Compute the cross-entropy (log) loss. Notes ----- This method returns the sum (not the average!) of the losses for e

Create a 'random' unweighted directed acyclic graph by pruning all the backward connections from a random graph. Parameters --------

Set the layer parameters from a dictionary of values. Parameters ---------- summary_dict : dict A dictio

Set the model parameters and hyperparameters using the settings in `summary_dict`. Parameters ---------- sum

Remove punctuation from a string

A naive :math:`O(N^2)` implementation of the 1D discrete cosine transform-II (DCT-II). Notes ----- For a signal :math:`\mathbf{x

A naive :math:`O(N^2)` implementation of the 1D discrete Fourier transform (DFT). Notes ----- The Fourier transform decomposes a sig

A generic directed edge object. Parameters ---------- fr: int The id of the vertex the edge goes from

Run the agent for an episode. Parameters ---------- W : :py:class:`ndarray <numpy.ndarray>` of shape `(obs_dim, n_ac

Run or train the agent on an episode. Parameters ---------- max_steps : int The maximum number of steps

Run or train the agent on an episode. Parameters ---------- max_steps : int The maximum number of steps

r""" Computes the forward probability trellis for an HMM parameterized by :math:`(A, B, \pi)`. Notes ----- Th

Actual computation of softmax forward pass

Helper function that computes indices into X in prep for columnization in :func:`im2col`. Code extended from Andrej Karpathy's `im2col.p

Return the unsmoothed log probability of the ngram

r""" Compute the (unnormalized) log posterior for each class. Parameters ---------- X: :py:class:`ndarray <numpy.ndar

Compute the current MAP estimate for an arm's payoff probability

Compute the probability that the slope of the OLS fit to the loss history is negative. Parameters ----------

The logistic sigmoid function

Returns the maximum number of timesteps for which `P(loss is decreasing) < 0.51`. Parameters ---------- robu

Actual N-gram training logic

Perform the ALS update

Update the priority queue by calculating the priority for (s, a) and inserting it into the queue if it exceeds a fixed (small) thresh

r""" Execute the behavior policy--an :math:`\epsilon`-soft policy used to generate actions during training. Parameters

r""" Execute the behavior policy--an :math:`\epsilon`-soft policy used to generate actions during training. Parameters

Find all simple paths between `s_i` and `e_i` in the graph. Notes ----- Uses breadth-first search. Ignores all paths

Backprop from layer outputs to inputs. Parameters ---------- dLdy : :py:class:`ndarray <numpy.ndarray>` of shape `(n

r""" Calculate the model cross-entropy on a sequence of words against the empirical distribution of words in a sample. Notes

Calc the frequency bin centers for a DFT with `N` coefficients. Parameters ---------- N : int The number of frequency bins i

Dilate the 4D volume `X` by `d`. Notes ----- For a visual depiction of a dilated convolution, see [1]. References ---------

Compute statistics for the current environment. Parameters ---------- env : ``gym.wrappers`` or ``gym.envs`` instance The en

Compute the posterior over model parameters using the data in `X` and `y`. Parameters ---------- X : :py:cla

Compute the posterior over model parameters using the data in `X` and `y`. Parameters ---------- X : :py:cla

Create `n_trees`-worth of bootstrapped samples from the training data and use each to fit a separate decision tree.

Fit the gradient boosted decision trees on a dataset. Parameters ---------- X : :py:class:`ndarray <numpy.ndarray>`

Fit the parameters of the GMM on some training data. Parameters ---------- X : :py:class:`ndarray <numpy.ndarray>` o

r""" Evaulate the ReLU function on the elements of input `z`. .. math:: \text{ReLU}(z_i) &= z_i \ \ \ \

r""" Evaluate the softplus activation on the elements of input `z`. .. math:: \text{SoftPlus}(z_i) = \log(1 + e^{z_i})

Compute the layer output given input volume `X`. Parameters ---------- X : :py:class:`ndarray <numpy.ndarray>` of sh

Freeze the layer parameters at their current values so they can no longer be updated.

r""" Evaluate the first derivative of the logistic sigmoid on the elements of `x`. .. math:: \frac{\partial \sigma}{\par

r""" Evaulate the first derivative of the ReLU function on the elements of input `x`. .. math:: \frac{\partial \text{ReL

r""" Evaluate the first derivative of the tanh function on the elements of input `x`. .. math:: \frac{\partial \

r""" Evaluate the first derivative of the ELU activation on the elements of input `x`. .. math:: \frac{\partial

r""" Evaluate the first derivative of the softplus activation on the elements of input `x`. .. math:: \frac{\par

Computes the gradient of the generator or critic loss with regard to its inputs. Parameters ---------- Y_fak

Compute the gradient of the NCE loss with regard to the inputs, weights, and biases. Parameters ---------- r

Take a greedy action. Returns ------- total_reward : float The total reward on the episode. n_st

Convert the frequency representaiton of a signal in Hz into the mel scale. Parameters ---------- hz : :py:class:`ndarray <numpy.ndar

Initialize the episode logs. Notes ----- Training logs are represented as a nested set of dictionaries with the

Return True if array `x` consists only of binary values

True if `X` contains probabilities that sum to 1 along the columns

Check that a matrix `X` is a symmetric and positive-definite.

Compute the filters in a Mel filterbank and return the corresponding transformation matrix Notes ----- The Mel scale is a percep

Compute the minibatch indices for a training dataset. Parameters ---------- X : :py:class:`ndarray <numpy.ndarray>` of shape `(N, \*

Find the `k` nearest neighbors in the ball tree to a query vector `x` using the KNS1 algorithm. Parameters ---------

Return the probability under the environment model of each outcome state after taking action `a` in state `s`. Parameters

Use the trained model to generate prediction probabilities on a new collection of data points. Parameters ----------

Use the trained model to generate predictions on a new collection of data points. Parameters ---------- X :

Return the log probability of each data point in `X` under each mixture components. Parameters ---------- X

Add a new (key, value) pair with priority `priority` to the queue. Notes ----- If the queue is at capacity and `prio

Generate a random Bernoulli multi-armed bandit environemt

Generate a random multinomial multi-armed bandit environemt

Remove stop words from a list of word strings

Reverse the edge direction

Return all (reward, next_state) pairs associated with taking action `a` in state `s`.

Generate random draws from the `probs` distribution over integers in [0, N). Parameters ---------- n_samples

Return all (state, action) pairs in the environment model

Return a function to encode the continous observations generated by `env` in terms of a collection of `n_tilings` overlapping tilings (each w

Return a dictionary representation of the node's contents

Convert a 1D signal x into overlapping windows of width `frame_width` using a hop length of `stride`. Notes ----- If ``(len(x) -

Returns a (non-unique) topological sort / linearization of the nodes IFF the graph is acyclic, otherwise returns None. Notes

Compile the n-gram counts for the text(s) in `corpus_fp`. Upon completion the `self.counts` attribute will store dictionaries of the

Standardize features by removing the mean and scaling to unit variance. For a sample `x`, the standardized score is calculated as:

Generate draws from a truncated normal distribution via rejection sampling. Notes ----- The rejection sampling regimen draws samples

Unfreeze the layer parameters so they can be updated.

Perform gradient updates

Perform gradient updates

r""" Incrementally update the linear least-squares coefficients for a set of new examples. Notes ----- The re

Return the variational lower bound associated with the current model parameters.

r""" Run a single E-step update for the Baum-Welch/Forward-Backward algorithm. This step estimates ``xi`` and ``gamma``, the excepted

Maximize the VLB with respect to the variational parameters, γ and ϕ