tff.learning.robust_aggregator

Creates aggregator for mean with adaptive zeroing and clipping.

tff.learning.robust_aggregator(
    *,
    zeroing: bool = True,
    clipping: bool = True,
    weighted: bool = True,
    debug_measurements_fn: Optional[Callable[[tff.aggregators.AggregationFactory], tff.aggregators.AggregationFactory]] = None
) -> tff.aggregators.AggregationFactory

Zeroes out extremely large values for robustness to data corruption on clients, and clips in the L2 norm to moderately high norm for robustness to outliers.

Adaptive clipping approach is as described in Andrew, Thakkar et al. (2021) https://arxiv.org/abs/1905.03871, which the robust_aggregator applies without the addition of noise.

For details on clipping and zeroing see tff.aggregators.clipping_factory and tff.aggregators.zeroing_factory. For details on the quantile-based adaptive algorithm see tff.aggregators.PrivateQuantileEstimationProcess.

Args
zeroing	Whether to enable adaptive zeroing for data corruption mitigation.
clipping	Whether to enable adaptive clipping in the L2 norm for robustness.
weighted	Whether the mean is weighted (vs. unweighted).
debug_measurements_fn	A callable to add measurements suitable for debugging learning algorithms. Often useful values include None, tff.learning.add_debug_measurements or tff.learning.add_debug_measurements_with_mixed_dtype.

Returns
A tff.aggregators.AggregationFactory.

Raises
TypeError	if debug_measurement_fn yields an aggregation factory whose weight type does not match weighted.