Mercurial > ift6266

comparison writeup/nipswp_submission.tex @ 598:a0fdc1f134da

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
minor changes to nips workshop submission
author Yoshua Bengio <bengioy@iro.umontreal.ca>
date Thu, 14 Oct 2010 22:09:55 -0400
parents 5ab605c9a7d9
children
comparison
equal deleted inserted replaced
597:5ab605c9a7d9 598:a0fdc1f134da
172 which is based on training with or without these transformed images and testing on 172 which is based on training with or without these transformed images and testing on
173 clean ones. We measure the relative advantage of out-of-distribution examples 173 clean ones. We measure the relative advantage of out-of-distribution examples
174 (perturbed or out-of-class) 174 (perturbed or out-of-class)
175 for a deep learner vs a supervised shallow one. 175 for a deep learner vs a supervised shallow one.
176 Code for generating these transformations as well as for the deep learning 176 Code for generating these transformations as well as for the deep learning
177 algorithms are made available at {\tt http://hg.assembla.com/ift6266}. 177 algorithms are made available at {\tt http://anonymous.url.net}.%{\tt http://hg.assembla.com/ift6266}.
178 We also estimate the relative advantage for deep learners of training with 178 We also estimate the relative advantage for deep learners of training with
179 other classes than those of interest, by comparing learners trained with 179 other classes than those of interest, by comparing learners trained with
180 62 classes with learners trained with only a subset (on which they 180 62 classes with learners trained with only a subset (on which they
181 are then tested). 181 are then tested).
182 The conclusion discusses 182 The conclusion discusses
225 \subfigure[Gaussian Noise]{\includegraphics[scale=0.6]{images/Distorsiongauss_only.png}} 225 \subfigure[Gaussian Noise]{\includegraphics[scale=0.6]{images/Distorsiongauss_only.png}}
226 \subfigure[Background Image Addition]{\includegraphics[scale=0.6]{images/background_other_only.png}} 226 \subfigure[Background Image Addition]{\includegraphics[scale=0.6]{images/background_other_only.png}}
227 \subfigure[Salt \& Pepper]{\includegraphics[scale=0.6]{images/Poivresel_only.png}} 227 \subfigure[Salt \& Pepper]{\includegraphics[scale=0.6]{images/Poivresel_only.png}}
228 \subfigure[Scratches]{\includegraphics[scale=0.6]{images/Rature_only.png}} 228 \subfigure[Scratches]{\includegraphics[scale=0.6]{images/Rature_only.png}}
229 \subfigure[Grey Level \& Contrast]{\includegraphics[scale=0.6]{images/Contrast_only.png}} 229 \subfigure[Grey Level \& Contrast]{\includegraphics[scale=0.6]{images/Contrast_only.png}}
230 \caption{Transformation modules} 230 \caption{Top left (a): example original image. Others (b-o): examples of the effect
231 of each transformation module taken separately. Actual perturbed examples are obtained by
232 a pipeline of these, with random choices about which module to apply and how much perturbation
233 to apply.}
231 \label{fig:transform} 234 \label{fig:transform}
232 \vspace*{-2mm} 235 \vspace*{-2mm}
233 \end{figure} 236 \end{figure}
234 237
235 \vspace*{-3mm} 238 \vspace*{-3mm}
245 248
246 The first step in constructing the larger datasets (called NISTP and P07) is to sample from 249 The first step in constructing the larger datasets (called NISTP and P07) is to sample from
247 a {\em data source}: {\bf NIST} (NIST database 19), {\bf Fonts}, {\bf Captchas}, 250 a {\em data source}: {\bf NIST} (NIST database 19), {\bf Fonts}, {\bf Captchas},
248 and {\bf OCR data} (scanned machine printed characters). Once a character 251 and {\bf OCR data} (scanned machine printed characters). Once a character
249 is sampled from one of these sources (chosen randomly), the second step is to 252 is sampled from one of these sources (chosen randomly), the second step is to
250 apply a pipeline of transformations and/or noise processes described in section \ref{s:perturbations}. 253 apply a pipeline of transformations and/or noise processes outlined in section \ref{s:perturbations}.
251 254
252 To provide a baseline of error rate comparison we also estimate human performance 255 To provide a baseline of error rate comparison we also estimate human performance
253 on both the 62-class task and the 10-class digits task. 256 on both the 62-class task and the 10-class digits task.
254 We compare the best Multi-Layer Perceptrons (MLP) against 257 We compare the best Multi-Layer Perceptrons (MLP) against
255 the best Stacked Denoising Auto-encoders (SDA), when 258 the best Stacked Denoising Auto-encoders (SDA), when
256 both models' hyper-parameters are selected to minimize the validation set error. 259 both models' hyper-parameters are selected to minimize the validation set error.
257 We also provide a comparison against a precise estimate 260 We also provide a comparison against a precise estimate
258 of human performance obtained via Amazon's Mechanical Turk (AMT) 261 of human performance obtained via Amazon's Mechanical Turk (AMT)
259 service (http://mturk.com). 262 service ({\tt http://mturk.com}).
260 AMT users are paid small amounts 263 AMT users are paid small amounts
261 of money to perform tasks for which human intelligence is required. 264 of money to perform tasks for which human intelligence is required.
262 Mechanical Turk has been used extensively in natural language processing and vision. 265 Mechanical Turk has been used extensively in natural language processing and vision.
263 %processing \citep{SnowEtAl2008} and vision 266 %processing \citep{SnowEtAl2008} and vision
264 %\citep{SorokinAndForsyth2008,whitehill09}. 267 %\citep{SorokinAndForsyth2008,whitehill09}.