Source code for secmlt.optimization.constraints

"""Constraints for tensors and the corresponding batch-wise projections."""

from abc import ABC, abstractmethod
from typing import Union

import torch
from secmlt.adv.evasion.perturbation_models import LpPerturbationModels
from secmlt.models.data_processing.data_processing import DataProcessing
from secmlt.models.data_processing.identity_data_processing import (
    IdentityDataProcessing,
)


[docs] class Constraint(ABC): """Generic constraint."""
[docs] def __call__(self, x: torch.Tensor, *args, **kwargs) -> torch.Tensor: """ Project onto the constraint. Parameters ---------- x : torch.Tensor Input tensor. Returns ------- torch.Tensor Tensor projected onto the constraint. """ x_transformed = x.detach().clone() return self._apply_constraint(x_transformed)
@abstractmethod def _apply_constraint(self, x: torch.Tensor, *args, **kwargs) -> torch.Tensor: ...
[docs] class InputSpaceConstraint(Constraint, ABC): """Input space constraint. Reverts the preprocessing, applies the constraint, and re-applies the preprocessing. """
[docs] def __init__(self, preprocessing: DataProcessing) -> None: """ Create InputSpaceConstraint. Parameters ---------- preprocessing : DataProcessing Preprocessing to invert to apply the constraint on the input space. """ if preprocessing is None: preprocessing = IdentityDataProcessing() self.preprocessing = preprocessing
[docs] def __call__(self, x: torch.Tensor, *args, **kwargs) -> torch.Tensor: """ Project onto the constraint in the input space. Parameters ---------- x : torch.Tensor Input tensor. Returns ------- torch.Tensor Tensor projected onto the constraint. """ x_transformed = x.detach().clone() x_transformed = self.preprocessing.invert(x_transformed) x_transformed = self._apply_constraint(x_transformed) return self.preprocessing(x_transformed)
[docs] class ClipConstraint(Constraint): """Box constraint, usually for the input space."""
[docs] def __init__(self, lb: float = 0.0, ub: float = 1.0) -> None: """ Create box constraint. Parameters ---------- lb : float, optional Lower bound of the domain, by default 0.0. ub : float, optional Upper bound of the domain, by default 1.0. """ self.lb = lb self.ub = ub
[docs] def _apply_constraint(self, x: torch.Tensor, *args, **kwargs) -> torch.Tensor: """ Call the projection function. Parameters ---------- x : torch.Tensor Input samples. Returns ------- torch.Tensor Tensor projected onto the box constraint. """ return x.clamp(self.lb, self.ub)
[docs] class LpConstraint(Constraint, ABC): """Abstract class for Lp constraint."""
[docs] def __init__( self, radius: float = 0.0, center: float = 0.0, p: str = LpPerturbationModels.LINF, ) -> None: """ Create Lp constraint. Parameters ---------- radius : float, optional Radius of the constraint, by default 0.0. center : float, optional Center of the constraint, by default 0.0. p : str, optional Value of p for Lp norm, by default LpPerturbationModels.LINF. """ self.p = LpPerturbationModels.get_p(p) self.center = center self.radius = radius
[docs] @abstractmethod def project(self, x: torch.Tensor) -> torch.Tensor: """ Project onto the Lp constraint. Parameters ---------- x : torch.Tensor Input tensor. Returns ------- torch.Tensor Tensor projected onto the Lp constraint. """ ...
[docs] def _apply_constraint(self, x: torch.Tensor, *args, **kwargs) -> torch.Tensor: """ Project the samples onto the Lp constraint. Parameters ---------- x : torch.Tensor Input tensor. Returns ------- torch.Tensor Tensor projected onto the Lp constraint. """ x = x + self.center with torch.no_grad(): norm = torch.linalg.norm(x.flatten(start_dim=1), ord=self.p, dim=1) to_normalize = (norm > self.radius).view(-1, 1) proj_delta = self.project(x).flatten(start_dim=1) delta = torch.where(to_normalize, proj_delta, x.flatten(start_dim=1)) return delta.view(x.shape)
[docs] class L2Constraint(LpConstraint): """L2 constraint."""
[docs] def __init__(self, radius: float = 0.0, center: float = 0.0) -> None: """ Create L2 constraint. Parameters ---------- radius : float, optional Radius of the constraint, by default 0.0. center : float, optional Center of the constraint, by default 0.0. """ super().__init__(radius=radius, center=center, p=LpPerturbationModels.L2)
[docs] def project(self, x: torch.Tensor) -> torch.Tensor: """ Project onto the L2 constraint. Parameters ---------- x : torch.Tensor Input tensor. Returns ------- torch.Tensor Tensor projected onto the L2 constraint. """ flat_x = x.flatten(start_dim=1) diff_norm = flat_x.norm(p=2, dim=1, keepdim=True).clamp_(min=1e-12) flat_x = torch.where(diff_norm <= 1, flat_x, flat_x / diff_norm) * self.radius return flat_x.reshape(x.shape)
[docs] class LInfConstraint(LpConstraint): """Linf constraint."""
[docs] def __init__(self, radius: float = 0.0, center: float = 0.0) -> None: """ Create Linf constraint. Parameters ---------- radius : float, optional Radius of the constraint, by default 0.0. center : float, optional Center of the constraint, by default 0.0. """ super().__init__(radius=radius, center=center, p=LpPerturbationModels.LINF)
[docs] def project(self, x: torch.Tensor) -> torch.Tensor: """ Project onto the Linf constraint. Parameters ---------- x : torch.Tensor Input tensor. Returns ------- torch.Tensor Tensor projected onto the Linf constraint. """ return x.clamp(min=-self.radius, max=self.radius)
