Mercurial > pylearn
annotate doc/v2_planning.txt @ 979:2a53384d9742
mcRBM - hacks to driver
| author | James Bergstra <bergstrj@iro.umontreal.ca> |
|---|---|
| date | Mon, 23 Aug 2010 16:05:31 -0400 |
| parents | 216f4ce969b2 |
| children |
| rev | line source |
|---|---|
| 941 | 1 |
| 2 Motivation | |
| 3 ========== | |
| 4 | |
| 5 Yoshua: | |
| 6 ------- | |
| 7 | |
| 8 We are missing a *Theano Machine Learning library*. | |
| 9 | |
| 10 The deep learning tutorials do a good job but they lack the following features, which I would like to see in a ML library: | |
| 11 | |
| 12 - a well-organized collection of Theano symbolic expressions (formulas) for handling most of | |
| 13 what is needed either in implementing existing well-known ML and deep learning algorithms or | |
| 14 for creating new variants (without having to start from scratch each time), that is the | |
| 15 mathematical core, | |
| 16 | |
| 17 - a well-organized collection of python modules to help with the following: | |
| 18 - several data-access models that wrap around learning algorithms for interfacing with various types of data (static vectors, images, sound, video, generic time-series, etc.) | |
| 19 - generic utility code for optimization | |
| 20 - stochastic gradient descent variants | |
| 21 - early stopping variants | |
| 22 - interfacing to generic 2nd order optimization methods | |
| 23 - 2nd order methods tailored to work on minibatches | |
| 24 - optimizers for sparse coefficients / parameters | |
| 25 - generic code for model selection and hyper-parameter optimization (including the use and coordination of multiple jobs running on different machines, e.g. using jobman) | |
| 26 - generic code for performance estimation and experimental statistics | |
| 27 - visualization tools (using existing python libraries) and examples for all of the above | |
| 28 - learning algorithm conventions and meta-learning algorithms (bagging, boosting, mixtures of experts, etc.) which use them | |
| 29 | |
| 30 [Note that many of us already use some instance of all the above, but each one tends to reinvent the wheel and newbies don't benefit from a knowledge base.] | |
| 31 | |
| 32 - a well-documented set of python scripts using the above library to show how to run the most | |
| 33 common ML algorithms (possibly with examples showing how to run multiple experiments with | |
| 34 many different models and collect statistical comparative results). This is particularly | |
| 35 important for pure users to adopt Theano in the ML application work. | |
| 36 | |
| 37 Ideally, there would be one person in charge of this project, making sure a coherent and | |
| 38 easy-to-read design is developed, along with many helping hands (to implement the various | |
| 39 helper modules, formulae, and learning algorithms). | |
| 40 | |
| 41 | |
| 42 James: | |
| 43 ------- | |
| 44 | |
| 45 I am interested in the design and implementation of the "well-organized collection of Theano | |
| 46 symbolic expressions..." | |
| 47 | |
| 48 I would like to explore algorithms for hyper-parameter optimization, following up on some | |
| 49 "high-throughput" work. I'm most interested in the "generic code for model selection and | |
| 50 hyper-parameter optimization..." and "generic code for performance estimation...". | |
| 51 | |
| 52 I have some experiences with the data-access requirements, and some lessons I'd like to share | |
| 53 on that, but no time to work on that aspect of things. | |
| 54 | |
| 55 I will continue to contribute to the "well-documented set of python scripts using the above to | |
| 56 showcase common ML algorithms...". I have an Olshausen&Field-style sparse coding script that | |
| 57 could be polished up. I am also implementing the mcRBM and I'll be able to add that when it's | |
| 58 done. | |
| 59 | |
| 60 | |
| 61 | |
| 62 Suggestions for how to tackle various desiderata | |
| 63 ================================================ | |
| 64 | |
| 65 | |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
66 Theano Symbolic Expressions for ML |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
67 ---------------------------------- |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
68 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
69 We could make this a submodule of pylearn: ``pylearn.nnet``. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
70 |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
71 Yoshua: I would use a different name, e.g., "pylearn.formulas" to emphasize that it is not just |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
72 about neural nets, and that this is a collection of formulas (expressions), rather than |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
73 completely self-contained classes for learners. We could have a "nnet.py" file for |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
74 neural nets, though. |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
75 |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
76 There are a number of ideas floating around for how to handle classes / |
| 947 | 77 modules (LeDeepNet, pylearn.shared.layers, pynnet, DeepAnn) so lets implement as much |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
78 math as possible in global functions with no classes. There are no models in |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
79 the wish list that require than a few vectors and matrices to parametrize. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
80 Global functions are more reusable than classes. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
81 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
82 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
83 Data access |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
84 ----------- |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
85 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
86 A general interface to datasets from the perspective of an experiment driver |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
87 (e.g. kfold) is to see them as a function that maps index (typically integer) |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
88 to example (whose type and nature depends on the dataset, it could for |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
89 instance be an (image, label) pair). This interface permits iterating over |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
90 the dataset, shuffling the dataset, and splitting it into folds. For |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
91 efficiency, it is nice if the dataset interface supports looking up several |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
92 index values at once, because looking up many examples at once can sometimes |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
93 be faster than looking each one up in turn. In particular, looking up |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
94 a consecutive block of indices, or a slice, should be well supported. |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
95 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
96 Some datasets may not support random access (e.g. a random number stream) and |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
97 that's fine if an exception is raised. The user will see a NotImplementedError |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
