Live Roundups Editor's Desk Insights Executive Ops Weather About 🔍
Home › Science › Poor supervision is pushing young researchers out of academia
Science

Poor supervision is pushing young researchers out of academia

Nature Monday 01 June 2026, 00:00 UTC By Mohana Basu 1 min read
Key Points

Nature, Published online: 01 June 2026; doi:10.1038/d41586-026-01693-4A survey finds that academic supervisors can have a huge impact on their students’ mental health.

Nature, Published online: 01 June 2026; doi:10.1038/d41586-026-01693-4

A survey finds that academic supervisors can have a huge impact on their students’ mental health.
academia Nature (ORG) doi:10.1038 (ORG)
Originally published by Nature Read original →

Related Stories

Systematic comparison of VMEC and HINT equilibrium calculations for finite-beta LHD plasmas

arXiv:2606.10490v2 Announce Type: new Abstract: A systematic comparison between VMEC and HINT equilibrium calculations has been carried out for Large Helical Device plasmas to clarify the influence of the assumption of the nested flux surfaces at finite beta. Three vacuum magnetic-axis configurations, $R_{\rm axV} = \SI{3.53}{\, m}$, $\SI{3.60}{\, m}$, $\SI{3.85}{\, m}$, are examined for the beta values on the axis in the range $\beta_0 \in [0.0\%, 5.0\%]$. The magnetic-axis position, the...

arXiv Physics 1h ago

Mitigating Manifold Departure: Uncertainty-Aware Subspace Rectification for Trustworthy MLLM Decoding

arXiv:2606.09859v1 Announce Type: new Abstract: MLLMs frequently hallucinate objects inconsistent with visual inputs. This issue is typically attributed to the over-reliance on language priors, which can override the visual context. Recent training-free decoding strategies address this by penalizing language priors.

arXiv CS 1h ago

Support sufficiency as action-sufficient compression: a single-cycle rate-regret formulation

arXiv:2606.09858v1 Announce Type: new Abstract: Robust decision-making requires compression. A system that forms a rich support state cannot usually preserve its full structure at the point of action. It must retain only those distinctions needed to act, verify, abstain, or defer under the current consequence geometry.

arXiv CS 1h ago

Uncertainty-aware Multi-fidelity Closure via Conditional Normalizing Flows

Announce Type: new Abstract: Reduced-order models (ROMs) provide an efficient surrogate for complex multiscale systems, but their predictive accuracy is often compromised by truncation errors and the inadequate representation of interactions between resolved and unresolved scales. The missing effect of truncated (unresolved) scales on ROM (resolved) scales is often denoted as the closure problem. In this work, we formulate ROM closure modeling as a multi-fidelity (MF) learning problem and...

arXiv CS 1h ago