(self, s, a)

(Mi, Mo)

(V, g)

(self)

(self)

(p_y, t)

(self, df)

(filename, activation)

(P, g)

(shape)

(Mi, Mo)

(self, action)

(self, s)

(self)

(X, W1, b1, W2, b2)

(X, W, b)

(self)

(self, X)

(self)

(y)

(w1, w2, w3, concat=True, we_file='w2v_model.npz', w2i_file='w2v_word2idx.json')

(d)

(p_y, t)

(self, X, Y)

()

(shape)

(y)

()

(X, W)

(w1, w2, w3, we_file='word_embeddings.npy', w2i_file='wikipedia_word2idx.json')

(self)

(N=20, T=100, init=big_init)

(step_cost=-0.1)

(X, K, max_iter=20, beta=1.0, show_plots=True)

(self, observations)

(self, X, y)

(shape)

(totalrewards)

(self, X, Y)

()

(bytestream)

(self)

(limit=None)

()

(self, x, go_backwards=False)

(self, X)

(a, eps=0.1)

(self, X, Y)

(self, X, Y)

(self, X, Y)

(self, observations)

(Z, T, Y, W2)

(T, Y)

(X, Z, T, Y, W2)

(Z, T, Y)

(A)

(A)

(A)

(p, t)

(self, X, Y, learning_rate=1e-4, mu=0.99, epochs=30, show_fig=True, activation=T.nnet.relu, RecurrentUnit=GRU,

(self, X, reuse, is_training)

(self, X, is_training)

()

(t, y, X)

(t, y)

(Y, R)

(self, fn)

(self, X, Y)

(nbit)

( input_, gamma, beta, running_mean, running_var, is_training, axes='per-activation')

(X, W, b, poolsize=(2, 2))

(self, X, Y)

(self, X, Y)

(w1, w2, w3, we_file='word_embeddings.npy', w2i_file='wikipedia_word2idx.json')

(w1, w2, w3, We, word2idx)

(self, X, learning_rate=1e-5, mu=0.99, epochs=10, show_fig=True, activation=T.nnet.relu, RecurrentUnit=GRU, no

(self, df)

(self, X)

(self, X)

()

()

(n_vocab=2000, keep_words=KEEP_WORDS)

(n_files, n_vocab, by_paragraph=False)

(shape, poolsz)

(shape)

(shape)

(self, i, j)

(X, deg)

(self, x)

(self, X, Y)

(X, model)

(self, s, a)

(self, X)

(self, X)

(X)

(self, n)

(self, filename)

(self, X, Y)

(self, X, Y)