http://asymptoticlabs.com/enContents © 2022 <a href="mailto:quidditymaster@gmail.com">Tim Anderton</a> Wed, 31 Aug 2022 21:28:47 GMTNikola (getnikola.com)http://blogs.law.harvard.edu/tech/rsshttp://asymptoticlabs.com/posts/visualizing-convolution-kernels.htmlTim Anderton<div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt"> </div><div class="inner_cell"> <div class="text_cell_render border-box-sizing rendered_html"> <p>I very rarely see any sort of inspection being done on the convolutional kernels of a CNN. In part this is because the parameters themselves are far more difficult to interpret than the outputs of a network (even intermediate outputs a.k.a the network activations). This difficulty of interpretation is worst for kernels with a small spatial footprint and unfortunately 3x3 kernels are the most performant and popular choice. Trying to understand the structure of a 3x3 convolution kernel by looking at all of the possible 3x3 spatial slices is somewhat like trying to guess what an full image looks like from being shown all the 3x3 chunks of it in random order.</p> <p>Despite the difficulties I think good kernel visualizations are a worthwile pursuit. Good visualization techniques can be powerful diagnostics and the better the visualizations of our models the more powerful and robust we can make them. As a motivational carrot here is a teaser plot of a visualization of a simple network which we generate in this post.</p> <p><img src="http://asymptoticlabs.com/images/visualizing_convolution_kernels_teaser.png" alt="teaser_plot"></p> <p><a href="http://asymptoticlabs.com/posts/visualizing-convolution-kernels.html">Read more…</a> (19 min remaining to read)</p></div></div></div></body></html> kernelsneural networksvisualizationhttp://asymptoticlabs.com/posts/visualizing-convolution-kernels.htmlFri, 14 Sep 2018 06:00:00 GMThttp://asymptoticlabs.com/posts/tips-for-visualizing-correlation-matrices.htmlTim Anderton<div tabindex="-1" id="notebook" class="border-box-sizing"> <div class="container" id="notebook-container"> <div class="cell border-box-sizing text_cell rendered"><div class="prompt input_prompt"> </div> <div class="inner_cell"> <div class="text_cell_render border-box-sizing rendered_html"> <p>When dealing with data with dozens or hundreds of features one important tool is to look at the correlations between different features as a heat map. Although it is easy to generate a correlation heat map not all such visualizations are created equal. Here are some rules of thumb to keep in mind,</p> <ul> <li>Limit the range of the color map to the middle 99.x% of the values</li> <li>Use symmetric magnitude bounds</li> <li>Use a divergent color map</li> <li>Make 0 correlation correspond to a dull dark color (dark grey), and high magnitude correlations high luminance</li> <li>Different orderings of the features can have a huge impact, pick wisely.</li> </ul> <p>Using these guidelines together almost always improves the overall quality of the visualization of a correlation or covariance matrix.</p> <p>We will apply these guidelines one by one to an example data set (see below) talking about the motivation for each guideline as we apply it.</p> <p><a href="http://asymptoticlabs.com/posts/tips-for-visualizing-correlation-matrices.html">Read more…</a> (28 min remaining to read)</p></div></div></div></div></div></body></html> color mapscovariancevisualizationhttp://asymptoticlabs.com/posts/tips-for-visualizing-correlation-matrices.htmlThu, 12 Apr 2018 19:33:31 GMT