Search

Background: Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related death in the world. Nanosecond pulsed electric fields (nsPEFs) have emerged as a new treatment for cancer. This study aims to identify the effectiveness of nsPEFs in the treatment of HCC and analyze the alterations in the gut microbiome and serum metabonomics after ablation.Methods: C57BL/6 mice were randomly divided into three groups: healthy control mice (n = 10), HCC mice (n = 10), and nsPEF-treated HCC mice (n = 23). Hep1-6 cell lines were used to establish the HCC model in situ. Histopathological staining was performed on tumor tissues. The gut microbiome was analyzed by 16S rRNA sequencing. Serum metabolites were analyzed by liquid chromatography–mass spectrometry (LC-MS) metabolomic analysis. Spearman’s correlation analysis was carried out to analyze the correlation between the gut microbiome and serum metabonomics.Results: The fluorescence image showed that nsPEFs were significantly effective. Histopathological staining identified nuclear pyknosis and cell necrosis in the nsPEF group. The expression of CD34, PCNA, and VEGF decreased significantly in the nsPEF group. Compared with normal mice, the gut microbiome diversity of HCC mice was increased. Eight genera including Alistipes and Muribaculaceae were enriched in the HCC group. Inversely, these genera decreased in the nsPEF group. LC-MS analysis confirmed that there were significant differences in serum metabolism among the three groups. Correlation analysis showed crucial relationships between the gut microbiome and serum metabolites that are involved in nsPEF ablation of HCC.Conclusion: As a new minimally invasive treatment for tumor ablation, nsPEFs have an excellent ablation effect. The alterations in the gut microbiome and serum metabolites may participate in the prognosis of HCC ablation.

Online marketplace is a digital platform that connects buyers (demand) and sellers (supply) and provides exposure opportunities that individual participants would not otherwise have access to. The KDD-22 Workshop on Decision Intelligence and Analytics for Online Marketplaces: Jobs, Ridesharing, Retail, and Beyond brought together academics and practitioners working on a broad array of online marketplaces to appreciate the diversity in applications, to draw connections to inform decision optimization across different industries, and to discover new problems that are fundamental to marketplaces of different domains. The workshop followed a dual-track format, covering marketplaces in general and job marketplaces specifically.

Both of the surface topographical features and distribution of biochemical cues can influence the cell-substrate interactions and thereby tissue regeneration in vivo. However, they have not been combined simultaneously onto a biodegradable scaffold to demonstrate the synergistic role so far. In this study, a proof-of-concept study is performed to prepare micropatterns and peptide gradient on the inner wall of a poly (D,L-lactide-co-caprolactone) (PLCL) guidance conduit and its advantages in regeneration of peripheral nerve in vivo. After linear ridges/grooves of 20/40 μm in width are created on the PLCL film, its surface is aminolyzed in a kinetically controlled manner to obtain the continuous gradient of amino groups, which are then transferred to CQAASIKVAV peptide density gradient via covalent coupling of glutaraldehyde. The Schwann cells are better aligned along with the stripes, and show a faster migration rate toward the region of higher peptide density. Implantation of the nerve guidance conduit made of the PLCL film having both the micropatterns and peptide gradient can significantly accelerate the regeneration of sciatic nerve in terms of rate, function recovery and microstructures, and reduction of fibrosis in muscle tissues. Moreover, this nerve conduit can also benefit the M2 polarization of macrophages and promote vascularization in vivo., Graphical abstract Dual gradient of nerve-affinitive CQAASIKVAV peptides and linear ridges/grooves on the inner surface of biodegradable guidance conduit synergistically promote the regeneration of sciatic nerve in terms of rate, function recovery and microstructures, and reduction of fibrosis in muscle tissues in vivo.Image 1, Highlights • CQAASIKVAV peptide gradient and micropatterns are integrated on inner surface of a nerve regeneration conduit. • Cell alignment and migration are promoted toward the region of a higher peptide density. • The peptide gradient and micropatterns can synergistically accelerate the regeneration of sciatic nerve in vivo.

ObjectiveWe investigate the characteristics of histological damage to myocardial cells in the ablation region and surrounding areas of the left ventricular epicardium in rabbits using our self-developed cardiac pulsed electric field (PEF) ablation instrument and ablation catheter.MethodsForty eight New Zealand rabbits underwent ablation on the left ventricular myocardium after open-heart exposure with a cardiac arrhythmia PEF ablation device and ablation catheter developed by the Medical Translation Laboratory of Pulsed Electric Field Technology in Zhejiang Province. The ablation parameters were set as biphasic electrical pulses; voltage, ±800 V; pulse width, 10 μs; interphase delay, 500 us. Six rabbits were included in the sham group and 42 other rabbits were randomly divided into immediately, 6-h, 1-, 3-day, 1-, 2-, and 4-week post-ablation groups, with six rabbits in each group. Creatine kinase- (CK)-MB isoenzyme (CK-MB), aspartate aminotransferase (AST), and lactate dehydrogenase (LDH) levels were measured before and at different time points after PEF ablation to analyze their dynamic evolution. Masson staining of tissue block sections of left ventricular myocardial ablation and adjacent tissue heart specimens was performed, and the occurrence of TUNEL apoptosis in myocardium tissue was analyzed.ResultsAll rabbits completed the PEF ablation procedure and the follow-up process. After PEF ablation, the levels of cardiac enzymes, including CK-MB, CK, and AST, increased significantly, peaking 1–3 days after the procedure. In particular, those of CK and CK-MB increased by 15–20 times but returned to the preoperative level after 2 weeks. Based on general observation, it was found that the myocardium in the ablation area was swollen immediately after PEF ablation. Masson staining analysis revealed that cardiomyocytes were broken and infiltrated by erythrocytes after 6 h. After 1 day, the cells started to experience atrophy and necrosis; after 3 days, fibrotic replacement of the necrotic area became obvious. Then, by 4 weeks, the myocardial cells were completely replaced by hyperplasia. Apoptosis occurred significantly at 6 h and peaked at 24 h post-ablation, demonstrating a 37.7-fold increase; apoptotic cell counts decreased significantly at 3 days post-ablation, and no significant apoptotic cardiomyocytes were seen after 1 week.ConclusionAfter PEF ablation, cardiomyocytes showed apoptotic process and dyed, at least partially, through a secondary necrosis, the ablation boundary was clear, the ablation area was replaced by structurally intact fibroblasts, no island myocardium tissue were seen, and the ablation area vessels and nerves were not affected.