[docs] class L1Constraint(LpConstraint): """L1 constraint."""
[docs] def __init__(self, radius: float = 0.0, center: float = 0.0) -> None: """ Create L1 constraint. Parameters ---------- radius : float, optional Radius of the constraint, by default 0.0. center : float, optional Center of the constraint, by default 0.0. """ super().__init__(radius=radius, center=center, p=LpPerturbationModels.L1)
[docs] def project(self, x: torch.Tensor) -> torch.Tensor: """ Compute Euclidean projection onto the L1 ball for a batch. Source: https://gist.github.com/tonyduan/1329998205d88c566588e57e3e2c0c55 min ||x - u||_2 s.t. ||u||_1 <= eps Inspired by the corresponding numpy version by Adrien Gaidon. Parameters ---------- x : torch.Tensor Input tensor. Returns ------- torch.Tensor Projected tensor. Notes ----- The complexity of this algorithm is in O(dlogd) as it involves sorting x. References ---------- [1] Efficient Projections onto the l1-Ball for Learning in High Dimensions John Duchi, Shai Shalev-Shwartz, Yoram Singer, and Tushar Chandra. International Conference on Machine Learning (ICML 2008) """ original_shape = x.shape x = x.view(x.shape[0], -1) mask = (torch.norm(x, p=1, dim=1) < self.radius).float().unsqueeze(1) mu, _ = torch.sort(torch.abs(x), dim=1, descending=True) cumsum = torch.cumsum(mu, dim=1) arange = torch.arange(1, x.shape[1] + 1, device=x.device) rho, _ = torch.max((mu * arange > (cumsum - self.radius)) * arange, dim=1) theta = (cumsum[torch.arange(x.shape[0]), rho.cpu() - 1] - self.radius) / rho proj = (torch.abs(x) - theta.unsqueeze(1)).clamp(min=0) x = mask * x + (1 - mask) * proj * torch.sign(x) return x.view(original_shape)
[docs] class L0Constraint(LpConstraint): """L0 constraint."""
[docs] def __init__(self, radius: float = 0.0, center: float = 0.0) -> None: """ Create L0 constraint. Parameters ---------- radius : float, optional Radius of the constraint, by default 0.0. center : float, optional Center of the constraint, by default 0.0. """ if int(radius) != radius: msg = ( f"Pass either an integer or a float with no decimals for " f"the radius of an L0 constraint (current value: {radius})." ) raise ValueError(msg) super().__init__(radius=radius, center=center, p=LpPerturbationModels.L0)
[docs] def project(self, x: torch.Tensor) -> torch.Tensor: """ Project the samples onto the L0 constraint. Returns the sample with the top-k components preserved, and the rest set to zero. Parameters ---------- x : torch.Tensor Input samples. Returns ------- torch.Tensor Samples projected onto L0 constraint. """ flat_x = x.flatten(start_dim=1) _, topk_indices = torch.topk(flat_x.abs(), k=int(self.radius), dim=1) # zero out all values and scatter the top k values back proj = torch.zeros_like(flat_x) proj.scatter_(1, topk_indices, flat_x.gather(1, topk_indices)) return proj.view_as(x)
[docs] class QuantizationConstraint(InputSpaceConstraint): """Constraint for ensuring quantized outputs into specified levels."""
[docs] def __init__( self, preprocessing: DataProcessing = None, levels: Union[list[float], torch.Tensor, int] = 255, ) -> None: """ Create the QuantizationConstraint. Parameters ---------- preprocessing: DataProcessing Preprocessing to apply the constraint in the input space. levels : int, list[float] | torch.Tensor Number of levels or specified levels. """ if isinstance(levels, int | float): if levels < 2: # noqa: PLR2004 msg = "Number of levels must be at least 2." raise ValueError(msg) if int(levels) != levels: msg = "Pass an integer number of levels." raise ValueError(msg) # create uniform levels if an integer is provided self.levels = torch.linspace(0, 1, int(levels)) elif isinstance(levels, list): self.levels = torch.tensor(levels, dtype=torch.float32) elif isinstance(levels, torch.Tensor): self.levels = levels.type(torch.float32) if len(self.levels) < 2: # noqa: PLR2004 msg = "Number of custom levels must be at least 2." raise ValueError(msg) else: msg = "Levels must be an integer, list, or torch.Tensor." raise TypeError(msg) # sort levels to ensure they are in ascending order self.levels = self.levels.sort().values # noqa: PD011 super().__init__(preprocessing)
def _apply_constraint(self, x: torch.Tensor) -> torch.Tensor: # reshape x to facilitate broadcasting with custom levels x_expanded = x.unsqueeze(-1) # calculate the absolute difference between x and each custom level distances = torch.abs(x_expanded - self.levels) # find the index of the closest custom level nearest_indices = torch.argmin(distances, dim=-1) # quantize x to the nearest custom level return self.levels[nearest_indices]
[docs] class MaskConstraint(Constraint): """Constraint for keeping components only on the given mask."""
[docs] def __init__(self, mask: torch.Tensor) -> None: """ Create the MaskConstraint. Parameters ---------- mask : torch.Tensor Mask (1=apply, 0=mask) to enforce where the components are kept. """ self.mask = mask.type(torch.bool) super().__init__()
[docs] def _apply_constraint(self, x: torch.Tensor) -> torch.Tensor: """ Enforce the mask constraint. Parameters ---------- x : torch.Tensor Masked input tensor. Returns ------- torch.Tensor Input active only on the non-masked components. """ if self.mask.shape != x.squeeze().shape: msg = ( f"Shape of input ({x.shape}) and mask {self.mask.shape} does not match." ) raise ValueError(msg) return torch.where(self.mask, x, torch.zeros_like(x))