Searcharxiv⌕ Search

arXiv subjects

Liangyu Zhu

Publications and source records attributed to Liangyu Zhu.

5 recordsLinked to original sources

A Note on the Cosupport of Balmer-Favi Idempotents

Let $g_x$ be the Balmer-Favi idempotent at a point $x$ of the spectrum of a rigidly-compactly generated tensor triangulated category. We study the equality $\operatorname{Cosupp}(g_x)=\operatorname{gen}(x)$ by expressing $g_x$ as a local dual. We prove the equality for the derived category of a Noetherian scheme and for the stable module category of a finite group scheme over a field of positive characteristic. For $E(n)$-local spectra we express the maps between distinct monochromatic layers of the sphere in terms of the chromatic fracture square. The height-two splitting gives the equality through height two at every prime. For a Noetherian local ring we characterize the Gorenstein property by the form of its closed-point idempotent.

math.CT↗

A Relative Telescope-Type Conjecture: Support and Generation in Half-Bounded Derived Categories

Let $R$ be a commutative noetherian ring. We ask whether complexes with finitely generated cohomology and a fixed support can be generated by relatively compact objects, using coproducts which exist in the half-bounded derived category. We prove that $\mathbf D^+_{\mathrm{fg},W}(R)=\operatorname{Loc}_{\mathrm{ex}}(\mathbf D^b_{\mathrm{fg},W}(R))$ whenever $W\subseteq\operatorname{Spec} R$ is specialization closed and has finite dimension. The same equality holds for every specialization-closed subset of $\operatorname{Spec} B$ when $B$ is a finite-type algebra over a noetherian domain $A$ of finite character and either $A$ or $B$ is regular. These equalities classify the thick Hom-submodules closed under existing coproducts. We also prove the corresponding bounded-above statement from the support results of Matsui--Takahashi. For an arbitrary noetherian ring, the bounded-below generation conjecture reduces to torsion complexes over quotient domains, with their support unchanged. Finally, over a regular noetherian domain of infinite dimension, we give a bounded-below object which is generated by bounded objects with the same support but has no bounded-below representative by finitely generated modules. The comparison functor from the bounded-below derived category of finite modules is therefore neither full nor essentially surjective.

math.AC↗

Spectra, Growth, and Localization in Half-Bounded Tensor Triangular Geometry

Over a noetherian scheme, classical cohomological support determines thick tensor ideals of perfect complexes. In half-bounded derived categories, complexes with the same support can generate different radical thick tensor ideals. We study how these distinctions affect the Balmer spectrum and localization. For a commutative noetherian ring $R$, we prove that $\operatorname{Spc}(\mathbf D^-_{\mathrm{fg}}(R))$ is noetherian if and only if $R$ is artinian. For a nonempty noetherian scheme $X$, the corresponding condition is $\dim X=0$. For such a ring $R$ and a finite group $G$, $\operatorname{Spc}(\mathbf D^-_{\mathrm{fg}}(RG))$ is noetherian if and only if $R$ is artinian and $|G|$ is invertible in $R$. Using the Sanders--Zhang classification over a discrete valuation ring, we compare objects whose annihilator exponents grow at different rates in different degrees. We lift the relevant generation relations before changing coefficients and obtain spectral retractions. These comparisons persist after thick quotients and algebraic localization under explicit uniform annihilation and nonvanishing conditions. A second argument uses nonnilpotent classes of positive degree and passes from cyclic subgroups to finite groups. Applications include relative results for schemes and bounded-below spectra with degreewise finitely generated homotopy. We also describe the cohomological information shared by the resulting ideals. The localization comparisons for bounded-below complexes of finitely generated modules and for bounded-below derived objects with finitely generated cohomology are established separately.

math.AG↗

Causal Effect Estimation and Optimal Dose Suggestions in Mobile Health

In this article, we propose novel structural nested models to estimate causal effects of continuous treatments based on mobile health data. To find the treatment regime that optimizes the expected short-term outcomes for patients, we define a weighted lag-K advantage as the value function. The optimal treatment regime is then defined to be the one that maximizes the value function. Our method imposes minimal assumptions on the data generating process. Statistical inference is provided for the estimated parameters. Simulation studies and an application to the Ohio type 1 diabetes dataset show that our method could provide meaningful insights for dose suggestions with mobile health data.

stat.ME↗

Kernel Assisted Learning for Personalized Dose Finding

An individualized dose rule recommends a dose level within a continuous safe dose range based on patient level information such as physical conditions, genetic factors and medication histories. Traditionally, personalized dose finding process requires repeating clinical visits of the patient and frequent adjustments of the dosage. Thus the patient is constantly exposed to the risk of underdosing and overdosing during the process. Statistical methods for finding an optimal individualized dose rule can lower the costs and risks for patients. In this article, we propose a kernel assisted learning method for estimating the optimal individualized dose rule. The proposed methodology can also be applied to all other continuous decision-making problems. Advantages of the proposed method include robustness to model misspecification and capability of providing statistical inference for the estimated parameters. In the simulation studies, we show that this method is capable of identifying the optimal individualized dose rule and produces favorable expected outcomes in the population. Finally, we illustrate our approach using data from a warfarin dosing study for thrombosis patients.

stat.ME↗