diff options
| -rw-r--r-- | assets/1423graph.jpg (renamed from assets/resources/1423graph.jpg) | bin | 50282 -> 50282 bytes | |||
| -rw-r--r-- | assets/dpmm.png (renamed from assets/resources/dpmm.png) | bin | 41295 -> 41295 bytes | |||
| -rw-r--r-- | assets/fully-bayesian-mm.png (renamed from assets/resources/fully-bayesian-mm.png) | bin | 37129 -> 37129 bytes | |||
| -rw-r--r-- | assets/hmm.png (renamed from assets/resources/hmm.png) | bin | 57074 -> 57074 bytes | |||
| -rw-r--r-- | assets/juggling.png (renamed from assets/resources/juggling.png) | bin | 4802462 -> 4802462 bytes | |||
| -rw-r--r-- | assets/lda.png (renamed from assets/resources/lda.png) | bin | 43373 -> 43373 bytes | |||
| -rw-r--r-- | assets/mixture-model.png (renamed from assets/resources/mixture-model.png) | bin | 18380 -> 18380 bytes | |||
| -rw-r--r-- | assets/plsa1.png (renamed from assets/resources/plsa1.png) | bin | 32191 -> 32191 bytes | |||
| -rw-r--r-- | assets/plsa2.png (renamed from assets/resources/plsa2.png) | bin | 22340 -> 22340 bytes | |||
| -rw-r--r-- | assets/qrsklatest.pdf (renamed from assets/resources/qrsklatest.pdf) | bin | 436474 -> 436474 bytes | |||
| -rw-r--r-- | assets/realdis.pdf (renamed from assets/resources/realdis.pdf) | bin | 698192 -> 698192 bytes | |||
| -rw-r--r-- | assets/sl-vs-ml.png (renamed from assets/resources/sl-vs-ml.png) | bin | 150277 -> 150277 bytes | |||
| -rw-r--r-- | assets/wangana.png (renamed from assets/resources/wangana.png) | bin | 1447628 -> 1447628 bytes | |||
| -rw-r--r-- | assets/wangana_screenshot_index.png (renamed from assets/resources/wangana_screenshot_index.png) | bin | 90981 -> 90981 bytes | |||
| -rw-r--r-- | posts/2014-04-01-q-robinson-schensted-symmetry-paper.md | 2 | ||||
| -rw-r--r-- | posts/2014-04-01-q-robinson-schensted-symmetry-paper.org | 2 | ||||
| -rw-r--r-- | posts/2015-04-02-juggling-skill-tree.md | 2 | ||||
| -rw-r--r-- | posts/2016-10-13-q-robinson-schensted-knuth-polymer.md | 2 | ||||
| -rw-r--r-- | posts/2016-10-13-q-robinson-schensted-knuth-polymer.org | 2 | ||||
| -rw-r--r-- | posts/2019-02-14-raise-your-elbo.md | 14 | ||||
| -rw-r--r-- | posts/2019-02-14-raise-your-elbo.org | 14 |
21 files changed, 19 insertions, 19 deletions
diff --git a/assets/resources/1423graph.jpg b/assets/1423graph.jpg Binary files differindex 1a6e739..1a6e739 100644 --- a/assets/resources/1423graph.jpg +++ b/assets/1423graph.jpg diff --git a/assets/resources/dpmm.png b/assets/dpmm.png Binary files differindex 7d41b41..7d41b41 100644 --- a/assets/resources/dpmm.png +++ b/assets/dpmm.png diff --git a/assets/resources/fully-bayesian-mm.png b/assets/fully-bayesian-mm.png Binary files differindex 5ce98c8..5ce98c8 100644 --- a/assets/resources/fully-bayesian-mm.png +++ b/assets/fully-bayesian-mm.png diff --git a/assets/resources/hmm.png b/assets/hmm.png Binary files differindex da902cc..da902cc 100644 --- a/assets/resources/hmm.png +++ b/assets/hmm.png diff --git a/assets/resources/juggling.png b/assets/juggling.png Binary files differindex 18ec08b..18ec08b 100644 --- a/assets/resources/juggling.png +++ b/assets/juggling.png diff --git a/assets/resources/lda.png b/assets/lda.png Binary files differindex 3eef014..3eef014 100644 --- a/assets/resources/lda.png +++ b/assets/lda.png diff --git a/assets/resources/mixture-model.png b/assets/mixture-model.png Binary files differindex b0787a5..b0787a5 100644 --- a/assets/resources/mixture-model.png +++ b/assets/mixture-model.png diff --git a/assets/resources/plsa1.png b/assets/plsa1.png Binary files differindex 8fa81a6..8fa81a6 100644 --- a/assets/resources/plsa1.png +++ b/assets/plsa1.png diff --git a/assets/resources/plsa2.png b/assets/plsa2.png Binary files differindex 6b38829..6b38829 100644 --- a/assets/resources/plsa2.png +++ b/assets/plsa2.png diff --git a/assets/resources/qrsklatest.pdf b/assets/qrsklatest.pdf Binary files differindex 01a042b..01a042b 100644 --- a/assets/resources/qrsklatest.pdf +++ b/assets/qrsklatest.pdf diff --git a/assets/resources/realdis.pdf b/assets/realdis.pdf Binary files differindex 92d9302..92d9302 100644 --- a/assets/resources/realdis.pdf +++ b/assets/realdis.pdf diff --git a/assets/resources/sl-vs-ml.png b/assets/sl-vs-ml.png Binary files differindex 11ad6a8..11ad6a8 100644 --- a/assets/resources/sl-vs-ml.png +++ b/assets/sl-vs-ml.png diff --git a/assets/resources/wangana.png b/assets/wangana.png Binary files differindex b8457dc..b8457dc 100644 --- a/assets/resources/wangana.png +++ b/assets/wangana.png diff --git a/assets/resources/wangana_screenshot_index.png b/assets/wangana_screenshot_index.png Binary files differindex ecaa8ac..ecaa8ac 100644 --- a/assets/resources/wangana_screenshot_index.png +++ b/assets/wangana_screenshot_index.png diff --git a/posts/2014-04-01-q-robinson-schensted-symmetry-paper.md b/posts/2014-04-01-q-robinson-schensted-symmetry-paper.md index 38874bb..a0432c8 100644 --- a/posts/2014-04-01-q-robinson-schensted-symmetry-paper.md +++ b/posts/2014-04-01-q-robinson-schensted-symmetry-paper.md @@ -13,7 +13,7 @@ generalisation of the growth diagram approach introduced by Fomin. This approach, which uses "growth graphs", can also be applied to a wider class of insertion algorithms which have a branching structure. - + Above is the growth graph of the \\(q\\)-weighted Robinson-Schensted algorithm for the permutation \\({1 2 3 4\\choose1 4 2 3}\\). diff --git a/posts/2014-04-01-q-robinson-schensted-symmetry-paper.org b/posts/2014-04-01-q-robinson-schensted-symmetry-paper.org index b1c967d..7411a5d 100644 --- a/posts/2014-04-01-q-robinson-schensted-symmetry-paper.org +++ b/posts/2014-04-01-q-robinson-schensted-symmetry-paper.org @@ -10,7 +10,7 @@ approach, which uses "growth graphs", can also be applied to a wider class of insertion algorithms which have a branching structure. #+caption: Growth graph of q-RS for 1423 -[[../assets/resources/1423graph.jpg]] +[[../assets/1423graph.jpg]] Above is the growth graph of the \(q\)-weighted Robinson-Schensted algorithm for the permutation \({1 2 3 4\choose1 4 2 3}\). diff --git a/posts/2015-04-02-juggling-skill-tree.md b/posts/2015-04-02-juggling-skill-tree.md index 11eb377..7708988 100644 --- a/posts/2015-04-02-juggling-skill-tree.md +++ b/posts/2015-04-02-juggling-skill-tree.md @@ -30,4 +30,4 @@ have therefore written a script using [Python](http://python.org), difficulties, which is the leftmost column) from the Library of Juggling (click the image for the full size): -<a href="../assets/resources/juggling.png"><img