Thanh Tu Nguyen

At diagnosis, metastatic prostate cancer (PC) is sensitive to androgen deprivation therapy (ADT), and patients are usually referred to as having castration-sensitive prostate cancer (mCSPC). The combination of ADT and androgen receptor pathway inhibitors (ARPI) is the current standard of care for mCSPC. This study aimed to review the literature on economic evaluations and health economic models related to mCSPC. A literature search was performed covering Medline, Embase, and Scopus with additional grey literature sources. Studies with data on health economic evaluations focusing on Europe or North America were relevant. 18 peer-reviewed articles and 10 grey literature documents were included. The majority (n = 23) had a deterministic Markov structure and applied either Markov cohort or partitioned survival models. Evaluations investigated various types of ADT-based combinations, comparing the addition of ARPI, chemotherapy agents, or radiation therapy to ADT alone. We concluded that economic evaluations in the field of PC are widely published, and there are a large number of publications even in the specific subgroup of mCSPC. Regardless of the investigated interventions, most studies applied similar methodologies and simulated patients from the mCSPC state until the development of mCRPC or death.

Adversarial machine learning defenses have primarily been focused on mitigating static, white-box attacks. However, it remains an open question whether such defenses are robust under an adaptive black-box adversary. In this paper, we specifically focus on the black-box threat model and make the following contributions: First we develop an enhanced adaptive black-box attack which is experimentally shown to be ≥ 30% more effective than the original adaptive black-box attack proposed by Papernot et al. For our second contribution, we test 10 recent defenses using our new attack and propose our own black-box defense (barrier zones). We show that our defense based on barrier zones offers significant improvements in security over state-of-the-art defenses. This improvement includes greater than 85% robust accuracy against black-box boundary attacks, transfer attacks and our new adaptive black-box attack, for the datasets we study. For completeness, we verify our claims through extensive experimentation with 10 other defenses using three adversarial models (14 different black-box attacks) on two datasets (CIFAR-10 and Fashion-MNIST).

The dressing regime parameters in the process of grinding are the most important enabling factors that need to be determined. In this study, the influence of dressing regime parameters (the rough dressing depth, the dressing feed rate, the fine dressing depth, rough dressing times, fine dressing times, none-feeding dressing times) on roundness tolerance for external grinding of SKD11 steel is introduced. The Taguchi technique in designing the experiment and analyzing of variance (ANOVA) have been applied to define the influence of dressing regime parameters on the roundness tolerance. The results show that the impact level of fine dressing times (n_f), none-feeding dressing times (n₀), fine dressing depth (a_f), rough dressing times (n_r), rough dressing depth (a_r), dressing feed rate (S_d), and errors on roundness tolerance (Rt) are 62.88%, 19.70%, 9.85%, 3.79%, 1.52%, 1.52% and 0.76% respectively. Finally, by analyzing experimental results, optimized dressing parameters with fine dressing times (n_f) of 3, none-feeding dressing times (n_o) of 3, fine dressing depth (a_f) of 0.01 mm, rough dressing times (n_r) of 3, rough dressing depth (a_r) of 0.03 mm and dressing feed rate (S_d) of 1.0 m/min have been determined, that allow getting the best roundness tolerance.