wassname

"""
A keras attention layer that wraps RNN layers.

Based on tensorflows [attention_decoder](https://github.com/tensorflow/tensorflow/blob/c8a45a8e236776bed1d14fd71f3b6755bd63cc58/tensorflow/python/ops/seq2seq.py#L506) 
and [Grammar as a Foreign Language](https://arxiv.org/abs/1412.7449).

date: 20161101
author: wassname
url: https://gist.github.com/wassname/5292f95000e409e239b9dc973295327a
"""

pannous

"""Sequence-to-sequence model with an attention mechanism."""
# see https://www.tensorflow.org/versions/r0.10/tutorials/seq2seq/index.html
# compare https://github.com/tflearn/tflearn/blob/master/examples/nlp/seq2seq_example.py
from __future__ import print_function
import numpy as np
import tensorflow as tf

vocab_size=256 # We are lazy, so we avoid fency mapping and just use one *class* per character/byte
target_vocab_size=vocab_size
learning_rate=0.1

gerbrand

Spreading activation in Neo4j

Introduction

How does the human brain work? How do thoughts pop up in our conscience? Just thinking about the question is confusion enough. Maybe our brain works as a semantic network as explained by Carole Yue at Kahn Academy.
Try for yourself: think of a bird? You probably think a sparrow, duck. Think of an extinct animal? Now you might think of the dodo, mammoth or dinosaur. Animal living on the south pole: a penguin? Or polar bear?

Facts, ideas, memories in our brain are interconnected. They’re not structured into tables like in relational databases or hierarchical in directories on a harddrive: if I’d ask you to name all the birds you know, you won’t give me a long list immediately. A better model of human m