Collaborative machine learning and related techniques such as federated learning allow multiple participants, each with his own training dataset, to build a joint model by training locally and periodically exchanging model updates. We demonstrate that these updates leak unintended information about participants' training data and develop passive and active inference attacks to exploit this leakage. First, we show that an adversarial participant can infer the presence of exact data points-for example, specific locations-in others' training data (i.e., membership inference). Then, we show how this adversary can infer properties that hold only for a subset of the training data and are independent of the properties that the joint model aims to capture. For example, he can infer when a specific person first appears in the photos used to train a binary gender classifier. We evaluate our attacks on a variety of tasks, datasets, and learning configurations, analyze their limitations, and discuss possible defenses.

Generative models estimate the underlying distribution of a dataset to generate realistic samples according to that distribution. In this paper, we present the first membership inference attacks against generative models: given a data point, the adversary determines whether or not it was used to train the model. Our attacks leverage Generative Adversarial Networks (GANs), which combine a discriminative and a generative model, to detect overfitting and recognize inputs that were part of training datasets, using the discriminator's capacity to learn statistical differences in distributions.We present attacks based on both white-box and black-box access to the target model, against several state-of-the-art generative models, over datasets of complex representations of faces (LFW), objects (CIFAR-10), and medical images (Diabetic Retinopathy). We also discuss the sensitivity of the attacks to different training parameters, and their robustness against mitigation strategies, finding that defenses are either ineffective or lead to significantly worse performances of the generative models in terms of training stability and/or sample quality.

Minoxidil, a vasodilator medication known for its ability to slow or stop hair loss and promote hair regrowth, was first introduced, exclusively as an oral drug, to treat high blood pressure. It was however discovered to have the important side-effect of increasing growth or darkening of fine body hairs; this led to the development of a topical formulation as a 2% concentration solution for the treatment of female androgenic alopecia or 5% for treating male androgenic alopecia. Measurable changes disappear within months after discontinuation of treatment. The mechanism by which it promotes hair growth is not fully understood. Minoxidil is a potassium channel opener, causing hyperpolarization of cell membranes and it is also a vasodilator, it is speculated that, by widening blood vessels and opening potassium channels, it allows more oxygen, blood and nutrients to the follicle. This can also cause follicles in the telogen phase to shed, usually soon to be replaced by new, thicker hairs in a new anagen phase. It needs to be applied regularly, once or twice daily, for hair gained to be maintained, and side effects are common. The most common adverse reactions of the topical formulation are limited to irritant and allergic contact dermatitis on the scalp. There have been cases of allergic reactions to the nonactive ingredient propylene glycol, which is found in some topical solution especially if they are galenic. Increased hair loss which can occur during Minoxidil use, is due to the synchronization of the hair cycle that the treatment induces. In this review, we described its mechanism of action, use in dermatology and some patents related to alternative treatment of allergic reactions due to its use.

The identification of prognostic and predictive markers is crucial for choosing the most appropriate management method for ovarian cancer patients. We aimed to assess the prognostic role of tumorassociated macrophage (TAM) polarization in advanced ovarian cancer patients. We carried out a prospective observational study that included 140 consecutive patients with advanced-stage highgrade serous ovarian cancer as well as patients with other histotypes of ovarian cancer and patients with ovarian metastasis from other sites between June 2013 and December 2018. Patients were enrolled at the time of laparoscopic surgery before receiving any antineoplastic treatment. We found that patients with high-grade serous papillary ovarian cancers had a prevalence of M1 TAMs, a higher M1/M2 ratio, and a longer overall survival (OS) and progression-free survival (PFS) than other patients. Regression analysis confirmed that there was a significant positive association between the M1/M2 ratio and an improved OS, PFS and platinum-free interval (PFI), both in the entire population and in patients stratified according to tumor type and initial surgery. Kaplan-Meier analysis was performed after the patients were divided into 2 groups according to the median M1/M2 ratio and revealed that patients with a high M1/M2 ratio had a higher OS, PFS and PFI than those with a low M1/M2 ratio. In conclusion, the prognostic and predictive role of TAM polarization in the tumor microenvironment could be of great clinical relevance and may allow the early identification of patients who are likely to respond to therapy. Further studies in a larger prospective sample are warranted.www.nature.com/scientificreports www.nature.com/scientificreports/ linear relationship (β coefficient = 0.504, 95% CI 0.493-10.334, p = 0.033) was found between the platinum-free interval and M1/M2 ratio (Fig. 2).The median value for the M1/M2 ratio was 1.4. Then, we analyzed the difference in OS and PFS by Kaplan-Meier curve analysis and log-rank analysis between patients with an M1/M2 ratio ≥ 1.4 and those with an M1/M2 ratio < 1.4. We found that patients with an M1/M2 ratio ≥ 1.4 had a significantly longer OS (34 months versus 18 months, HR 2.7483, 95% CI: 1.1667-6.4736; p = 0.0207), PFS (24 months versus 9 months, HR 2.1285, 95% CI: 1.0461-4.3309; p = 0.0371) and PFI (12 months versus 6 months, HR 3.3959, 95% CI 1.2471-9.2469; p = 0.0168) ( Fig. 3). Scientific RepoRtS |(2020) 10:6096 | https://doi.

No abstract