Search papers, labs, and topics across Lattice.
3
0
5
0
Forget hyperparameter tuning for multi-treatment effect estimation: this work provides a theoretical estimator for optimal balancing weights, achieving O(1) scalability without sacrificing accuracy.
Overcome data scarcity in causal matrix completion with multiple treatments by borrowing information across treatment levels, achieving better accuracy without sacrificing theoretical guarantees.
Representing combinatorial treatments with permutation-invariant embeddings boosts uplift estimation accuracy and stability, especially for rare policies.
