kkjkok.blogspot.com

ift6266h15.wordpress.com

Project Blogs and Repos – IFT6266 – H2015 Representation Learning

https://ift6266h15.wordpress.com/project-blogs-and-repos

IFT6266 – H2015 Representation Learning. A mostly deep learning course. Project Blogs and Repos. Project Blogs and Repos. Https:/ gyomift6266h15.wordpress.com/. GIT: https:/ github.com/gyom/ift6266h15. Blog: http:/ adbrebs.wordpress.com. GIT: https:/ github.com/adbrebs/dogs vs cats. Blog: https:/ cesarlaurent.wordpress.com/. GIT: https:/ github.com/Thrandis/ift6266h15. Blog: https:/ jychung.wordpress.com/. GIT: https:/ github.com/jych/ift6266h15. GIT: https:/ github.com/cdansereau/ift6266. José Manuel Ro...

kastnerkyle.github.io

Archive | Garbage In Garbage Out

http://kastnerkyle.github.io/archive.html

Skip to main content. Garbage In Garbage Out. Projects I Am Involved In. Contents 2016 Kyle Kastner.

jychung.wordpress.com

Junyoung | IFT6266

https://jychung.wordpress.com/author/jychung

My trial and errors. Model configuration and classification and generation results. April 7, 2015. April 20, 2015. I used Adam [1] as an optimizer, and the initial learning rate was set to 0.0005. Above figure shows how the model generate cat images (if you click it changes into full size). At the below is the figure showing the write operation of dog images. The model always start writing from the top-left corner of the canvas and asymptotically fills out the remaining region of the canvas. The results ...

jychung.wordpress.com

First stage complete | IFT6266

https://jychung.wordpress.com/2015/04/02/first-stage-achieved

My trial and errors. April 2, 2015. April 19, 2015. I have just finished doing experiments with my baseline. It is a very naive CNN model with 3 hidden and pooling layers and 2 fully connected layers with dropout. I’m using binary cross entropy to train my model. I used my personal research library so called clé (you can find the library here. And the script file of the model in here. It takes 1.5 2 mins to achieve 80% accuracy using Titan Black and cuDNN v2. Here is the pipeline with more details. You a...

kastnerkyle.github.io

Wavelets | Garbage In Garbage Out

http://kastnerkyle.github.io/posts/wavelets

Skip to main content. Garbage In Garbage Out. Projects I Am Involved In. Wavelets are a fundamental part of modern signal processing and feature engineering. Utilizing well developed basis functions with certain mathematical properties, rather than the more typical sines and cosines used for the DFT (discrete fourier transform) and DCT (discrete cosine transform), wavelet analysis has many interesting applications. Wavelets have been used extensively for denoising. Sunspots.csv already downloaded! Functi...

kastnerkyle.github.io

Single Speaker Word Recognition With Hidden Markov Models | Garbage In Garbage Out

http://kastnerkyle.github.io/posts/single-speaker-word-recognition-with-hidden-markov-models

Skip to main content. Garbage In Garbage Out. Projects I Am Involved In. Single Speaker Word Recognition With Hidden Markov Models. Hidden Markov Models (HMMs) are powerful, flexible methods for representing and classifying data with trends over time, and have been a key component in speech recognition systems for many years. Which holds the HMM implementations that were formerly a part of sklearn. Other available datasets are largely multispeaker, but the simple frequency peak features used in this exam...

kastnerkyle.github.io

Robust Matrix Decomposition | Garbage In Garbage Out

http://kastnerkyle.github.io/posts/robust-matrix-decomposition

Skip to main content. Garbage In Garbage Out. Projects I Am Involved In. There has been a heavy focus on random projections in recent algorithms, which can often lead to increased stability and computationally efficient solutions. Below is a link to the GoDec algorithm output, as applied to the "Hall" video (shown below) found in this zip file. And we will set up a simple download function for ease-of-access. Special thanks to @kuantkid. PRETTY. DANG. AWESOME. Download files from http:/ perception.i2...

kastnerkyle.github.io

Bad Speech Synthesis Made Simple | Garbage In Garbage Out

http://kastnerkyle.github.io/posts/bad-speech-synthesis-made-simple

Skip to main content. Garbage In Garbage Out. Projects I Am Involved In. Bad Speech Synthesis Made Simple. Speech synthesis is a core component of many modern products - turn by turn navigation, automatic personal assistants/search like Siri and Alexa, and even the voices of AI/robots/people in computer games and movies are (in some cases) generated by speech synthesis systems. But how are these systems built? How can we make computers reply to our requests (and demands) using speech? This seemed to be t...

kastnerkyle.github.io

Using PyTables for Larger-Than-RAM Data Processing | Garbage In Garbage Out

http://kastnerkyle.github.io/posts/using-pytables-for-larger-than-ram-data-processing

Skip to main content. Garbage In Garbage Out. Projects I Am Involved In. Using PyTables for Larger-Than-RAM Data Processing. HDF5 is a direct, easy path to "big" (or just annoyingly larger than RAM) data in scientific python. Using the HDF5 file format and careful coding of your algorithms, it is quite possible to process "big-ish" data on modest hardware, and really push your resources to the limit before moving to larger, fancier distributed platforms. Make random cluster points. PyTables is a package ...