| Heterogeneous policy effects | example | example 1 | arxiv |
| Kernel approximation to Riesz representer | definition | definition 1 | arxiv |
| Lower bound; Theorem 2 | statement | lemma 6 | arxiv |
| Equivalence in a special case; Proposition 27 | statement | corollary 5 | arxiv |
| High probability events in $H$ | definition | definition 7 | arxiv |
| Nonasymptotic bound | statement | proposition 5 | arxiv |
| Special case: Kernel ridge regression; Theorem 1 | statement | corollary 7 | arxiv |
| Optimality | statement | corollary 8 | arxiv |
| Standard learning theory quantities | definition | definition 5 | arxiv |
| Towards a loss for KRRR | statement | lemma 3 | arxiv |
| generalization error to $H$ norm; c.f. Lemma 12 | statement | lemma 8 | arxiv |
| Loss for kernel ridge balancing weights | definition | definition 3 | arxiv |
| Residuals; Proposition S9 | statement | lemma 13 | arxiv |
| Characterization of key quantities; Theorem 2 | statement | lemma 10 | arxiv |
| Concentration; Proposition 2 | statement | lemma 15 | arxiv |
| Main corollary | statement | corollary 1 | arxiv |
| A practical result | statement | proposition 1 | arxiv |
| Oracle concentration | statement | lemma 14 | arxiv |
| Residual expansion | statement | lemma 12 | arxiv |
| New learning theory quantity | definition | definition 6 | arxiv |
| Inference for causal functions; Corollary S1 | statement | corollary 9 | arxiv |
| High probability events in $H$ | statement | proposition 4 | arxiv |
| Standalone closed form for KRRR | statement | proposition 2 | arxiv |
| Foldwise target and oracle | definition | definition 8 | arxiv |
| Shorter notation | definition | definition 10 | arxiv |
| Main theoretical result | statement | theorem 1 | arxiv |
| Debiased machine learning | definition | definition 4 | arxiv |
| Overall target and oracle | definition | definition 9 | arxiv |
| Loss for KRRR | definition | definition 2 | arxiv |
| Main lemma | statement | lemma 2 | arxiv |
| Riesz representation; Lemma S3.1 | statement | lemma 1 | arxiv |
| Standard learning theory bounds; Proposition 3 | statement | lemma 11 | arxiv |
| Balance-variance trade-off | statement | corollary 2 | arxiv |
| First order condition | statement | lemma 4 | arxiv |
| Heterogeneous treatment effects | example | example 2 | arxiv |
| Upper bound for KRRR | statement | theorem 2 | arxiv |
| Inference; Corollary 1 | statement | lemma 7 | arxiv |
| New learning theory bound | statement | lemma 9 | arxiv |
| Special case: Kernel ridge balancing weights; eq. 8 | statement | corollary 4 | arxiv |
| Non-equivalence in general | statement | corollary 6 | arxiv |
| Consistency under mis-specification | statement | proposition 3 | arxiv |
| Nonasymptotic bound under mis-specification | statement | proposition 6 | arxiv |
| Special case: Kernel ridge regression | statement | corollary 3 | arxiv |
| Standalone closed form exists; c.f. Lemma 2 | statement | lemma 5 | arxiv |
