autointent.modules.decision.TunableDecision

class autointent.modules.decision.TunableDecision(n_trials=320, seed=0, tags=None)

Bases: autointent.modules.abc.DecisionModule

Tunable predictor module.

TunableDecision uses an optimization process to find the best thresholds for predicting labels in single-label or multi-label classification tasks. It is designed for datasets with varying score distributions and supports out-of-scope (OOS) detection.

Variables:

• name – Name of the predictor, defaults to “tunable”.

• multilabel – Whether the task is multi-label classification.

• n_classes – Number of classes determined during fitting.

• tags – Tags for predictions, if any.

Parameters:

• n_trials (int)

• seed (int)

• tags (list[autointent.schemas.Tag] | None)

Examples

Single-label classification

import numpy as np
from autointent.modules import TunableDecision
scores = np.array([[0.2, 0.8], [0.6, 0.4], [0.1, 0.9]])
labels = [1, 0, 1]
predictor = TunableDecision(n_trials=100, seed=42)
predictor.fit(scores, labels)
test_scores = np.array([[0.3, 0.7], [0.5, 0.5]])
predictions = predictor.predict(test_scores)
print(predictions)

[1 0]

Multi-label classification

labels = [[1, 0], [0, 1], [1, 1]]
predictor = TunableDecision(n_trials=100, seed=42)
predictor.fit(scores, labels)
test_scores = np.array([[0.3, 0.7], [0.6, 0.4]])
predictions = predictor.predict(test_scores)
print(predictions)

[[1 1]
 [1 1]]

name = 'tunable'

multilabel: bool

n_classes: int

tags: list[autointent.schemas.Tag] | None

n_trials = 320

seed = 0

classmethod from_context(context, n_trials=320)

Initialize from context.

Parameters:

• context (autointent.context.Context) – Context

• n_trials (int) – Number of trials

Return type:

TunableDecision

fit(scores, labels, tags=None)

Fit module.

When data doesn’t contain out-of-scope utterances, using TunableDecision imposes unnecessary

computational overhead.

Parameters:

• scores (numpy.typing.NDArray[Any]) – Scores to fit

• labels (list[autointent.custom_types.LabelType]) – Labels to fit

• tags (list[autointent.schemas.Tag] | None) – Tags to fit

Return type:

None

predict(scores)

Predict the best score.

Parameters:

scores (numpy.typing.NDArray[Any]) – Scores to predict

Return type:

numpy.typing.NDArray[Any]

dump(path)

Dump all data needed for inference.

Parameters:

path (str) – Path to dump

Return type:

None

load(path)

Load data from dump.

Parameters:

path (str) – Path to load

Return type:

None