Search papers, labs, and topics across Lattice.
Affiliation:
5
0
7
Property-specific probing unlocks efficient, confidential multi-group fairness auditing, revealing bias structure without model reconstruction and resolving open sample complexity questions.
Dri-MED adapts to user preferences and context drifts, achieving significantly lower regret than traditional methods in dynamic environments.
Database tuning just got easier: DOT dynamically identifies and optimizes key parameters on-the-fly, outperforming existing methods without the need for costly warm-up phases.
Optimal sequential hypothesis testing with Markovian data is now possible, thanks to a new lower bound that incorporates both stationary distribution and transition structure, leading to an asymptotically optimal test.
Knowing the sequence of model updates lets you pinpoint when a specific data point was added to the training set, making membership inference attacks far more potent than previously thought.