src="../assets/resources/juggling.png" alt="The juggling skill tree" style="width:38em" /></a> +<a href="../assets/juggling.png"><img src="../assets/juggling.png" alt="The juggling skill tree" style="width:38em" /></a> diff --git a/posts/2016-10-13-q-robinson-schensted-knuth-polymer.md b/posts/2016-10-13-q-robinson-schensted-knuth-polymer.md index 4d31e37..4a5f793 100644 --- a/posts/2016-10-13-q-robinson-schensted-knuth-polymer.md +++ b/posts/2016-10-13-q-robinson-schensted-knuth-polymer.md @@ -12,7 +12,7 @@ This article is available at [arXiv](https://arxiv.org/abs/1610.03692). It seems to me that one difference between arXiv and Github is that on arXiv each preprint has a few versions only. In Github many projects have a "dev" branch hosting continuous updates, whereas the master branch is where the stable releases live. -[Here]({{ site.url }}/assets/resources/qrsklatest.pdf) is a "dev" version of the article, which I shall push to arXiv when it stablises. Below is the changelog. +[Here]({{ site.url }}/assets/qrsklatest.pdf) is a "dev" version of the article, which I shall push to arXiv when it stablises. Below is the changelog. * 2017-01-12: Typos and grammar, arXiv v2. * 2016-12-20: Added remarks on the geometric \\(q\\)-pushTASEP. Added remarks on the converse of the Burke property. Added natural language description of the \\(q\\)RSK. Fixed typos. diff --git a/posts/2016-10-13-q-robinson-schensted-knuth-polymer.org b/posts/2016-10-13-q-robinson-schensted-knuth-polymer.org index 93da639..206a4ab 100644 --- a/posts/2016-10-13-q-robinson-schensted-knuth-polymer.org +++ b/posts/2016-10-13-q-robinson-schensted-knuth-polymer.org @@ -28,7 +28,7 @@ few versions only. In Github many projects have a "dev" branch hosting continuous updates, whereas the master branch is where the stable releases live. -[[file:%7B%7B%20site.url%20%7D%7D/assets/resources/qrsklatest.pdf][Here]] +[[file:/assets/qrsklatest.pdf][Here]] is a "dev" version of the article, which I shall push to arXiv when it stablises. Below is the changelog. diff --git a/posts/2019-02-14-raise-your-elbo.md b/posts/2019-02-14-raise-your-elbo.md index 4080d0b..36bd364 100644 --- a/posts/2019-02-14-raise-your-elbo.md +++ b/posts/2019-02-14-raise-your-elbo.md @@ -138,7 +138,7 @@ $p(x_{1 : m}; \theta)$. Represented as a DAG (a.k.a the plate notations), the model looks like this: -{style="width:250px"} +{style="width:250px"} where the boxes with $m$ mean repitition for $m$ times, since there $m$ indepdent pairs of $(x, z)$, and the same goes for $\eta$. @@ -298,7 +298,7 @@ $$p(d_i = u, x_i = w | z_i = k; \theta) = p(d_i ; \xi_k) p(x_i; \eta_k) = \xi_{k The model can be illustrated in the plate notations: -{style="width:350px"} +{style="width:350px"} So the solution of the M-step is @@ -365,7 +365,7 @@ pLSA1, $(x | z = k) \sim \text{Cat}(\eta_{k, \cdot})$. Illustrated in the plate notations, pLSA2 is: -{style="width:350px"} +{style="width:350px"} The computation is basically adding an index $\ell$ to the computation of SMM wherever applicable. @@ -411,7 +411,7 @@ $$p(z_{i1}) = \pi_{z_{i1}}$$ So the parameters are $\theta = (\pi, \xi, \eta)$. And HMM can be shown in plate notations as: -{style="width:350px"} +{style="width:350px"} Now