methods software as standards e g lda
play

Methods/Software as Standards e.g., LDA Lead: All Participants: - PowerPoint PPT Presentation

Breakout session on Methods/Software as Standards e.g., LDA Lead: All Participants: Andre Skupin, Margaret Clements, Katy Borner, Ying Ding, Stasa Milojevic State of the Art: Please List Existing Standards LDA for topic identification,


  1. Breakout session on Methods/Software as Standards e.g., LDA Lead: All Participants: Andre Skupin, Margaret Clements, Katy Borner, Ying Ding, Stasa Milojevic

  2. State of the Art: Please List Existing Standards LDA for topic identification, dimensionality reduction Quality: Can human users make sense of topics Can you use topics to support retrieval and predictions

  3. Identify Most Needed Standards Methods/Software to • Disambiguate, ddupe (http://www.cs.umd.edu/projects/linqs/ddupe) • Extract networks • Dimensionality reduction LDA (http://mallet.cs.umass.edu) NNF • Layout algorithms (KK, FR) • Rendering (GIS, Pajek) LDA computes from probability distribution, http://en.wikipedia.org/wiki/Latent_Dirichlet_allocation Non negative factorization does matrix analysis, http://en.wikipedia.org/wiki/Non-negative_matrix_factorization Arrive at same result

  4. Processes to Get these Standards De facto standards: ISI, .net data formats WC3 Standards – submit candidate standards proposals, review, approval. Democratization of standards to those than can pay membership. Bottom up – to get emergent areas Top-down – to impose structure, resolve mapping issues, good labels. Upper-level ontologies to meet in middle.

  5. Processes to Update/Maintain/Promote/Align These Standards Map must expose sockets Any artifact should have sockets. Need to know how to do proper cosine similarity calculation. • Identity (PaperID) • similarity (vis n-dim vector calculation) • semantic sockets (same semantic class/category) Take novel record and see probability landscape of where the new record should go. --- Drop in county data to state US ma. Map then aggregated to states. – Semantic socket. Yahoo geocoding. Yahoo gets address, delivers lat/long, goes to map that has Mercator projected system, gives image back. In geography, we can take maps from different projections and drag and drop. Hard to design sockets (resource) and easy to design plugs (users).

  6. Joseph, Katy et al build a giant socket system (standard) and some prototypical plugs in anticipation of future plugs/plugins.

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend