Configuration spaces for computer systems can be challenging for traditional and automatic tuning strategies. Injecting task-specific knowledge into the tuner for a task may allow for more efficient exploration of candidate configurations. We apply this idea to the task of index set selection to accelerate database workloads. Index set selection has been amenable to recent applications of vanilla deep RL, but real deployments remain out of reach. In this paper, we explore how learning index selection can be enhanced with task-specific inductive biases, specifically by encoding these inductive biases in better action structures. Index selection-specific action representations arise when the problem is reformulated in terms of permutation learning and we rely on recent work for learning RL policies on permutations. Through this approach, we build an indexing agent that is able to achieve improved indexing and validate its behavior with task-specific statistics. Early experiments reveal that our agent can find configurations that are up to 40% smaller for the same levels of latency as compared with other approaches and indicate more intuitive indexing behavior.
AKI with incomplete epithelial repair is a major contributor to CKD characterized by tubulointerstitial fibrosis. Injury-induced epithelial secretion of profibrotic factors is hypothesized to underlie this link, but the identity of these factors and whether epithelial injury is required remain undefined. We previously showed that activation of the canonical Wnt signaling pathway in interstitial pericytes cell autonomously drives myofibroblast activation in vivo. Here, we show that inhibition of canonical Wnt signaling also substantially prevented TGFβ-dependent myofibroblast activation in vitro. To investigate whether Wnt ligand derived from proximal tubule is sufficient for renal fibrogenesis, we generated a novel mouse strain with inducible proximal tubule Wnt1 secretion. Adult mice were treated with vehicle or tamoxifen and euthanized at 12 or 24 weeks postinjection. Compared with vehicle-treated controls, kidneys with tamoxifen-induced Wnt1 expression from proximal tubules displayed interstitial myofibroblast activation and proliferation and increased matrix protein production. PDGF receptor β-positive myofibroblasts isolated from these kidneys exhibited increased canonical Wnt target gene expression compared with controls. Notably, fibrotic kidneys had no evidence of inflammatory cytokine expression, leukocyte infiltration, or epithelial injury, despite the close histologic correlation of each with CKD. These results provide the first example of noninflammatory renal fibrosis. The fact that epithelial-derived Wnt ligand is sufficient to drive interstitial fibrosis provides strong support for the maladaptive repair hypothesis in the AKI to CKD transition.