we apply EM to HMM, which is called the [Baum-Welch algorithm](https://en.wikipedia.org/wiki/Baum%E2%80%93Welch_algorithm). @@ -569,7 +569,7 @@ later in this section that the posterior $q(\eta_k)$ belongs to the same family as $p(\eta_k)$. Represented in a plate notations, a fully Bayesian mixture model looks like: -{style="width:450px"} +{style="width:450px"} Given this structure we can write down the mean-field approximation: @@ -672,7 +672,7 @@ As the second example of fully Bayesian mixture models, Latent Dirichlet allocation (LDA) (Blei-Ng-Jordan 2003) is the fully Bayesian version of pLSA2, with the following plate notations: -{style="width:450px"} +{style="width:450px"} It is the smoothed version in the paper. @@ -782,7 +782,7 @@ $$L(p, q) = \sum_{k = 1 : T} \mathbb E_{q(\theta_k)} \log {p(\theta_k) \over q(\ The plate notation of this model is: -{style="width:450px"} +{style="width:450px"} As it turns out, the infinities can be tamed in this case. diff --git a/posts/2019-02-14-raise-your-elbo.org b/posts/2019-02-14-raise-your-elbo.org index f0de7d1..9cc79a7 100644 --- a/posts/2019-02-14-raise-your-elbo.org +++ b/posts/2019-02-14-raise-your-elbo.org @@ -137,7 +137,7 @@ $p(x_{1 : m}; \theta)$. Represented as a DAG (a.k.a the plate notations), the model looks like this: -[[/assets/resources/mixture-model.png]] +[[/assets/mixture-model.png]] where the boxes with $m$ mean repitition for $m$ times, since there $m$ indepdent pairs of $(x, z)$, and the same goes for $\eta$. @@ -309,7 +309,7 @@ $$p(d_i = u, x_i = w | z_i = k; \theta) = p(d_i ; \xi_k) p(x_i; \eta_k) = \xi_{k The model can be illustrated in the plate notations: -[[/assets/resources/plsa1.png]] +[[/assets/plsa1.png]] So the solution of the M-step is @@ -380,7 +380,7 @@ pLSA1, $(x | z = k) \sim \text{Cat}(\eta_{k, \cdot})$. Illustrated in the plate notations, pLSA2 is: -[[/assets/resources/plsa2.png]] +[[/assets/plsa2.png]] The computation is basically adding an index $\ell$ to the computation of SMM wherever applicable. @@ -429,7 +429,7 @@ $$p(z_{i1}) = \pi_{z_{i1}}$$ So the parameters are $\theta = (\pi, \xi, \eta)$. And HMM can be shown in plate notations as: -[[/assets/resources/hmm.png]] +[[/assets/hmm.png]] Now we apply EM to HMM, which is called the [[https://en.wikipedia.org/wiki/Baum%E2%80%93Welch_algorithm][Baum-Welch @@ -592,7 +592,7 @@ later in this section that the posterior $q(\eta_k)$ belongs to the same family as $p(\eta_k)$. Represented in a plate notations, a fully Bayesian mixture model looks like: -[[/assets/resources/fully-bayesian-mm.png]] +[[/assets/fully-bayesian-mm.png]] Given this structure we can write down the mean-field approximation: @@ -701,7 +701,7 @@ As the second example of fully Bayesian mixture models, Latent Dirichlet allocation (LDA) (Blei-Ng-Jordan 2003) is the fully Bayesian version of pLSA2, with the following plate notations: -[[/assets/resources/lda.png]] +[[/assets/lda.png]] It is the smoothed version in the paper. @@ -813,7 +813,7 @@ $$L(p, q) = \sum_{k = 1 : T} \mathbb E_{q(\theta_k)} \log {p(\theta_k) \over q(\ The plate notation of this model is: -[[/assets/resources/dpmm.png]] +[[/assets/dpmm.png]] As it turns out, the infinities can be tamed in this case. |