Keratin 8 (KRT8) is the major component of the intermediate filament cytoskeleton and aberrant expression in multiple tumors. However, the role of KRT8 in lung adenocarcinoma (LUAD) remains unclear. In the present study, KRT8 expression was found to be upregulated along with prognosis and metastasis in LUAD. Kaplan–Meier analysis presented that the 5-year OS and DSS rates were significantly better among patients with low KRT8 expression compared to those with high expression. Correlation analysis showed that KRT8 expression was significantly associated with gender (P = 0.027), advanced T stage (P = 0.001), advanced N stage (P = 0.048), and advanced pathologic stage (P = 0.025). Univariate Cox analysis demonstrated that KRT8 was a predictor of OS [hazard ratio (HR) = 1.526; 95% confidence interval (CI) 1.141–2.040; P = 0.004] and DSS (HR = 1.625; 95% CI 1.123–2.353; P = 0.010) in the TCGA database. Importantly, downregulation of KRT8 obviously suppressed cell proliferation, cell migration, invasion, and EMT as well as induced cell apoptosis. KRT8 knockdown significantly inhibited NF-κB signaling, suggesting a potential mechanism. Overall, our results indicated that KRT8 could regulate lung carcinogenesis and may serve as a potential target for antineoplastic therapies.

Chemodynamic therapy (CDT)-activated apoptosis is a potential anticancer strategy. However, CDT encounters a bottleneck in clinical translation due to its serious side effects and low efficacy. Here, we first reveal that surface engineering of ginsenoside Rg3 dramatically alters the organ distribution and tumor enrichment of systematically administered nanocatalysts using the orthotopic pancreatic tumor model while avoiding toxicity and increasing efficacy in vivo to address the key and universal toxicity problems encountered in nanomedicine. Compared with nanocatalysts alone, Rg3-sheltered dynamic nanocatalysts form hydrophilic nanoclusters, prolonging their circulation lifespan in the blood, protecting the internal nanocatalysts from leakage while allowing their specific release at the tumor site. Moreover, the nanoclusters provide a drug-loading platform for Rg3 so that more Rg3 reaches the tumor site to achieve obvious synergistic effect with nanocatalysts. Rg3-sheltered dynamic nanocatalysts can simultaneously activate ferroptosis and apoptosis to significantly improve anticancer efficacy. Systematic administration of ginsenoside Rg3-sheltered nanocatalysts inhibited 86.6% of tumor growth without toxicity and prolonged the survival time of mice. This study provides a promising approach of nanomedicine with high biosafety and a new outlook for catalytic ferroptosis-apoptosis combined antitumor therapies. STATEMENT OF SIGNIFICANCE: Chemodynamic therapy (CDT) has limited clinical efficacy in cancer. In this study, we developed Rg3-sheltered dynamic nanocatalysts, which could simultaneously activate ferroptosis based on CDT-activated apoptosis, and ultimately form a combined therapy of ferroptosis-apoptosis to kill tumors. Studies have shown that the nanocatalysts after Rg3 surface engineering dramatically alters the pharmacokinetics and organ distribution of the nanocatalysts after being systematically administered, resulting in avoiding the toxicity of the nanocatalysts. Nanocatalysts also act as a drug-loading platform, guiding more Rg3 into the tumor site. This study emphasizes that nanocatalysts after Rg3 surface engineering improve the safety and effectiveness of ferroptosis-apoptosis combined therapy, providing an effective idea for clinical practices.

Autophagy triggered by reactive oxygen species (ROS) in photodynamic therapy (PDT) generally exhibits an anti-apoptotic effect to promote cell survival. Herein, an innovative supramolecular nanoplatform was fabricated for enhanced PDT by converting the role of autophagy from pro-survival to pro-death. The respiration inhibitor 3-bromopyruvate (3BP), which can act as an autophagy promoter and hypoxia ameliorator, was integrated into photosensitizer chlorin e6 (Ce6)-encapsulated nanoparticles to combat hypoxic tumor. 3BP could inhibit respiration by down-regulating HK-II and GAPDH expression to significantly reduce intracellular oxygen consumption rate, which could relieve tumor hypoxia for enhanced photodynamic cancer therapy. More importantly, the autophagy level was significantly elevated by the combination of 3BP and PDT determined by Western blot, immunofluorescent imaging, and transmission electron microscopy. It was very surprising that excessively activated autophagy promoted cell apoptosis, leading to the changeover of autophagy from pro-survival to pro-death. Therefore, PDT combined with 3BP could achieve efficient cell proliferation inhibition and tumor regression. Furthermore, hypoxia-inducible factor-1α (HIF-1α) could be down-regulated after tumor hypoxia was relieved by 3BP. Tumor metastasis could then be effectively inhibited by eliminating primary tumors and down-regulating HIF-1α expression. These results provide an inspiration for future innovative approaches of cancer therapy by triggering pro-death autophagy.

Regeneration and functional recovery of peripheral nerves remain formidable due to the inefficient physical and chemical cues in the available nerve guidance conduits (NGCs). Introducing micropatterns and bioactive substances into the inner wall of NGCs can effectively regulate the behavior of Schwann cells, the elongation of axons, and the phenotype of macrophages, thereby aiding the regeneration of injured nerve. In this study, linear micropatterns with ridges and grooves of 3/3, 5/5, 10/10, and 30/30 μm were created on poly(d,l-lactide-co-caprolactone) (PLCL) films following with surface aminolysis and electrostatic adsorption of graphene oxide (GO) nanosheets. The GO-modified micropatterns could significantly accelerate the collective migration of Schwann cells (SCs) and migration of SCs from their spheroids in vitro. Moreover, the SCs migrated directionally along the stripes with a fastest rate on the 3/3-GO film that had the largest cell adhesion force. The neurites of N2a cells were oriented along the micropatterns, and the macrophages tended to differentiate into the M2 type on the 3/3-GO film judged by the higher expression of Arg 1 and IL-10. The systematic histological and functional assessments of the regenerated nerves at 4 and 8 weeks post-surgery in vivo confirmed that the 3/3-GO NGCs had better performance to promote the nerve regeneration, and the CMAP, NCV, wet weight of gastrocnemius muscle, positive S100β and NF200 area percentages, and average myelinated axon diameter were more close to those of the autograft group at 8 weeks. This type of NGCs thus has a great potential for nerve regeneration.

