya, U.V. Catalyurek, Diversifying Citation Recommendations, arxiv:1209.5809, Sep 2012.
    6. O. Kucuktunc, E. Saule, K. Kaya, U.V. Catalyurek, Recommendation on Academic Networks using Direction Aware Citation Analysis, arxiv:1205.1143, May 2012.

  • API

    API provides users and third-party application developers an interface to query our citation database. API reference

  • Contact Us

    For any comments or if you have any questions, do not hesitate to contact us by email at:

    hpc@bmi.osu.edu

    Want to give feedback?

    We would like to get some feedback about usability of the service and the recommendation quality. The followings are the questions we ask to our friends, if you want to keep in mind while testing the service.

    1. How was the usability of the website? Did you think it would be better if there were more instructions?
    2. How was the response time between submitting your bibliography file and getting all the recommendations?
    3. Were the citation recommendations good? Would you expand your reference list with some of the recommended papers?
    4. How relevant were the venue recommendations? If you had actually submitted the paper for that reference list, was the venue in the list?
    5. Did you get relevant names in your field from reviewer recommendation? If you (your advisor) have a couple of papers in the field, did you see your (your advisor's) name in the list?
    6. Did you select relevant papers in the first page and give feedback to refine the search? How many times did you do that? Did you get some other relevant papers?
    7. What new features would you like in the service?
    8. Would you use theadvisor in the future?

  • Frequently Asked Questions

    A. Field, Area, and Data Related

    1. Papers from which fields can be searched and explored in your data?
      We have mainly data-mined publications in Computer Science and Biomedical literature. We also have some limited coverage of Mathematics and Theoretical Physics.
    2. Are you planning to include papers in field X in the near future?
      As we find open access databases and reliable resources for other fields, we will continue to expand our service. If you come across such data, let us know.

    B. Input and Output Related

    1. Which bibliography format can I use with theadvisor?
      theadvisor can currently work with bibliography in BibTeX (bbl optionally), RIS and EndNote (exported to XML).
    2. Will you support bibliography format X?
      If you wish us to include another format, let us know.
    3. I would like to search for papers within the database!
      We will soon include a feature to search the database based on author names and keywords. The user could then add these papers to his/her bibliography.
    4. Can I export the papers I found in any bibliography format?
      theadvisor can not export the bibliography in BibTeX or RIS format yet. However, you can access the list of papers that have been selected as relevant by using the "Export relevant papers so far" button. This list features links to BibTeX entries on DBLP website when possible.

    C. Search related

    1. I submitted a bibliography file but theadvisor does not suggest anything!
      It seems that none of the papers included in your bibliography file are known by theadvisor. theadvisor's database does not contain all the academic papers. Follow this link to see which datasets are included. If you can provide us access to more data, let us know! It is also possible that the bibliography file you provided does not spell the information in the same way our database do.
    2. After refining the search a few times, no more suggestions are provided!
      When you refine a search, the papers that are checked are declared as relevant and are added to your bibliography. The ones that are not checked are declared irrelevant and removed from the database (as far as your search is concerned). After removing many papers from the database, there might be no more papers that are reachable from your bibliography.

theadvisor

is a graph-based citation, venue, and reviewer recommendation service developed by researchers of High Performance Computing Lab at The Ohio State University. more info

Frequently Asked

Have question? We have answers.

Contact

Let us know how you like the service: contact us