Phil 10.10.18

7:00 – 4:30 ASRC BD

  • Starting to add content to the proposal. Going to put together a section on game theory that ties together Beyond Individual Choice, The Evolution of Cooperation, and Consensus and Cooperation in Networked Multi-Agent Systems
    • Got some good writing done, but didn’t upload!
  • And also, voter influencing from this post
  • And I just saw this! Structure of Decision: The Cognitive Maps of Political Elites. It’s another book by Robert Axelrod. Ordered.
  • Putting together my notes on the Evolution of Cooperation. Can’t believe I haven’t done that yet
  • Got a good response from Antonio. Need to respond
  • Found some good stuff on market-oriented programming for Antonio’s workshop. The person who seems to really own this space is Michael Wellman (Scholar). Downloaded several of his papers.
  • From Benjamin Schmidt, via Twitter:
    • I have a new article in CA on dimensionality reduction for digital libraries. . Let me walk through one figure, Eames power-of-10 style, that shows a machine clustering of all 14 million books in the collection–including most of the books you’ve read.
    • 1-hathi_zoom_marked-683x1024 This is very close to mapping as I understand it. There is the ability to zoom in and out at different levels of structure. His repo is here, but its for [R]
    • Random Projection in Scikit-learn
    • Here’s the paper its based on: Visualizing Large-scale and High-dimensional Data
      • We study the problem of visualizing large-scale and high-dimensional data in a low-dimensional (typically 2D or 3D) space. Much success has been reported recently by techniques that first compute a similarity structure of the data points and then project them into a low-dimensional space with the structure preserved. These two steps suffer from considerable computational costs, preventing the state-of-the-art methods such as the t-SNE from scaling to large-scale and high-dimensional data (e.g., millions of data points and hundreds of dimensions). We propose the LargeVis, a technique that first constructs an accurately approximated K-nearest neighbor graph from the data and then layouts the graph in the low-dimensional space. Comparing to t-SNE, LargeVis significantly reduces the computational cost of the graph construction step and employs a principled probabilistic model for the visualization step, the objective of which can be effectively optimized through asynchronous stochastic gradient descent with a linear time complexity. The whole procedure thus easily scales to millions of high-dimensional data points. Experimental results on real-world data sets demonstrate that the LargeVis outperforms the state-of-the-art methods in both efficiency and effectiveness. The hyper-parameters of LargeVis are also much more stable over different data sets.