Ischemic stroke is an acute and serious cerebral vascular disease, which greatly affects people's health and brings huge economic burden to society. Microglia, as important innate immune components in central nervous system (CNS), are double-edged swords in the battle of nerve injury, considering their polarization between pro-inflammatory M1 or anti-inflammatory M2 phenotypes. High mobility group box 1 (HMGB1) is one of the potent pro-inflammatory mediators that promotes the M1 polarization of microglia. 18β-glycyrrhetinic acid (GA) is an effective intracellular inhibitor of HMGB1, but of poor water solubility and dose-dependent toxicity. To overcome the shortcomings of GA delivery and to improve the efficacy of cerebral ischemia therapy, herein, we designed reactive oxygen species (ROS) responsive polymer-drug conjugate nanoparticles (DGA) to manipulate microglia polarization by suppressing the translocation of nuclear HMGB1. DGA presented excellent therapeutic efficacy in stroke mice, as evidenced by the reduction of infarct volume, recovery of motor function, suppressed of M1 microglia activation and enhanced M2 activation, and induction of neurogenesis. Altogether, our work demonstrates a close association between HMGB1 and microglia polarization, suggesting potential strategies for coping with inflammatory microglia-related diseases.

Keratin 8 (KRT8) is the major component of the intermediate filament cytoskeleton and aberrant expression in multiple tumors. However, the role of KRT8 in lung adenocarcinoma (LUAD) remains unclear. In the present study, KRT8 expression was found to be upregulated along with prognosis and metastasis in LUAD. Kaplan-Meier analysis presented that the 5-year OS and DSS rates were significantly better among patients with low KRT8 expression compared to those with high expression. Correlation analysis showed that KRT8 expression was significantly associated with gender (P = 0.027), advanced T stage (P = 0.001), advanced N stage (P = 0.048), and advanced pathologic stage (P = 0.025). Univariate Cox analysis demonstrated that KRT8 was a predictor of OS [hazard ratio (HR) = 1.526; 95% confidence interval (CI) 1.141-2.040; P = 0.004] and DSS (HR = 1.625; 95% CI 1.123-2.353; P = 0.010) in the TCGA database. Importantly, downregulation of KRT8 obviously suppressed cell proliferation, cell migration, invasion, and EMT as well as induced cell apoptosis. KRT8 knockdown significantly inhibited NF-κB signaling, suggesting a potential mechanism. Overall, our results indicated that KRT8 could regulate lung carcinogenesis and may serve as a potential target for antineoplastic therapies.

The coupling of nanozymes with ginsenoside Rg3 to form a nanodrug completely compensates for their respective defects and results in obvious anticancer synergistic effects, which provides a clear direction for the subsequent clinical application.

Background: Nanosecond pulsed electric field (nsPEF) is a novel ablative technique for treating tumours of the liver, kidneys, and other organs. This study investigated the effect of nsPEF on hepatic veins of different diameters and evaluated the feasibility and safety of using nsPEF to ablate tissues in porcine livers.Methods: A total of 15 pigs were included in the study. B-mode ultrasonography was used to guide two-needle nsPEF electrodes inserted adjacent to the left, middle, and right hepatic veins. The animals were followed up at 1 hour and at 1, 3, 7, and 14 days post treatment. Results: During and after the procedure of nsPEF treatment, electrocardiographs (ECGs) detected no cardiovascular events, and B-mode ultrasonography found no hepatic venous thrombosis formation. Follow-up at 1 hour to 7 days post treatment revealed complete ablation of the perivascular tissue within the ablated area. There was no apparent structural damage to the hepatic veins of different diameters, but there appeared to be red blood cells and residual inflammatory cell infiltration in the subendothelial layer from 1 hour to 7 days post treatment. Follow-up blood biochemical tests showed transient damage to the liver and myocardial function, and immediate recovery, which is consistent with the pathological changes.Conclusion: NsPEF achieved complete liver tissue ablation while causing little damage to the hepatic veins. Thus, nsPEF could be used to ablate central tumour lesions that are located nearing the major hepatic veins, thereby avoiding some of the limitations of conventional surgical therapies and thermal ablations.

The widespread usage of e-commerce websites in daily life and the resulting wealth of implicit feedback data form the foundation for systems that train and test e-commerce search ranking algorithms. While convenient to collect, implicit feedback data inherently suffers from various types of bias since user feedback is limited to products they are exposed to by existing search ranking algorithms and impacted by how the products are displayed. In the literature, a vast majority of existing methods have been proposed towards unbiased learning to rank for list-based web search scenarios. However, such methods cannot be directly adopted by e-commerce websites mainly for two reasons. First, in e-commerce websites, search engine results pages (SERPs) are displayed in 2-dimensional grids. The existing methods have not considered the difference in user behavior between list-based web search and grid-based product search. Second, there can be multiple types of implicit feedback (e.g., clicks and purchases) on e-commerce websites. We aim to utilize all types of implicit feedback as the supervision signals. In this work, we extend unbiased learning to rank to the world of e-commerce search via considering a grid-based product search scenario. We propose a novel framework which (1) forms the theoretical foundations to allow multiple types of implicit feedback in unbiased learning to rank and (2) incorporates the row skipping and slower decay click models to capture unique user behavior patterns in grid-based product search for inverse propensity scoring. Through extensive experiments on real-world e-commerce search log datasets across browsing devices and product taxonomies, we show that the proposed framework outperforms the state of the art unbiased learning to rank algorithms. These results also reveal important insights on how user behavior patterns vary in e-commerce SERPs across browsing devices and product taxonomies.

The excessive reactive oxygen species (ROS) and hypoxia deteriorate the inflammation-related diseases such as myocardial infarction (MI), and thereby deter the normal tissue repair and recovery and further lead to severe fibrosis and malfunction of tissues and organs. In particular, the MI has become one of the leading causes of death nowadays. In this study, a novel type of injectable hydrogel with dual functions of ROS scavenging and O

There is an increasing attention on next-item recommendation systems to infer the dynamic user preferences with sequential user interactions. While the semantics of an item can change over time and across users, the item correlations defined by user interactions in the short term can be distilled to capture such change, and help in uncovering the dynamic user preferences. Thus, we are motivated to develop a novel next-item recommendation framework empowered by sequential hypergraphs. Specifically, the framework: (i) adopts hypergraph to represent the short-term item correlations and applies multiple convolutional layers to capture multi-order connections in the hypergraph; (ii) models the connections between different time periods with a residual gating layer; and (iii) is equipped with a fusion layer to incorporate both the dynamic item embedding and short-term user intent to the representation of each interaction before feeding it into the self-attention layer for dynamic user modeling. Through experiments on datasets from the ecommerce sites Amazon and Etsy and the information sharing platform Goodreads, the proposed model can significantly outperform the state-of-the-art in predicting the next interesting item for each user.

