User Tools
latentartists
====== Differences ====== This shows you the differences between two versions of the page.
|
latentartists [2013/06/30 16:46] ben |
latentartists [2013/06/30 17:09] (current) ben |
||
|---|---|---|---|
| Line 7: | Line 7: | ||
| - | Assume a fixed vocabulary $V$, which in our experiments is a company internal list | + | Assume a fixed vocabulary $V$, which in our experiments is a list compiled by the Echonest |
| of music related multiword terms. | of music related multiword terms. | ||
| Line 52: | Line 52: | ||
| $KL(\mathcal{N}_0||\mathcal{N}_1) \propto (\mathbb{E}[z_0] - \mathbb{E}[z_1])\Sigma^{-1}(\mathbb{E}[z_0] - \mathbb{E}[z_1])^T + C$ | $KL(\mathcal{N}_0||\mathcal{N}_1) \propto (\mathbb{E}[z_0] - \mathbb{E}[z_1])\Sigma^{-1}(\mathbb{E}[z_0] - \mathbb{E}[z_1])^T + C$ | ||
| - | if the covariance matrix $\Sigma$ is the same for both Gaussians. This shows that if \Sigma^{-1} is a multiple | + | if the covariance matrix $\Sigma$ is the same for both Gaussians. This shows that if $\Sigma^{-1}$ is a multiple |
| of the identity matrix, the ranking retrieved will be the same as that of Euclidian distance between posterior means. | of the identity matrix, the ranking retrieved will be the same as that of Euclidian distance between posterior means. | ||
| Line 69: | Line 69: | ||
| during training. | during training. | ||
| - | {{::1000.ps|}} | + | == Hottt 1000 == |
| - | {{::full.ps|}} | + | {{::1000.jpg?600|}} |
| + | |||
| + | == Full Dataset == | ||
| + | |||
| + | {{::full.jpg?600|}} | ||
| The results do not support our hypothesis that taking uncertainty into account would create a more robust notion of similarity. | The results do not support our hypothesis that taking uncertainty into account would create a more robust notion of similarity. | ||
| - | While both methods effectively capture the artist similars in the Echonest database, the area under the ROC curve is clearly | + | While both methods clearly capture the information in the Echonest artist similar lists, the area under the ROC curve is clearly |
| greater for the simple Euclidean distance based approach. | greater for the simple Euclidean distance based approach. | ||
| - | + | The reason that the experimental results do not match our intuition is unclear. One possibility is that KL divergence | |
| - | + | is not an appropriate metric for similarity. | |
latentartists.1372625211.txt.gz · Last modified: 2013/06/30 16:46 by ben
Page Tools
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 3.0 Unported
