MoËT: Mixture of Expert Trees and its application to verifiable reinforcement learning.

Rapid advancements in deep learning have led to many recent breakthroughs. While deep learning models achieve superior performance, often statistically better than humans, their adoption into safety-critical settings, such as healthcare or self-driving cars is hindered by their inability to provide safety guarantees or to expose the inner workings of the model in a human understandable form. We present MoËT, a novel model based on Mixture of Experts, consisting of decision tree experts and a generalized linear model gating function. Thanks to such gating function the model is more expressive than the standard decision tree. To support non-differentiable decision trees as experts, we formulate a novel training procedure. In addition, we introduce a hard thresholding version, MoËTh, in which predictions are made solely by a single expert chosen via the gating function. Thanks to that property, MoËTh allows each prediction to be easily decomposed into a set of logical rules in a form which can be easily verified. While MoËT is a general use model, we illustrate its power in the reinforcement learning setting. By training MoËT models using an imitation learning procedure on deep RL agents we outperform the previous state-of-the-art technique based on decision trees while preserving the verifiability of the models. Moreover, we show that MoËT can also be used in real-world supervised problems on which it outperforms other verifiable machine learning models.

Chelidonium majus crude extract inhibits migration and induces cell cycle arrest and apoptosis in tumor cell lines.

ETHNOPHARMACOLOGICAL RELEVANCE: Chelidonium majus L (Papaveraceae) is widely used in alternative medicine for treatment of various disorders. Antitumor activities of alkaloids isolated from this plant have been reviewed, while there are only a few studies that examine properties of the whole extract. AIM OF THE STUDY: The aim of the present study was to investigate direct cytotoxic effects, as well as indirect antitumor effects of Chelidonium majus ethanolic extract against different tumor cell lines,. MATERIALS AND METHODS: MTT and SRB assays were performed to estimate cytotoxic effects of Chelidonium majus extract against human tumor cell lines A549, H460, HCT 116, SW480, MDA-MB 231 and MCF-7 and peripheral blood mononuclear cells from healthy individuals. Cell cycle analysis was performed by flow cytometry. Type of cell death induced by extract was determined by flow cytometry and cell morphology assessment. Inhibitory effect on migration of cancer cells was assessed by wound healing assay. RESULTS: Chelidonium majus extract showed selective time- and dose-dependent increase of cytotoxicity in all six cell lines, with individual cell line sensitivities. Extract promoted cell cycle arrest and induced apoptosis. Cotreatment with doxorubicin enhanced cytotoxicity of the drug. Also, inhibitory effect on migration was shown with non-toxic extract concentration. CONCLUSIONS: These results indicate possible usefulness of Chelidonium majus crude extract in antitumor therapy, whether through its direct cytotoxic effect, by prevention of metastasis, or as adjuvant therapy.