The combination of different modalities greatly enhances the anticancer efficacy of each treatment by combining their merits, showing promising potential in clinical translation. Herein, we fabricated cancer cell membrane-coated gold nanorods (GNR@Mem) possessing excellent photothermal transfer ability in the second near-infrared window and radiosensitizing ability under X-ray irradiation. The cancer cell membrane coating endowed the nanomedicine with stability in the physiological environment and selective homotypic targeting to specific cancer cells in vitro. Under NIR light and X-ray irradiation, the gold nanorods induced a temperature increase, reactive oxygen generation, and subsequent damage to the DNA helix structure, leading to enhanced cell apoptosis. Benefitting from its relative long circulation time in the blood and homotypic targeting effect, the tumor accumulation of GNR@Mem significantly increased. The in vivo results demonstrate that the combination of photothermal therapy and radiotherapy effectively suppresses tumor growth without noticeable systemic toxicity.

Users exhibit different intents across e-commerce services (e.g. discovering items, purchasing gifts, etc.) which drives them to interact with a wide variety of items in multiple ways (e.g. click, add-to-cart, add-to-favorites, purchase). To give better recommendations, it is important to capture user intent, in addition to considering their historic interactions. However these intents are by definition latent, as we observe only a user’s interactions, and not their underlying intent. To discover such latent intents, and use them effectively for recommendation, in this paper we propose an Attentive Sequential model of Latent Intent (ASLI in short). Our model first learns item similarities from users’ interaction histories via a self-attention layer, then uses a Temporal Convolutional Network layer to obtain a latent representation of the user’s intent from her actions on a particular category. We use this representation to guide an attentive model to predict the next item. Results from our experiments show that our model can capture the dynamics of user behavior and preferences, leading to state-of-the-art performance across datasets from two major e-commerce platforms, namely Etsy and Alibaba.

For large-scale online marketplaces with over millions of items, users come to rely on personalized recommendations to find relevant items from their massive inventory. One hallmark of the shopping experience in such online marketplaces is the many ways a user can interact with an item they are interested in: they can click it, favorite it, add it to cart, purchase it, etc. We hypothesize that the different ways in which a user interacts with an item indicates different kinds of intent. Consequently, a user’s recommendations should be based not only on items from their past activity, but also the way in which they interacted with these items. Co-occurrence based methods have been successfully used to give recommendations that incorporate interaction types, such as the popular “Because you purchased X, you may also purchase Y”. In this paper, we propose a novel method that generalizes upon the co-occurrence methods to learn interaction-based item embeddings that encode the co-occurrence patterns of not only the item itself, but also the interaction type. The learned embeddings provide a convenient way of approximating the likelihood that one item-interaction pair would co-occur with another by way of a simple inner product. To show their effectiveness, we deploy the interaction-based embeddings in an industry-scale recommendation system that serves live traffic on Etsy.com. We find that taking interaction types into account shows significant improvements in accurately modeling user shopping behavior for both online and offline metrics.

Currently, most sequence-based recommendation models aim to predict a user's next actions (e.g. next purchase) based on their past actions. These models either capture users' intrinsic preference (e.g. a comedy lover, or a fan of fantasy) from their long-term behavior patterns or infer their current needs by emphasizing recent actions. However, in e-commerce, intrinsic user behavior may be shifted by occasions such as birthdays, anniversaries, or gifting celebrations (Valentine's Day or Mother's Day), leading to purchases that deviate from long-term preferences and are not related to recent actions. In this work, we propose a novel next-item recommendation system which models a user's default, intrinsic preference, as well as two different kinds of occasion-based signals that may cause users to deviate from their normal behavior. More specifically, this model is novel in that it: (1) captures a personal occasion signal using an attention layer that models reoccurring occasions specific to that user (e.g. a birthday); (2) captures a global occasion signal using an attention layer that models seasonal or reoccurring occasions for many users (e.g. Christmas); (3) balances the user's intrinsic preferences with the personal and global occasion signals for different users at different timestamps with a gating layer. We explore two real-world e-commerce datasets (Amazon and Etsy) and show that the proposed model outperforms state-of-the-art models by 7.62% and 6.06% in predicting users' next purchase.

Regeneration and functional recovery of peripheral nerves remain formidable due to the inefficient physical and chemical cues in the available nerve guidance conduits (NGCs). Introducing micropatterns and bioactive substances into the inner wall of NGCs can effectively regulate the behavior of Schwann cells, the elongation of axons, and the phenotype of macrophages, thereby aiding the regeneration of injured nerve. In this study, linear micropatterns with ridges and grooves of 3/3, 5/5, 10/10, and 30/30 μm were created on poly(d,l-lactide

To promote the clinical theranostic performances of platinum-based anticancer drugs, imaging capability is urgently desired, and their chemotherapeutic efficacy needs to be upgraded. Herein, a theranostic metallacycle (M) is developed for imaging-guided cancer radio-chemotherapy using perylene bisimide fluorophore (PPy) and tetraphenylethylene-based di-Pt(II) organometallic precursor (TPE-Pt) as building blocks. The formation of this discrete supramolecular coordination complex facilitates the encapsulation of M by a glutathione (GSH)-responsive amphiphilic block copolymer to prepare M-loaded nanoparticles (MNPs). TPE-Pt acts as a chemotherapeutic drug and also an excellent radiosensitizer, thus incorporating radiotherapy into the nanomedicine to accelerate the therapeutic efficacy and overcome drug resistance. The NIR-emission of PPy is employed to detect the intracellular delivery and tissue distribution of MNPs in real time. In vitro and in vivo investigations demonstrate the excellent anticancer efficacy combining chemotherapy and radiotherapy; the administration of this nanomedicine effectively inhibits the tumor growth and greatly extends the survival rate of cisplatin-resistant A2780CIS-tumor-bearing mice. Guided by in vivo fluorescence imaging, radio-chemotherapy is precisely carried out, which facilitates boosting of the therapeutic outcomes and minimizing undesired side effects. The success of this theranostic system brings new hope to supramolecular nanomedicines for their potential clinical translations.

