University of Findlay

Accurate 3D medical segmentation is limited by anatomical variability and high computational costs. While knowledge distillation (KD) offers a route for model compression, conventional methods often fail to preserve complex structures and are overwhelmed by background noise. We propose Displacement-Preserving Relational Distillation (DPRD), which distills latent anatomical trajectories via vector based alignment to preserve the orientation and relative scale of the teacher's manifold, and prevents signal dilution by anchoring distillation in task-relevant structures. Integrated into nnU-Net, DPRD outperforms established baselines on ISLES 2022 and AMOS 2022 benchmarks. Notably, on the AMOS dataset, DPRD achieves a Dice score of 85.46

On-policy self-distillation (OPSD) gives language agents dense token-level supervision from a privileged self-teacher on the policy's own trajectories. Existing methods allocate this supervision mainly by teacher trust, but trust does not reveal whether emphasizing a token supports the current policy objective. We call this the trust-utility mismatch and introduce Influence Calibration for Self-Distillation (ICSD). For each supervised token, ICSD measures the first-order response of its importance-weighted RL surrogate contribution to a teacher-directed output perturbation. Batch-adaptive calibration converts this non-stationary signal into a bounded allocation weight while preserving the original auxiliary-loss mass within each action turn. These detached weights affect only the distillation loss and require no additional model pass. Across ALFWorld, WebShop, and Search-QA, ICSD improves all matched aggregate metrics over trust-only allocation under Group Relative Policy Optimization (GRPO) and Group-in-Group Policy Optimization (GiGPO), across two model families spanning 1.5B to 7B. At 7B, it reaches 96.1

Using linear morphometric measurements and applying the size dimorphism index (SDI), allometric analysis, along with both univariate and multivariate approaches, we evaluated sexual shape dimorphism in adults of the Mesquite Lizard (Sceloporus grammicus) from a population in central Mexico. Most SDI values (tail length, head length, head width, head height, femur length, and largest toe length) were biased towards males; the SDI for interlimb distance was biased towards females. In our population, this lizard does not show sexual dimorphism in SVL. Females showed an isometric relationship in five traits, and hypoallometry in femur and largest toe length. Males also showed an isometric relationship in five traits, and hypoallometry in the head length and largest toe length. Body proportions differed between sexes: tail length and femur length increased more rapidly with SVL for males than for females. Additionally, length, width and height of head, and largest toe length were greater in males than in females when accounting for SVL. The interlimb distance was greater for females than for males when accounting for SVL. PCA revealed clear separation between sexes along PCI (accounting 75%. Tail length showed the highest loading associated with males, while the interlimb distance had the highest loading associated with females. A MANOVA showed significant multivariate sexual dimorphism over all seven morphometric variables. Although this population is not sexually dimorphic in SVL, the observed sexual dimorphism in body proportions was similar to those reported in other lizard species and may be influenced by selective pressures linked to both sexual and fecundity selection.