{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"# Course Outline and Administrative Issues\n",
"\n",
"### Bayesian Machine Learning and Information Processing (5SSD0)"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"### Logistic Issues\n",
"\n",
"- **When**: 3rd quartile, at $8$ weeks of $4$ hours per week.\n",
"\n",
"- **Load**: Total workload is 5 ECTS $\\Rightarrow 5\\times 28 \\text{[hrs/ECTS]} = 140$ hours or $140/32 \\approx 4.4$ study hours per lecture.\n",
"\n",
"- **Web** \n",
" - home at [http://bmlip.nl](http://bmlip.nl) (or goto teaching tab at [http://biaslab.org](http://biaslab.org) ) \n",
" - source materials at github repo at [https://github.com/bertdv/BMLIP](https://github.com/bertdv/BMLIP)\n",
" - Please file a github issue if something is wrong or just unclear in these notes.\n",
" \n",
"- **Instructors**\n",
" - [Bert de Vries](http://bertdv.nl), rm. FLUX-7.101 (on Wednesdays), responsible for full course\n",
" - [Wouter Kouw](https://biaslab.github.io/member/wouter/), rm. FLUX-7.060, responsible for Probabilistic Programming mini-course\n",
" - Teaching assistents: [Magnus Koudahl](https://biaslab.github.io/member/magnus/) and [Ismail Senoz](https://biaslab.github.io/member/ismail/), rm. FLUX-7.060\n",
" - Please contact the TA's first for any questions regarding [Julia programming](http://julialang.org) examples and exercises"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"### Why Take This Class?\n",
"\n",
"- Suppose you need to develop an algorithm for a complex DSP task, e.g., a speech recognition engine. This is what you'll do:\n",
"\n",
" 1. Choose a set of candidate algorithms $y=H_k(x;\\theta)$ where $k \\in \\{1,2,\\ldots,K\\}$ and $\\theta \\in \\Theta_k$; (you think that) there's an algorithm $H_{k^*}(x;\\theta^*)$ that performs according to your liking. \n",
" 2. You collect a set of examples $D=\\{(x_1,y_1),(x_2,y_2),\\ldots,(x_N,y_N)\\}$ that are consistent with the correct algorithm behavior."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "subslide"
}
},
"source": [
" \n",
"- Using the methods from this class, you will be able to design a suitable algorithm through learning from the data set, thus achieving:\n",
" 1. **model selection**, i.e., find $k^*$ \n",
" 2. **parameter estimation**, i.e., find $\\theta^*$ "
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"- Better yet, we will discuss methods that find distributions $p(k|D)$ and $p(\\theta|D)$ that represent your knowledge about the best models and parameters, given the data set. "
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"### Materials\n",
"\n",
"- Book ([free download link](https://www.microsoft.com/en-us/research/people/cmbishop/#!prml-book)): \n",
"\n",
"- Mostly used for background reading as the (mandatory) slides are the main resource.\n",
"- Book theme: Whatever you do in machine learning, you can do it better with Bayesian methods.\n",
"- Contains much more material; great for future study and reference."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"### Exam Guide\n",
"- Tested material consists of these lecture notes, reading assignments (as assigned in the first cell/slide of each lecture notebook) and exercises (see class website)."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"- Advice: Make [Exercises](https://nbviewer.jupyter.org/github/bertdv/BMLIP/blob/master/lessons/notebooks/Exercises.ipynb), (to be) posted and regularly updated on the course website."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"- Advice: download (and make free use of) Sam Roweis' cheat sheets for [Matrix identities](https://github.com/bertdv/BMLIP/blob/master/lessons/notebooks/files/Roweis-1999-matrix-identities.pdf) and [Gaussian identities](https://github.com/bertdv/BMLIP/blob/master/lessons/notebooks/files/Roweis-1999-gaussian-identities.pdf)."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"- You are not allowed to use books nor bring printed or handwritten formula sheets to the exam. Difficult-to-remember formulas are supplied at the exam sheet (see old exams)."
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "fragment"
}
},
"source": [
"- You may use a simple pocket calculator, but no smartphones (only arithmetic assistance is allowed.)"
]
},
{
"cell_type": "markdown",
"metadata": {
"slideshow": {
"slide_type": "slide"
}
},
"source": [
"#