Emergent and long-term hemorrhage control is requisite and beneficial for reducing global mortality and postoperative complications (e.g., second bleeding and adverse tissue adhesion). Despite recent advance in injectable hydrogels for hemostasis, achieving rapid gelation, strong tissue-adhesive property and stable mechanical strength under fluid physiological environment is still challenging. Herein, we developed a novel chitosan hydrogel (CCS@gel) via dynamic Schiff base reaction and mussel-inspired catechol chemistry. The hydrogel possessed high gelation rate (10 s), strong wet adhesiveness, excellent self-healing performance and biocompatibility. More importantly, the CCS@gel exhibited saline-induced contractile performance and mechanical enhancement, promoting its mechanical property in moist internal conditions. In vivo studies demonstrated its superior hemostatic efficacy for diverse anticoagulated visceral and carotid bleeding scenarios, compared to commercialized fibrin glue. The hydrogel-treated rats survived for 8 weeks with minimal inflammation and postoperative adhesion. These results revealed that the promising CCS@gel would be a facile, efficient and safe sealant for clinical hemorrhage control.

Abstract Background Worldwide, hepatocellular carcinoma (HCC) is a common solid tumor with a poor prognosis. HCC is often due to hepatitis B virus (HBV) infection. As yet, efficacious HCC treatment regimens for late-stage HCC patients are lacking. Therefore, the identification of more specific and sensitive biomarkers for its early diagnosis and treatment remains an urgent need. Methods Total RNAs from paired HBV-derived HCC tumors and adjacent peritumor tissues (APTs) were subjected to RNA sequencing (RNA-seq), and differentially expressed genes (DEGs) between HCC tumors and APTs were selected and verified. Results We identified 166 DEGs and found that eight top-ranked and verified DEGs (TK1, CTTN, CEP72, TRIP13, FTH1, FLAD1, CHRM2, AMBP) all contained putative OCT4 binding motifs in their promoter regions. TK1, TRIP13 and OCT4 were found to exhibit concurrent higher expression levels in HCC tumors than in APTs. The mRNA levels of TK1, TRIP13 and OCT4 in a cohort of 384 HCC samples from the TCGA database were all found to be negatively correlated with patient overall survival, relapse-free survival and progression-free survival, underscoring the HCC biomarker status of TK1 and TRIP13 on one hand, and implicating their association with OCT4 on the other hand. Furthermore, OCT4 proteins were found to bind to the promoters of both genes in vitro and in vivo. Knocking out OCT4 in HCC-derived cell lines reduced the expression of TK1 and TRIP13 and significantly decreased their tumorigenicity. Conclusions Using RNA-seq, we identified several novel HCC signature genes that may serve as biomarkers for its diagnosis and prognosis. Their common transcriptional regulation by OCT4 suggests key roles in the development of HCC, and indicates that OCT4 may serve as a potential therapeutic target.

Hepatocellular carcinoma (HCC) is the third leading cause of cancer-related death worldwide. The prognosis of HCC remains very poor; thus, an effective treatment remains urgent. Herein, a type of nanomedicine is developed by conjugating Fe@Fe3 O4 nanoparticles with ginsenoside Rg3 (NpRg3), which achieves an excellent coupling effect. In the dimethylnitrosamine-induced HCC model, NpRg3 application significantly prolongs the survival of HCC mice. Further research indicates that NpRg3 application significantly inhibits HCC development and eliminates HCC metastasis to the lung. Notably, NpRg3 application delays HCC-induced ileocecal morphology and gut microbial alterations more than 12 weeks during HCC progression. NpRg3 administration elevates the abundance of Bacteroidetes and Verrucomicrobia, but decreases Firmicutes. Twenty-nine predicted microbial gene functions are enriched, while seven gene functions are reduced after NpRg3 administration. Moreover, the metabolomics profile presents a significant progression during HCC development, but NpRg3 administration corrects tumor-dominant metabolomics. NpRg3 administration decreases 3-indolepropionic acid and urea, but elevates free fatty acids. Importantly, NpRg3 application remodels the unbalanced correlation networks between gut microbiota and metabolism during HCC therapy. In conclusion, nanoparticle conjugation of ginsenoside Rg3 inhibits HCC development and metastasis via the remodeling of unbalanced gut microbiota and metabolism in vivo, providing an antitumor therapy strategy.

Reactive oxygen species (ROS) play an important role in the pathogenesis of numerous diseases including atherosclerosis, diabetes, inflammation and myocardial infarction (MI). In this study, a ROS-responsive biodegradable elastomeric polyurethane containing thioketal (PUTK) linkages was synthesized from polycaprolactone diol (PCL-diol ), 1,6-hexamethylene diisocyanate (HDI), and ROS-cleavable chain extender. The PUTK was electrospun into fibrous patches with the option to load glucocorticoid methylprednisolone (MP), which were then used to treat MI of rats in vivo. The fibrous patches exhibited suitable mechanical properties and high elasticity. The molecular weight of PUTK was decreased significantly after incubation in 1 mM H2O2 solution for 2 weeks due to the degradation of thioketal bonds on the polymer backbone. Both the PUTK and PUTK/MP fibrous patches showed good antioxidant property in an oxidative environment in vitro. Implantation of the ROS-responsive polyurethane patches in MI of rats in vivo could better protect cardiomyocytes from death in the earlier stage (24 h) than the non ROS-responsive ones. Implantation of the PUTK/MP fibrous patches for 28 days could effectively improve the reconstruction of cardiac functions including increased ejection fraction, decreased infarction size, and enhanced revascularization of the infarct myocardium.

Aesthetic style is the crux of many purchasing decisions. When considering an item for purchase, buyers need to be aligned not only with the functional aspects (e.g. description, category, ratings) of an item's specification, but also its stylistic and aesthetic aspects (e.g. modern, classical, retro) as well. Style becomes increasingly important on e-commerce sites like Etsy, an online marketplace for handmade and vintage goods, where hundreds of thousands of items can differ by style and aesthetic alone. As such, it is important for industry recommender systems to properly model style when understanding shoppers' buying preference. In past work, because of its abstract nature, style is often approached in an unsupervised manner, represented by nameless latent factors or embeddings. As a result, there has been no previous work on predictive models nor analysis devoted to understanding how style, or even the presence of style, impacts a buyer's purchase decision. In this paper, we discuss a novel process by which we leverage 43 named styles given by merchandising experts in order to bootstrap large-scale style prediction and analysis of how style impacts purchase decision. We train a supervised, style-aware deep neural network that is shown to predict item style with high accuracy, while generating style-aware embeddings that can be used in downstream recommendation tasks. We share in our analysis, based on over a year's worth of transaction data and show that these findings are crucial to understanding how to more explicitly leverage style signal in industry-scale recommender systems.

