Type WALSMatrixFactorization
Namespace tensorflow.contrib.factorization
Parent Estimator
Interfaces IWALSMatrixFactorization
An Estimator for Weighted Matrix Factorization, using the WALS method. WALS (Weighted Alternating Least Squares) is an algorithm for weighted matrix
factorization. It computes a lowrank approximation of a given sparse (n x m)
matrix `A`, by a product of two matrices, `U * V^T`, where `U` is a (n x k)
matrix and `V` is a (m x k) matrix. Here k is the rank of the approximation,
also called the embedding dimension. We refer to `U` as the row factors, and
`V` as the column factors.
See tensorflow/contrib/factorization/g3doc/wals.md for the precise problem
formulation. The training proceeds in sweeps: during a row_sweep, we fix `V` and solve for
`U`. During a column sweep, we fix `U` and solve for `V`. Each one of these
problems is an unconstrained quadratic minimization problem and can be solved
exactly (it can also be solved in minibatches, since the solution decouples
across rows of each matrix).
The alternating between sweeps is achieved by using a hook during training,
which is responsible for keeping track of the sweeps and running preparation
ops at the beginning of each sweep. It also updates the global_step variable,
which keeps track of the number of batches processed since the beginning of
training.
The current implementation assumes that the training is run on a single
machine, and will fail if `config.num_worker_replicas` is not equal to one.
Training is done by calling `self.fit(input_fn=input_fn)`, where `input_fn`
provides two tensors: one for rows of the input matrix, and one for rows of
the transposed input matrix (i.e. columns of the original matrix). Note that
during a row sweep, only row batches are processed (ignoring column batches)
and viceversa.
Also note that every row (respectively every column) of the input matrix
must be processed at least once for the sweep to be considered complete. In
particular, training will not make progress if some rows are not generated by
the `input_fn`. For prediction, given a new set of input rows `A'`, we compute a corresponding
set of row factors `U'`, such that `U' * V^T` is a good approximation of `A'`.
We call this operation a row projection. A similar operation is defined for
columns. Projection is done by calling
`self.get_projections(input_fn=input_fn)`, where `input_fn` satisfies the
constraints given below. The input functions must satisfy the following constraints: Calling `input_fn`
must return a tuple `(features, labels)` where `labels` is None, and
`features` is a dict containing the following keys: TRAIN:
* `WALSMatrixFactorization.INPUT_ROWS`: float32 SparseTensor (matrix).
Rows of the input matrix to process (or to project).
* `WALSMatrixFactorization.INPUT_COLS`: float32 SparseTensor (matrix).
Columns of the input matrix to process (or to project), transposed. INFER:
* `WALSMatrixFactorization.INPUT_ROWS`: float32 SparseTensor (matrix).
Rows to project.
* `WALSMatrixFactorization.INPUT_COLS`: float32 SparseTensor (matrix).
Columns to project.
* `WALSMatrixFactorization.PROJECT_ROW`: Boolean Tensor. Whether to project
the rows or columns.
* `WALSMatrixFactorization.PROJECTION_WEIGHTS` (Optional): float32 Tensor
(vector). The weights to use in the projection. EVAL:
* `WALSMatrixFactorization.INPUT_ROWS`: float32 SparseTensor (matrix).
Rows to project.
* `WALSMatrixFactorization.INPUT_COLS`: float32 SparseTensor (matrix).
Columns to project.
* `WALSMatrixFactorization.PROJECT_ROW`: Boolean Tensor. Whether to project
the rows or columns.
Methods
 get_col_factors
 get_col_factors_dyn
 get_projections
 get_projections_dyn
 get_row_factors
 get_row_factors_dyn
 NewDyn
Properties
 COMPLETED_SWEEPS_dyn
 config
 config_dyn
 INPUT_COLS_dyn
 INPUT_ROWS_dyn
 LOSS_dyn
 model_dir
 model_dir_dyn
 model_fn
 model_fn_dyn
 PROJECT_ROW_dyn
 PROJECTION_RESULT_dyn
 PROJECTION_WEIGHTS_dyn
 PythonObject
 RWSE_dyn
Fields
Public instance methods
IList<object> get_col_factors()
Returns the column factors of the model, loading them from checkpoint. Should only be run after training.
Returns

IList<object>
 A list of the column factors of the model.
object get_col_factors_dyn()
Returns the column factors of the model, loading them from checkpoint. Should only be run after training.
Returns

object
 A list of the column factors of the model.
IEnumerator<object> get_projections(PythonFunctionContainer input_fn)
Computes the projections of the rows or columns given in input_fn. Runs predict() with the given input_fn, and returns the results. Should only
be run after training.
Parameters

