kwrobel

25 karma · joined February 22, 2015

I agree. Perplexities (probability of a text) can be compared using different tokenization after normalization.

kwrobel··on What I Learned from Working 32 Hours a Week

In some countries it is mandatory by law to have 14 days in one block per year.

kwrobel··on Deep learning beginner to expert: techniques from Fast.ai

"As Jeremy Howard points out, even academic papers often use softmax for multi-class classification, and I too have already seen it used incorrectly in blogs and papers during my short time studying DL."

AFAIK softmax should be used with mutli-class classification and sigmoid can be used with mutli-label classification.

kwrobel··on I had to learn Python to pay my taxes

AFAIK in Europe you need to pay taxes only when you convert (sell) crypto currency to real currency.

kwrobel··on High Expectations for Low Precision at Cern

Also in NLP domain a precision reduction may be applied: https://arxiv.org/abs/1706.06363

kwrobel··on Scientists have detected a major change to the Earth’s oceans

Read the paper and convince yourself.

kwrobel··on FastText – Library for fast text representation and classification

Is it multi label text classification or only multi class?

kwrobel··on Bag of Tricks for Efficient Text Classification

How a sequence of words is feeded to non-recurrent network?

kwrobel··on TensorFlow Simplified Interface

How it can be compared to keras?

kwrobel··on Google Knowledge Graph Search API

Talking about Cyc ontology, we are working on automatic Wikipedia articles classification: http://cycloped.io/