ml-recall

#long-range-dependencies

3 questions from Recurrent Neural Networks, Attention & Transformers, Language Modeling

Timed test

3 questions · 3 min · score and explanations at the end

Take a test

Q 1 / 3Score 0 / 0

Shuffle questions

Question 1: Short answer

mediumAI-checkedRecurrent Neural Networks

A linear RNN with a scalar state, $h_t = w_h h_{t-1} + w_x x_t$ , starts from $h_0 = 0$ , with $w_h = 0.9$ and $w_x = 1$ . After $T = 50$ steps the state is a weighted sum $h_{50} = \sum_{t=1}^{50} c_t x_t$ . What is the weight $c_1$ of the first input?

Enter: check / next