ABSTRACT
To alleviate the sparsity issue, many recommender systems have been proposed to consider the review text as the auxiliary information to improve the recommendation quality. Despite success, they only use the ratings as the ground truth for error backpropagation. However, the rating information can only indicate the users' overall preference for the items, while the review text contains rich information about the users' preferences and the attributes of the items. In real life, reviews with the same rating may have completely opposite semantic information. If only the ratings are used for error backpropagation, the latent factors of these reviews will tend to be consistent, resulting in the loss of a large amount of review information. In this article, we propose a novel deep model termed deep rating and review neural network (DRRNN) for recommendation. Specifically, compared with the existing models that adopt the review text as the auxiliary information, DRRNN additionally considers both the target rating and target review of the given user-item pair as ground truth for error backpropagation in the training stage. Therefore, we can keep more semantic information of the reviews while making rating predictions. Extensive experiments on four publicly available datasets demonstrate the effectiveness of the proposed DRRNN model in terms of rating prediction.
Subject(s)
Neural Networks, Computer , SemanticsABSTRACT
Emeriones A-C (1-3), three highly methylated polyketides with bicyclo[4.2.0]octene and 3,6-dioxabicyclo[3.1.0]hexane functionalities, were isolated from Emericella nidulans. An additional peroxide bridge in compound 3 led to the construction of an unexpected 7,8-dioxatricyclo[4.2.2.02,5]decene scaffold. The structures of 1-3 were elucidated by comprehensive spectroscopic techniques, and their absolute configurations were confirmed by single-crystal X-ray crystallographic analyses and ECD calculations. Compound 1 shows weak inhibitory effects on NO production in LPS-induced RAW264.7 cells.
Subject(s)
Alkenes/chemistry , Emericella/chemistry , Hexanes/chemistry , Polyketides/chemistry , Methylation , Models, Molecular , Molecular ConformationABSTRACT
(±)-Peniorthoesters A and B (±1 and ±2), two pairs of unprecedented spiro-orthoester enantiomers with a 1,4,6-trioxaspiro[4. 5]decane-7-one unit, were obtained from Penicillium minioluteum. Their structures were determined by spectroscopic methods, X-ray diffraction analyses, and ECD calculations. (±)-Peniorthoesters A and B are the first examples of spiro-orthoester enantiomers, and they represent the first spiro-orthoesters originating from fungi. All compounds showed potential inhibitory activities comparable to dexamethasone against NO production with IC50 values ranging from 14.2 to 34.5 µM.
ABSTRACT
Brasilanones A-F and asperterreusines A-C, undescribed brasilane sesquiterpenoids and dihydrobenzofuran derivatives, were isolated from the marine-derived fungus Aspergillus terreus [CFCC 81836]. Their structures with absolute configurations were elucidated on the basis of spectroscopic data, X-ray crystallographic analyses, and electronic circular dichroism (ECD) calculations. Brasilanones A-F are unusual brasilane sesquiterpenoids with an α,ß-unsaturated ketone unit, interestingly, brasilanones B-D are stereo isomers. All of the isolates were evaluated for their inhibitory activities against NO production and cytotoxic activities against five human cancer cell lines (HL-60, SW-480, A-549, MCF-7, and SMMC-7721). Brasilanones A and E showed moderate inhibitory effect with NO inhibition rates of 47.7% (pâ¯<â¯0.001) and 37.3% (pâ¯<â¯0.001) at the concentration of 40⯵M. Asperterreusines A showed cytotoxicity against HL-60 and SW-480â¯cell lines with IC50 values of 15.3 and 25.7⯵M, respectively.