cymr.operations.p_recall

cymr.operations.p_recall(int start, int n_f, int[:] recalls, double[:, :] w_fc_exp, const double[:, :] w_fc_pre, double[:, :] w_cf_exp, const double[:, :] w_cf_pre, double[:, :] w_ff_exp, const double[:, :] w_ff_pre, double[:] f, double[:] f_in, double[:] c, double[:] c_in, int[:, :] c_ind, int[:] exclude, double amin, double[:, :] B, double T, const double[:] p_stop, double[:] p)

Calculate the likelihood of each recall in a sequence.

Parameters:

• start – Start unit for the item segment to recall from.

• n_f – Number of units in the item segment.

• recalls – Item indices of recalls in output order.

• w_fc_exp – Weight matrix \(\mathbf{M}^{FC}_\mathrm{exp}\).

• w_fc_pre – Weight matrix \(\mathbf{M}^{FC}_\mathrm{pre}\).

• w_cf_exp – Weight matrix \(\mathbf{M}^{CF}_\mathrm{exp}\).

• w_cf_pre – Weight matrix \(\mathbf{M}^{CF}_\mathrm{pre}\).

• w_ff_exp – Weight matrix \(\mathbf{M}^{FF}_\mathrm{exp}\).

• w_ff_pre – Weight matrix \(\mathbf{M}^{FF}_\mathrm{pre}\).

• f – Item representation \(\mathbf{f}\).

• f_in – Item activation input \(\mathbf{f}^\mathrm{IN}\).

• c – Context state \(\mathbf{c}\).

• c_in – Context input \(\mathbf{c}^\mathrm{IN}\).

• c_ind – Start and end indices of the context sublayer.

• exclude – Vector of item indices to exclude from recall.

• amin – Minimum item activation for non-excluded items.

• B – \(\beta\) parameter.

• T – Temperature parameter of the softmax function.

• p_stop – Probability of stopping by output position.

• p – Likelihood of each recall and the stopping event.