98 or similar, and try something else. We might want to have a way to test |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
99 that a dataset is random-access or not without having to load an example. |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
100 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
101 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
102 A more intuitive interface for many datasets (or subsets) is to load them as |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
103 matrices or lists of examples. This format is more convenient to work with at |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
104 an ipython shell, for example. It is not good to provide only the "dataset |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
105 as a function" view of a dataset. Even if a dataset is very large, it is nice |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
106 to have a standard way to get some representative examples in a convenient |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
107 structure, to be able to play with them in ipython. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
108 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
109 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
110 Another thing to consider related to datasets is that there are a number of |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
111 other efforts to have standard ML datasets, and we should be aware of them, |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
112 and compatible with them when it's easy: |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
113 - mldata.org (they have a file format, not sure how many use it) |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
114 - weka (ARFF file format) |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
115 - scikits.learn |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
116 - hdf5 / pytables |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
117 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
118 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
119 pylearn.datasets uses a DATA_ROOT environment variable to locate a filesystem |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
120 folder that is assumed to have a standard form across different installations. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
121 That's where the data files are. The correct format of this folder is currently |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
122 defined implicitly by the contents of /data/lisa/data at DIRO, but it would be |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
123 better to document in pylearn what the contents of this folder should be as |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
124 much as possible. It should be possible to rebuild this tree from information |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
125 found in pylearn. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
126 |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
127 Yoshua (about ideas proposed by Pascal Vincent a while ago): |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
128 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
129 - we may want to distinguish between datasets and tasks: a task defines |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
130 not just the data but also things like what is the input and what is the |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
131 target (for supervised learning), and *importantly* a set of performance metrics |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
132 that make sense for this task (e.g. those used by papers solving a particular |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
133 task, or reported for a particular benchmark) |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
134 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
135 - we should discuss about a few "standards" that datasets and tasks may comply to, such as |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
136 - "input" and "target" fields inside each example, for supervised or semi-supervised learning tasks |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
137 (with a convention for the semi-supervised case when only the input or only the target is observed) |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
138 - "input" for unsupervised learning |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
139 - conventions for missing-valued components inside input or target |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
140 - how examples that are sequences are treated (e.g. the input or the target is a sequence) |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
141 - how time-stamps are specified when appropriate (e.g., the sequences are asynchronous) |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
142 - how error metrics are specified |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
143 * example-level statistics (e.g. classification error) |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
144 * dataset-level statistics (e.g. ROC curve, mean and standard error of error) |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
145 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
146 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
147 Model Selection & Hyper-Parameter Optimization |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
148 ---------------------------------------------- |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
149 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
150 Driving a distributed computing job for a long time to optimize |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
151 hyper-parameters using one or more clusters is the goal here. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
152 Although there might be some library-type code to write here, I think of this |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
153 more as an application template. The user would use python code to describe |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
154 the experiment to run and the hyper-parameter space to search. Then this |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
155 application-driver would take control of scheduling jobs and running them on |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
156 various computers... I'm imagining a potentially ugly brute of a hack that's |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
157 not necessarily something we will want to expose at a low-level for reuse. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
158 |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
159 Yoshua: We want both the library-defined driver that takes instructions about how to generate |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
160 new hyper-parameter combinations (e.g. implicitly providing a prior distribution from which |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
161 to sample them), and examples showing how to use it in typical cases. |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
162 Note that sometimes we just want to find the best configuration of hyper-parameters, |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
163 but sometimes we want to do more subtle analysis. Often a combination of both. |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