User engagement plays a central role in companies operating online services, such as search engines, news portals, e-commerce sites, entertainment services, and social networks. A main challenge is to leverage collected knowledge about the daily online behavior of millions of users to understand what engage them short-term and more importantly long-term. Two critical steps of improving user engagement are metrics and their optimization. The most common way that engagement is measured is through various online metrics, acting as proxy measures of user engagement. This tutorial will review these metrics, their advantages and drawbacks, and their appropriateness to various types of online services. Once metrics are defined, how to optimize them will become the key issue. We will survey methodologies including machine learning models and experimental designs that are utilized to optimize these metrics via direct or indirect ways. As case studies, we will focus on four types of services, news, search, entertainment, and e-commerce. We will end with lessons learned and a discussion on the most promising research directions.

Online A/B tests play an instrumental role for Internet companies to improve products and technologies in a data-driven manner. An online A/B test, in its most straightforward form, can be treated as a static hypothesis test where traditional statistical tools such as p-values and power analysis might be applied to help decision makers determine which variant performs better. However, a static A/B test presents both time cost and the opportunity cost for rapid product iterations. For time cost, a fast-paced product evolution pushes its shareholders to consistently monitor results from online A/B experiments, which usually invites peeking and altering experimental designs as data collected. It is recognized that this flexibility might harm statistical guarantees if not introduced in the right way, especially when online tests are considered as static hypothesis tests. For opportunity cost, a static test usually entails a static allocation of users into different variants, which prevents an immediate roll-out of the better version to larger audience or risks of alienating users who may suffer from a bad experience. While some works try to tackle these challenges, no prior method focuses on a holistic solution to both issues. In this paper, we propose a unified framework utilizing sequential analysis and multi-armed bandit to address time cost and the opportunity cost of static online tests simultaneously. In particular, we present an imputed sequential Girshick test that accommodates online data and dynamic allocation of data. The unobserved potential outcomes are treated as missing data and are imputed using empirical averages. Focusing on the binomial model, we demonstrate that the proposed imputed Girshick test achieves Type-I error and power control with both a fixed allocation ratio and an adaptive allocation such as Thompson Sampling through extensive experiments. In addition, we also run experiments on historical Etsy.com A/B tests to show the reduction in opportunity cost when using the proposed method.

Matching between two information objects is the core of many different information retrieval (IR) applications including Web search, question answering, and recommendation. Recently, deep learning methods have yielded immense success in speech recognition, computer vision, and natural language processing, significantly advancing state-of-the-art of these areas. In the IR community, deep learning has also attracted much attention, and researchers have proposed a large number of deep matching models to tackle the matching problem for different IR applications. Despite the fact that deep matching models have gained significant progress in these areas, there are still many challenges to be addressed when applying these models to real IR scenarios. In this workshop, we focus on the applicability of deep matching models to practical applications. We aim to discuss the issues of applying deep matching models to production systems, as well as to shed some light on the fundamental characteristics of different matching tasks in IR. website : https://wsdm2019-dapa.github.io/index.html

Background: Hepatocellular carcinoma (HCC) is a common solid tumor worldwide with a very poor prognosis. Identifying more specific and sensitive HCC biomarkers remains an urgent need for its early diagnosis and treatment. Methods: Total RNAs from paired HBV-derived HCC tumors and adjacent peritumor tissues (APTs) were amplified and subjected to RNA sequencing analysis, and differentially expressed genes (DEGs) between tumors and APTs were selected and verified. Findings: 166 DEGs were identified, and remarkably, eight top-ranked and verified DEGs (TK1, CTTN, CEP72, TRIP13, FTH1, FLAD1, CHRM2, AMBP) all contained putative OCT4 binding motifs at their promoter regions. TK1, TRIP13 and OCT4 were exemplified to have concurrent higher expression in HCC tumors than in APTs, and OCT4 proteins were bound to the promoters of both genes in vitro and in vivo. Knocking out OCT4 gene in hepatoma cell lines reduced the expression of TK1 and TRIP13 and significantly dampened their tumorigenicity. Interpretation: We identified multiple novel HCC signature genes that may serve as promising biomarkers for HCC diagnosis and prognosis. Their common transcriptional regulation by OCT4 implicates its key roles in the occurrence and development of HCC, and thus positions OCT4 as a potential therapeutic target for treating HCC. Funding Statement: This work was supported by grants from the National Key Research and Development Program of China (grant numbers 2016YFA0101201, 2016YFA0100303), the National Natural Science Foundation of China (grant number 31601103), the National High Technology Research and Development Program of China (grant number 2012AA020204) and the Fundamental Research Funds for the Central Universities (grant number WK9110000023). Declaration of Interests: The authors declare that they have no competing interests. Ethics Approval Statement: All studies on mice were conducted in accordance with the National Institute Guide for the Care and Use of Laboratory Animal. The animal protocol has been approved by the Committee of the Ethics of Animal Experiments, Zhejiang University.

The excessive reactive oxygen species (ROS) and hypoxia deteriorate the inflammation-related diseases such as myocardial infarction (MI), and thereby deter the normal tissue repair and recovery and further lead to severe fibrosis and malfunction of tissues and organs. In particular, the MI has become one of the leading causes of death nowadays. In this study, a novel type of injectable hydrogel with dual functions of ROS scavenging and O2 generating is fabricated for MI treatment in vivo. The hydrogel is formed within 3 s from the synthetic ROS-cleavable hyperbranched polymers and methacrylate hyaluronic acid (HA-MA) under UV-irradiation. Addition of biocompatible and applicable catalase in vivo enables the further transition of H2 O2 , a major type of ROS, to O2 and H2 O. Results of rat MI model demonstrate that this hydrogel can significantly remove excessive ROS, inhibit cell apoptosis, increase M2/M1 macrophage ratio, promote angiogenesis, reduce infarcted area, and improve cardiac functions. With the appropriate degradation rate, simple structure and composition without cell seeding, and very excellent MI therapeutic effect, this ROS scavenging and O2 generating hydrogel has a great promise to be applied clinically.