PythonFunctionContainer
input_fn  Input function which specifies the rows or columns to project.
Returns

IEnumerator<object>
 A generator of the projected factors.
object get_projections_dyn(object input_fn)
Computes the projections of the rows or columns given in input_fn. Runs predict() with the given input_fn, and returns the results. Should only
be run after training.
Parameters

object
input_fn  Input function which specifies the rows or columns to project.
Returns

object
 A generator of the projected factors.
IList<object> get_row_factors()
Returns the row factors of the model, loading them from checkpoint. Should only be run after training.
Returns

IList<object>
 A list of the row factors of the model.
object get_row_factors_dyn()
Returns the row factors of the model, loading them from checkpoint. Should only be run after training.
Returns

object
 A list of the row factors of the model.
Public static methods
WALSMatrixFactorization NewDyn(object num_rows, object num_cols, object embedding_dimension, ImplicitContainer<T> unobserved_weight, object regularization_coeff, ImplicitContainer<T> row_init, ImplicitContainer<T> col_init, ImplicitContainer<T> num_row_shards, ImplicitContainer<T> num_col_shards, ImplicitContainer<T> row_weights, ImplicitContainer<T> col_weights, ImplicitContainer<T> use_factors_weights_cache_for_training, ImplicitContainer<T> use_gramian_cache_for_training, object max_sweeps, object model_dir, object config)
Creates a model for matrix factorization using the WALS method.
Parameters

object
num_rows  Total number of rows for input matrix.

object
num_cols  Total number of cols for input matrix.

object
embedding_dimension  Dimension to use for the factors.

ImplicitContainer<T>
unobserved_weight  Weight of the unobserved entries of matrix.

object
regularization_coeff  Weight of the L2 regularization term. Defaults to None, in which case the problem is not regularized.

ImplicitContainer<T>
row_init  Initializer for row factor. Must be either:  A tensor: The row factor matrix is initialized to this tensor,  A numpy constant,  "random": The rows are initialized using a normal distribution.

ImplicitContainer<T>
col_init  Initializer for column factor. See row_init.

ImplicitContainer<T>
num_row_shards  Number of shards to use for the row factors.

ImplicitContainer<T>
num_col_shards  Number of shards to use for the column factors.

ImplicitContainer<T>
row_weights  Must be in one of the following three formats:  None: In this case, the weight of every entry is the unobserved_weight and the problem simplifies to ALS. Note that, in this case, col_weights must also be set to "None".  List of lists of nonnegative scalars, of the form \\([[w_0, w_1,...], [w_k,... ], [...]]\\), where the number of inner lists equal to the number of row factor shards and the elements in each inner list are the weights for the rows of that shard. In this case, \\(w_ij = unonbserved_weight + row_weights[i] * col_weights[j]\\).  A nonnegative scalar: This value is used for all row weights. Note that it is allowed to have row_weights as a list and col_weights as a scalar, or viceversa.

ImplicitContainer<T>
col_weights  See row_weights.

ImplicitContainer<T>
use_factors_weights_cache_for_training  Boolean, whether the factors and weights will be cached on the workers before the updates start, during training. Defaults to True. Note that caching is disabled during prediction.

ImplicitContainer<T>
use_gramian_cache_for_training  Boolean, whether the Gramians will be cached on the workers before the updates start, during training. Defaults to True. Note that caching is disabled during prediction.

object
max_sweeps  integer, optional. Specifies the number of sweeps for which to train the model, where a sweep is defined as a full update of all the row factors (resp. column factors). If `steps` or `max_steps` is also specified in model.fit(), training stops when either of the steps condition or sweeps condition is met.

object
model_dir  The directory to save the model results and log files.

object
config  A Configuration object. See Estimator.
Public properties
object COMPLETED_SWEEPS_dyn get; set;
object config get;
object config_dyn get;
object INPUT_COLS_dyn get; set;
object INPUT_ROWS_dyn get; set;
object LOSS_dyn get; set;
string model_dir get;
object model_dir_dyn get;
object model_fn get;
object model_fn_dyn get;
object PROJECT_ROW_dyn get; set;
object PROJECTION_RESULT_dyn get; set;
object PROJECTION_WEIGHTS_dyn get; set;
object PythonObject get;
object RWSE_dyn get; set;
Public fields
string PROJECT_ROW
return string

string PROJECTION_WEIGHTS
return string

string PROJECTION_RESULT
return string

string COMPLETED_SWEEPS
return string

string LOSS
return string

string INPUT_ROWS
return string

string RWSE
return string

string INPUT_COLS
return string