164 In this respect it could be useful for the user to define hyper-parameters over |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
165 which scientific questions are sought (e.g. depth of an architecture) vs |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
166 hyper-parameters that we would like to marginalize/maximize over (e.g. learning rate). |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
167 This can influence both the sampling of configurations (we want to make sure that all |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
168 combinations of question-driving hyper-parameters are covered) and the analysis |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
169 of results (we may be willing to estimate ANOVAs or averaging or quantiles over |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
170 the non-question-driving hyper-parameters). |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
171 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
172 Python scripts for common ML algorithms |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
173 --------------------------------------- |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
174 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
175 The script aspect of this feature request makes me think that what would be |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
176 good here is more tutorial-type scripts. And the existing tutorials could |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
177 potentially be rewritten to use some of the pylearn.nnet expressions. More |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
178 tutorials / demos would be great. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
179 |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
180 Yoshua: agreed that we could write them as tutorials, but note how the |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
181 spirit would be different from the current deep learning tutorials: we would |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
182 not mind using library code as much as possible instead of trying to flatten |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
183 out everything in the interest of pedagogical simplicity. Instead, these |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
184 tutorials should be meant to illustrate not the algorithms but *how to take |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
185 advantage of the library*. They could also be used as *BLACK BOX* implementations |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
186 by people who don't want to dig lower and just want to run experiments. |
| 941 | 187 |
| 188 Functional Specifications | |
| 189 ========================= | |
| 190 | |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
191 TODO: |
| 941 | 192 Put these into different text files so that this one does not become a monster. |
| 193 For each thing with a functional spec (e.g. datasets library, optimization library) make a | |
| 194 separate file. | |
| 195 | |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
196 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
197 |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
198 pylearn.formulas |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
199 ---------------- |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
200 |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
201 Directory with functions for building layers, calculating classification |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
202 errors, cross-entropies with various distributions, free energies, etc. This |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
203 module would include for the most part global functions, Theano Ops and Theano |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
204 optimizations. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
205 |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
206 Yoshua: I would break it down in module files, e.g.: |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
207 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
208 pylearn.formulas.costs: generic / common cost functions, e.g. various cross-entropies, squared error, |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
209 abs. error, various sparsity penalties (L1, Student) |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
210 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
211 pylearn.formulas.linear: formulas for linear classifier, linear regression, factor analysis, PCA |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
212 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
213 pylearn.formulas.nnet: formulas for building layers of various kinds, various activation functions, |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
214 layers which could be plugged with various costs & penalties, and stacked |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
215 |
| 947 | 216 pylearn.formulas.ae: formulas for auto-encoders and denoising auto-encoder variants |
| 217 | |
| 218 pylearn.formulas.noise: formulas for corruption processes | |
|
946
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
219 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
220 pylearn.formulas.rbm: energies, free energies, conditional distributions, Gibbs sampling |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
221 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
222 pylearn.formulas.trees: formulas for decision trees |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
223 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
224 pylearn.formulas.boosting: formulas for boosting variants |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
225 |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
226 etc. |
|
7c4504a4ce1a
additions to formulas, data access, hyper-params, scripts
Yoshua Bengio <bengioy@iro.umontreal.ca>
parents:
945
diff
changeset
|
227 |
| 947 | 228 Fred: It seam that the DeepANN git repository by Xavier G. have part of this as function. |
| 229 | |
|
945
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
230 Indexing Convention |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
231 ~~~~~~~~~~~~~~~~~~~ |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
232 |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
233 Something to decide on - Fortran-style or C-style indexing. Although we have |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
234 often used c-style indexing in the past (for efficiency in c!) this is no |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
235 longer an issue with numpy because the physical layout is independent of the |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
236 indexing order. The fact remains that Fortran-style indexing follows linear |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
237 algebra conventions, while c-style indexing does not. If a global function |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
238 includes a lot of math derivations, it would be *really* nice if the code used |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
239 the same convention for the orientation of matrices, and endlessly annoying to |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
240 have to be always transposing everything. |
|
cafa16bfc7df
additions to v2_planning
James Bergstra <bergstrj@iro.umontreal.ca>
parents:
941
diff
changeset
|
241 |