Fabien Benetou's PIM | Tools / MachineLearning

From personal use

~~deeplearn.js~~ tensorflow.js discovery over several VRLab Brussels meetups
- could use emoji-scavenger-hunt 400 classes object recognition for Overlay
running TensorFlow Udacity tutorial on AWS
- AWS
- cf console https://eu-central-1.console.aws.amazon.com/ec2/v2/home?region=eu-central-1#Instances:sort=instanceId with docker
  - running on http://52.29.3.107:8888/notebooks/1_notmnist.ipynb there
    - no persistent EBS volume so all data lost, notebook saved locally though cf ~/Prototypes/AWS/TensorFlowUdacityMooc/
- docker https://github.com/tensorflow/tensorflow/tree/master/tensorflow/examples/udacity
- exploring http://playground.tensorflow.org with ReLU cf Wikipedia Rectifier neural networks and https://www.tensorflow.org while loading datasets
- done during http://www.meetup.com/MLGeek-Brussels/events/229482789/ while most people ran it on their own machines
- see also http://ufldl.stanford.edu/tutorial/supervised/ConvolutionalNeuralNetwork/
  - recommending instead http://cs231n.stanford.edu/syllabus.html
PIM page categorization
- cf several proposal on how to keep the PIM plastic yet scalable
- defining features on each page
  - incoming links
  - potential internal links (anchors)
  - titles (e.g. /^!/)
  - bags of words
AuctionPricing
- importance of feature selection

To explore

Note

My notes on Tools gather what I know or want to know. Consequently they are not and will never be complete references. For this, official manuals and online communities provide much better answers.

Machine Learning {Tools}

From personal use

See also

To explore

Note