Building price- and budget-aware recommender systems is critical in settings where one wishes to produce recommendations that balance users' preferences (what they like) with a model of purchase likelihood (what they will buy). A trivial solution consists of learning global budget terms for each user based on their past expenditure. To more accurately model user budgets, we also consider a user's within-session budget, which may deviate from their global budget depending on their shopping context. In this paper, we find that users implicitly reveal their session-specific budgets through the sequence of items they browse within that session. Specifically, we find that some users "browse down," by purchasing the cheapest item among alternatives under consideration, others "browse up" (selecting the most expensive), and others ultimately purchase items around the middle. Surprisingly, this mixture of behaviors is difficult to observe globally, as individual users tend to belong firmly to one of the three segments. To model this behavior, we develop an interpretable budget model that combines a clustering component to detect different user segments, with a model of segment-specific purchase profiles. We apply our model on a dataset of browsing and purchasing sessions from Etsy, a large e-commerce website focused on handmade and vintage goods, where it outperforms strong baselines and existing production systems.

Cell division cycle associated 5 (CDCA5) is an important element for the interaction between cohesin and chromatin in interphase. It is abnormally expressed in many types of cancer and works as an indicator of poor prognosis, but little is known about its activity in hepatocellular carcinoma (HCC). In the present study, we found that the expression of CDCA5 was upregulated in HCC tissues compared to paracancerous tissues and had a negative correlation with patient survival. Cell proliferation and tumorigenesis were inhibited and cell apoptosis was induced with the knockdown of CDCA5, suggesting an oncogenic role of CDCA5 in liver cancer. Luciferase reporter assay and chromatin immunoprecipitation showed that CDCA5 was transcribed by E2F1. Furthermore, we confirmed that CDCA5 interrupted cell behavior via the AKT pathway. These findings demonstrated that CDCA5 plays an important role in HCC progression.

In recent years, product search engines have emerged as a key factor for online businesses. According to a recent survey, over 55% of online customers begin their online shopping journey by searching on an E-Commerce (EC) website like Amazon as opposed to a generic web search engine like Google. Information retrieval research to date has been focused on optimizing search ranking algorithms for web documents while little attention has been paid to product search. There are several intrinsic differences between web search and product search that make the direct application of traditional search ranking algorithms to EC search platforms difficult. First, the success of web and product search is measured differently; one seeks to optimize for relevance while the other must optimize for both relevance and revenue. Second, when using real-world EC transaction data, there is no access to manually annotated labels. In this paper, we address these differences with a novel learning framework for EC product search called LETORIF (LEarning TO Rank with Implicit Feedback). In this framework, we utilize implicit user feedback signals (such as user clicks and purchases) and jointly model the different stages of the shopping journey to optimize for EC sales revenue. We conduct experiments on real-world EC transaction data and introduce a a new evaluation metric to estimate expected revenue after re-ranking. Experimental results show that LETORIF outperforms top competitors in improving purchase rates and total revenue earned.

User engagement plays a central role in companies operating online services, such as search engines, news portals, e-commerce sites, and social networks. A main challenge is to leverage collected knowledge about the daily online behavior of millions of users to understand what engage them short-term and more importantly long-term. The most common way that engagement is measured is through various online metrics, acting as proxy measures of user engagement. This tutorial will review these metrics, their advantages and drawbacks, and their appropriateness to various types of online services. As case studies, we will focus on three types of services, news, search and e-commerce. We will also briefly discuss how to develop better machine learning models to optimize online metrics, and design experiments to test these models.

In this work, we propose to improve long-term user engagement in a recommender system from the perspective of sequential decision optimization, where users' click and return behaviors are directly modeled for online optimization. A bandit-based solution is formulated to balance three competing factors during online learning, including exploitation for immediate click, exploitation for expected future clicks, and exploration of unknowns for model estimation. We rigorously prove that with a high probability our proposed solution achieves a sublinear upper regret bound in maximizing cumulative clicks from a population of users in a given period of time, while a linear regret is inevitable if a user's temporal return behavior is not considered when making the recommendations. Extensive experimentation on both simulations and a large-scale real-world dataset collected from Yahoo frontpage news recommendation log verified the effectiveness and significant improvement of our proposed algorithm compared with several state-of-the-art online learning baselines for recommendation.

Recommending personalized content to users is a long-standing challenge to many online services including Facebook, Yahoo, Linkedin and Twitter. Traditional recommendation models such as latent factor models and feature-based models are usually trained for all users and optimize an "average" experience for them, yielding sub-optimal solutions. Although multi-task learning provides an opportunity to learn personalized models per user, learning algorithms are usually tailored to specific models (e.g., generalized linear model, matrix factorization and etc.), creating obstacles for a unified engineering interface, which is important for large Internet companies. In this paper, we present an empirical framework to learn user-specific personal models for content recommendation by utilizing gradient information from a global model. Our proposed method can potentially benefit any model that can be optimized through gradients, offering a lightweight yet generic alternative to conventional multi-task learning algorithms for user personalization. We demonstrate the effectiveness of the proposed framework by incorporating it in three popular machine learning algorithms including logistic regression, gradient boosting decision tree and matrix factorization. Our extensive empirical evaluation shows that the proposed framework can significantly improve the efficiency of personalized recommendation in real-world datasets.

Latent factor models and decision tree based models are widely used in tasks of prediction, ranking and recommendation. Latent factor models have the advantage of interpreting categorical features by a low-dimensional representation, while such an interpretation does not naturally fit numerical features. In contrast, decision tree based models enjoy the advantage of capturing the nonlinear interactions of numerical features, while their capability of handling categorical features is limited by the cardinality of those features. Since in real-world applications we usually have both abundant numerical features and categorical features with large cardinality (e.g. geolocations, IDs, tags etc.), we design a new model, called GB-CENT, which leverages latent factor embedding and tree components to achieve the merits of both while avoiding their demerits. With two real-world data sets, we demonstrate that GB-CENT can effectively (i.e. fast and accurately) achieve better accuracy than state-of-the-art matrix factorization, decision tree based models and their ensemble.

