目录

Spatial signals

Temporal signals

# Sequence

Sequence embedding

• [b,seq_len,feature_len] # 1个句子10个单词每个单词4个意思，[1,10,4]

• e.g. I like it.

import tensorflow as tftf.convert_to_tensor([[1,0,0],[0,1,0],[0,0,1]])

[b,100,1]

• [price,scalar,1] # 1-->1个点表示一个价格

[b,28,28]

• 对图片扫描的次数变成了时间的概念

Batch

• [b,word num,word vec]

• [word num,b,word vec]

[words,word vec]

• How to represent a word
  • [Rome, Italy, ...]
• one hot

• sparse

• high-dim

• semantic similarity

• trainable

Word embedding

• Word2Vec vs GloVe

Embedding Layer

• Random initialized embedding

from tensorflow.keras import layersx = tf.range(5)x = tf.random.shuffle(x)x

net = layers.Embedding(10,4)net(x)

net.trainable

True

net.trainable_variables

[
    
     ]