引导式测验 · 中等 等级
长短期记忆细胞测验
测试 LSTM 门角色、单元状态保留、输入和输出控制、附加更新、梯度流和序列限制。
问题 1 的 8
In a common modern three-gate LSTM variant, which gates regulate cell information and exposed output?
预览每个问题8 问题
当您参加测验时,就会出现答案和解释。
- In a common modern three-gate LSTM variant, which gates regulate cell information and exposed output?
- 读、写和执行门
- 启动、停止和暂停门
- 遗忘门、输入门和输出门
- 编码、解码和值守登机口
- What training difficulty was LSTM designed to mitigate?
- 需要存储的参数太多
- Vanishing gradients during learning over long sequences
- 无法使用 GPU
- 缺乏激活函数
- 谁在哪一年引入了 LSTM?
- Vaswani 和同事在 2017 年
- 鲁梅尔哈特和辛顿,1986 年
- LeCun 和 Bengio,2012 年
- Hochreiter 和 Schmidhuber,1997 年
- In a common LSTM cell update, why can the additive cell-state path help gradient flow?
- It reduces reliance on repeated multiplicative transformations across time steps
- 每一步都会将状态乘以一个大数
- 它每一步都会删除状态
- 它在训练期间冻结所有重量
- LSTM 门通常使用什么激活函数来充当软开关?
- ReLU,输出0到无穷大
- Softmax 对词汇表的影响
- 精确输出 0 或 1 的阶跃函数
- Sigmoid,输出0到1
- 为什么 Transformer 很大程度上取代了 LSTM 来进行大规模语言建模?
- Transformer 根本无法处理文本
- LSTM 不需要训练
- Transformer 跨序列并行化并通过注意力捕获远程上下文
- 变形金刚总共使用的数量要少得多
- If an LSTM forget gate is near one and its input gate near zero, what happens to the cell state?
- It largely retains the previous state
- It is forced to zero
- It becomes the raw input token
- It must reverse the sequence
- Which gate controls how much of an LSTM cell’s stored state is exposed as output?
- The forget gate alone
- The output gate
- The optimizer momentum term
- The positional encoding gate