Etsy1 is a global marketplace where people across the world connect to make, buy and sell unique goods. Sellers at Etsy can promote their product listings via advertising campaigns similar to traditional sponsored search ads. Click-Through Rate (CTR) prediction is an integral part of online search advertising systems where it is utilized as an input to auctions which determine the final ranking of promoted listings to a particular user for each query. In this paper, we provide a holistic view of Etsy's promoted listings' CTR prediction system and propose an ensemble learning approach which is based on historical or behavioral signals for older listings as well as content-based features for new listings. We obtain representations from texts and images by utilizing state-of-the-art deep learning techniques and employ multimodal learning to combine these different signals. We compare the system to non-trivial baselines on a large-scale real world dataset from Etsy, demonstrating the effectiveness of the model and strong correlations between offline experiments and online performance. The paper is also the first technical overview to this kind of product in e-commerce context.

Background Biliary tract obstruction is a common clinical problem. In this study, we attempted to understand the change in intestinal glucose absorption after biliary tract obstruction. Methods Experimental models of murine biliary duct ligation and external biliary drainage were established. Murine intestinal mucosal glucose absorption was examined with Ussing chambers according to the increase in the short-circuit current in vitro and blood glucose measurement after oral glucose in vivo. The protein expression of the sodium-glucose cotransporter (SGLT1) and the facilitated glucose transporter, member 2 (GLUT2) was analyzed by Western blot and immunohistochemistry. Results The results from Ussing chamber experiments showed that duodenal mucosal glucose absorption levels were significantly higher in biliary duct ligation and biliary drainage mice than those in normal control mice at 1 and 2 weeks after the operation. Gastrointestinal bile acid administration almost reversed the elevated duodenal mucosal glucose absorption to the normal level in biliary drainage mice. The results from the experiments in vivo further confirmed that the glucose absorption increased in biliary duct ligation and biliary drainage mice. The protein expression levels of SGLT1 in the duodenal mucosae of both biliary duct ligation and biliary drainage mice were markedly higher than those in control mice, and the protein expression of GLUT2 was not significantly altered, compared with control mice. Conclusion Bile deficiency in the intestine upregulates the expression of intestinal mucosal SGLT1 and enhances intestinal mucosal glucose absorption capacity, which contributes to the understanding of intestinal physiologic function for patients with biliary duct obstruction and external biliary drainage.

Feature-based collaborative filtering models, such as state-of-the-art factorization machines and regression-based latent factor models, rarely consider features' structural information, ignoring the heterogeneity of inter-type and intra-type relationships. Naively treating all feature pairs equally would potentially deteriorate the overall recommendation performance. In addition, human prior knowledge and other hierarchical or graphical structures are often available for some features, e.g., the country-state-city hierarchy for geographic features and the topical taxonomy for article features. It is a challenge to utilize the prior knowledge to further boost performance of state-of-the-art models. In this paper we employ rich features from both user and item sides to enhance latent factors learnt from interaction data, uncovering hidden structures from features' relationships and learning sparse pairwise and tree structural connections among features. Our framework borrows the modeling strengh from both structural sparsity modeling and latent factor models. Experiments on a real-world large-scale recommendation data set demonstrated that the proposed model outperforms several strong state-of-the-art baselines.

Many internet companies, such as Yahoo, Facebook, Google and Twitter, rely on content recommendation systems to deliver the most relevant content items to individual users through personalization. Delivering such personalized user experiences is believed to increase the long term engagement of users. While there has been a lot of progress in designing effective personalized recommender systems, by exploiting user interests and historical interaction data through implicit (item click) or explicit (item rating) feedback, directly optimizing for users' satisfaction with the system remains challenging. In this paper, we explore the idea of using item-level dwell time as a proxy to quantify how likely a content item is relevant to a particular user. We describe a novel method to compute accurate dwell time based on client-side and server-side logging and demonstrate how to normalize dwell time across different devices and contexts. In addition, we describe our experiments in incorporating dwell time into state-of-the-art learning to rank techniques and collaborative filtering models that obtain competitive performances in both offline and online settings.

We propose a novel probabilistic topic model that jointly models authors, documents, cited authors, and venues simultaneously in one integrated framework, as compared to previous work which embeds fewer components. This model is designed for three typical applications in academic network analysis: the problems of expert ranking, cited author prediction and venue prediction. Experiments based on two real world data sets demonstrate the model to be effective, and it outperforms several state-of-the-art algorithms in all three applications.

With the availability of cheap location sensors, geotagging of messages in online social networks is proliferating. For instance, Twitter, Facebook, Foursquare, and Google+ provide these services both explicitly by letting users choose their location or implicitly via a sensor. This paper presents an integrated generative model of location and message content. That is, we provide a model for combining distributions over locations, topics, and over user characteristics, both in terms of location and in terms of their content preferences. Unlike previous work which modeled data in a flat pre-defined representation, our model automatically infers both the hierarchical structure over content and over the size and position of geographical locations. This affords significantly higher accuracy --- location uncertainty is reduced by 40% relative to the best previous results [21] achieved on location estimation from Tweets. We achieve this goal by proposing a new statistical model, the nested Chinese Restaurant Franchise (nCRF), a hierarchical model of tree distributions. Much statistical structure is shared between users. That said, each user has his own distribution over interests and places. The use of the nCRF allows us to capture the following effects: (1) We provide a topic model for Tweets; (2) We obtain location specific topics; (3) We infer a latent distribution of locations; (4) We provide a joint hierarchical model of topics and locations; (5) We infer personalized preferences over topics and locations within the above model. In doing so, we are both able to obtain accurate estimates of the location of a user based on his tweets and to obtain a detailed estimate of a geographical language model.

Users of popular services like Twitter and Facebook are often simultaneously overwhelmed with the amount of information delivered via their social connections and miss out on much content that they might have liked to see, even though it was distributed outside of their social circle. Both issues serve as difficulties to the users and drawbacks to the services.Social media service providers can benefit from understanding user interests and how they interact with the service, potentially predicting their behaviors in the future. In this paper, we address the problem of simultaneously predicting user decisions and modeling users' interests in social media by analyzing rich information gathered from Twitter. The task differs from conventional recommender systems as the cold-start problem is ubiquitous, and rich features, including textual content, need to be considered. We build predictive models for user decisions in Twitter by proposing Co-Factorization Machines (CoFM), an extension of a state-of-the-art recommendation model, to handle multiple aspects of the dataset at the same time. Additionally, we discuss and compare ranking-based loss functions in the context of recommender systems, providing the first view of how they vary from each other and perform in real tasks. We explore an extensive set of features and conduct experiments on a real-world dataset, concluding that CoFM with ranking-based loss functions is superior to state-of-the-art methods and yields interpretable latent factors.