Search for: All records

Maraging steels are known for their exceptional strength but suffer from limited work hardening and ductility. Here, we report an intermittent printing approach to tailor the microstructure and mechanical properties of maraging 250 steel via engineering of the thermal history during plasma arc additive manufacturing (PAAM). Through introducing a dwell time between adjacent layers, the maraging 250 steel is cooled below the martensite start temperature, triggering a thermally driven, in-situ martensitic transformation during the printing process. Re-heating or thermal cycling during subsequent layer deposition impedes complete martensitic transformation, enabling coexistence of martensite and retained austenite phases with elemental segregation. The enrichment of Ni in the austenite phase promotes stabilization of the retained austenite upon cooling down to room temperature. The retained austenite is yet metastable during deformation, leading to stress-induced martensitic transformation under loading. Specifically, a 3 min interlayer dwell time produces a maraging 250 steel with approximately 8% retained austenite, resulting in improved work hardening via martensitic transformation induced plasticity (TRIP) during deformation. Meanwhile, the higher cooling rate induced by the dwell time results in substantially refined grain structures with an increased dislocation density, leading to a simultaneously improved yield strength. Notably, the yield strength increases from 836 MPa (0 min dwell) to 990 MPa (3 min dwell), and the uniform elongation increases from 3.2% (0 min dwell) to 6.5% (3 min dwell). This intermittent deposition strategy demonstrates the potential to tune the microstructure and mechanical properties of maraging steels through engineering the thermal history during additive manufacturing. 

Federated Learning (FL) is a technique that allows multiple parties to train a shared model collaboratively without disclosing their private data. It has become increasingly popular due to its distinct privacy advantages. However, FL models can suffer from biases against certain demographic groups (e.g., racial and gender groups) due to the heterogeneity of data and party selection. Researchers have proposed various strategies for characterizing the group fairness of FL algorithms to address this issue. However, the effectiveness of these strategies in the face of deliberate adversarial attacks has not been fully explored. Although existing studies have revealed various threats (e.g., model poisoning attacks) against FL systems caused by malicious participants, their primary aim is to decrease model accuracy, while the potential of leveraging poisonous model updates to exacerbate model unfairness remains unexplored. In this paper, we propose a new type of model poisoning attack, EAB-FL, with a focus on exacerbating group unfairness while maintaining a good level of model utility. Extensive experiments on three datasets demonstrate the effectiveness and efficiency of our attack, even with state-of-the-art fairness optimization algorithms and secure aggregation rules employed. We hope this work will help the community fully understand the attack surfaces of current FL systems and facilitate corresponding mitigation to improve their resilience.

 

In this paper we derive the best constant for the following$L^{\infty }$-type Gagliardo-Nirenberg interpolation inequality$\begin{equation*} \|u\|_{L^{\infty }}\leq C_{q,\infty ,p} \|u\|^{1-\theta }_{L^{q+1}}\|\nabla u\|^{\theta }_{L^p},\quad \theta =\frac {pd}{dp+(p-d)(q+1)}, \end{equation*}$where parameters$q$and$p$satisfy the conditions$p>d\geq 1$,$q\geq 0$. The best constant$C_{q,\infty ,p}$is given by$\begin{equation*} C_{q,\infty ,p}=\theta ^{-\frac {\theta }{p}}(1-\theta )^{\frac {\theta }{p}}M_c^{-\frac {\theta }{d}},\quad M_c≔\int _{\mathbb {R}^d}u_{c,\infty }^{q+1} dx, \end{equation*}$where$u_{c,\infty }$is the unique radial non-increasing solution to a generalized Lane-Emden equation. The case of equality holds when$u=Au_{c,\infty }(\lambda (x-x_0))$for any real numbers$A$,$\lambda >0$and$x_{0}\in \mathbb {R}^d$. In fact, the generalized Lane-Emden equation in$\mathbb {R}^d$contains a delta function as a source and it is a Thomas-Fermi type equation. For$q=0$or$d=1$,$u_{c,\infty }$have closed form solutions expressed in terms of the incomplete Beta functions. Moreover, we show that$u_{c,m}\to u_{c,\infty }$and$C_{q,m,p}\to C_{q,\infty ,p}$as$m\to +\infty$for$d=1$, where$u_{c,m}$and$C_{q,m,p}$are the function achieving equality and the best constant of$L^m$-type Gagliardo-Nirenberg interpolation inequality, respectively.