derivative_noiser

Provides derivative noisers.

Derivative noisers are used to add noise to the derivative function of a prediction task. This way, noise can be propageted through time.

class AdditiveDerivNoiser(distribution: Distribution[float])

Bases: object

The AdditiveDerivNoiser is a noiser, that adds random-values to the derivative.

distribution

A distribution to generate random noise.

Inits AdditiveDerivNoiser with the provided params.

Parameters:

distribution – A distribution to generate random noise from.

noisify(deriv: Callable[[float, list[float], dict[str, float]], tuple[float, ...]], max_t: float) → Callable[[float, list[float], dict[str, float]], tuple[float, ...]]

Applies noise to the provided derivative function.

Parameters:

• deriv – Derivative function.

• max_t – Adds noise up to this timestep.

Returns:

Noised derivative function.

class DerivNoiser(*args, **kwargs)

Bases: Protocol[KineticParameterType]

A DerivNoiser is a Noiser, that noises a derivative function.

abstract noisify(deriv: Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]], max_t: float) → Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]]

Noises the derivative.

Parameters:

• deriv – The derivative function, that needs to be noised.

• max_t – Adds noise up to this timestep.

class MultiDerivNoiser(noisers: list[simba_ml.simulation.derivative_noiser.derivative_noiser.DerivNoiser[KineticParameterType]])

Bases: object

Applies one randomly selected DerivNoiser to noise a derivative function.

noisers

A list of derivative_noiser.DerivNoiser to choose from.

Inits MultiDerivNoiser with the provided arguments.

Parameters:

noisers – A list of derivative_noiser.DerivNoiser to choose from.

noisify(deriv: Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]], max_t: float) → Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]]

Applies noise to the provided derivative function.

Parameters:

• deriv – Derivative function.

• max_t – Adds noise up to this timestep.

Returns:

Noised derivative function.

class MultiplicativeDerivNoiser(distribution: Distribution[float])

Bases: object

Multiplies each element individually with a randomly generated number.

The number is generated using a selected InitialCondition.

distribution

A distribution to generate random noise.

Inits MultiplicativeDerivNoiser with the provided params.

Parameters:

distribution – A distribution to generate random noise from.

noisify(deriv: Callable[[float, list[float], dict[str, float]], tuple[float, ...]], max_t: float) → Callable[[float, list[float], dict[str, float]], tuple[float, ...]]

Applies noise to the provided derivative function.

Parameters:

• deriv – Derivative function.

• max_t – Adds noise up to this timestep.

Returns:

Noised derivative function.

class NoDerivNoiser

Bases: object

The NoDerivNoiser is a dummy DerivNoiser, that applies no noise.

noisify(deriv: Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]], _max_t: float) → Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]]

Returns the input signal.

Parameters:

deriv – Derivative function.

Returns:

Not noised derivative function.

class SequentialDerivNoiser(noisers: list[simba_ml.simulation.derivative_noiser.derivative_noiser.DerivNoiser[KineticParameterType]])

Bases: object

The SequentialNoiser applies multiple given DerivNoiser sequentially.

noisers

A list of DerivNoiser to be applied.

Inits SequentialNoiser with the provided arguments.

Parameters:

noisers – A list of DerivNoiser to be applied.

noisify(deriv: Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]], max_t: float) → Callable[[float, list[float], dict[str, KineticParameterType]], tuple[float, ...]]

Applies noise to the provided derivative function.

Parameters:

• deriv – Derivative function.

• max_t – Adds noise up to this timestep.

Returns:

Noised derivative function.

simba_ml.simulation.derivative_noiser.additive_deriv_noiser	Provides the AdditiveDerivNoiser.
simba_ml.simulation.derivative_noiser.derivative_noiser	Defines multiple classes for applying noise to a derivative.
simba_ml.simulation.derivative_noiser.multi_deriv_noiser	Provides the MultiDerivNoiser.
simba_ml.simulation.derivative_noiser.multiplicative_deriv_noiser	Provides the MultiplicativeDerivNoiser.
simba_ml.simulation.derivative_noiser.no_deriv_noiser	Provides the NoDerivNoiser.
simba_ml.simulation.derivative_noiser.sequential_deriv_noiser	Provides the SequentialDerivNoiser.