No abstract

It is often desirable to model multiple objectives in real-world web applications, such as user satisfaction and user engagement in recommender systems. Multi-task learning has become the standard approach for such applications recently. While most of the multi-task recommendation model architectures proposed to date are focusing on using non-sequential input features (e.g., query and context), input data is often sequential in real-world web application scenarios. For example, user behavior streams, such as user search logs in search systems, are naturally a temporal sequence. Modeling user sequential behaviors as explicit sequential representations can empower the multi-task model to incorporate temporal dependencies, thus predicting future user behavior more accurately. Furthermore, user activity streams can come from heterogeneous data sources, such as user search logs and user browsing logs. They typically possess very different properties such as data sparsity and thus need careful treatment when being modeled jointly. In this work, we study the challenging problem of how to model sequential user behavior in the neural multi-task learning settings. Our major contribution is a novel framework, Mixture of Sequential Experts (MoSE). It explicitly models sequential user behavior using Long Short-Term Memory (LSTM) in the state-of-art Multi-gate Mixture-of-Expert multi-task modeling framework. In experiments, we show the effectiveness of the MoSE architecture over seven alternative architectures on both synthetic and noisy real-world user data in G Suite. We also demonstrate the effectiveness and flexibility of the MoSE architecture in a real-world decision making engine in GMail that involves millions of users, balancing between search quality and resource costs.

The immune system initiates robust immune responses to defend against invading pathogens or tumor cells and protect the body from damage, thus acting as a fortress of the body. However, excessive responses cause detrimental effects, such as inflammation and autoimmune diseases. To balance the immune responses and maintain immune homeostasis, there are immune checkpoints to terminate overwhelmed immune responses. Pathogens and tumor cells can also exploit immune checkpoint pathways to suppress immune responses, thus escaping immune surveillance. As a consequence, therapeutic antibodies that target immune checkpoints have made great breakthroughs, in particular for cancer treatment. While the overall efficacy of immune checkpoint blockade (ICB) is unsatisfactory since only a small group of patients benefited from ICB treatment. Hence, there is a strong need to search for other targets that improve the efficacy of ICB. Ubiquitination is a highly conserved process which participates in numerous biological activities, including innate and adaptive immunity. A growing body of evidence emphasizes the importance of ubiquitination and its reverse process, deubiquitination, on the regulation of immune responses, providing the rational of simultaneous targeting of immune checkpoints and ubiquitination/deubiquitination pathways to enhance the therapeutic efficacy. Our review will summarize the latest findings of ubiquitination/deubiquitination pathways for anti-tumor immunity, and discuss therapeutic significance of targeting ubiquitination/deubiquitination pathways in the future of immunotherapy.

Ti-24Nb-4Zr-7.9Sn (TNZS) alloy is a newly developed β-titanium alloy considered suitable for dental implant applications due to its low elastic modulus and high strength. The aim of this study was to investigate the corrosion behavior of TNZS alloy through a static immersion test in various simulated physiological solutions, namely, artificial saliva, lactic acid solution, fluoridated saliva, and fluoridated acidified saliva for 7 days. The corrosion behavior of commercially pure titanium and Ti-6Al-4V alloy were also examined for comparison. The elemental release was measured with inductively coupled plasma mass spectroscopy, and the changes of alloy surface were observed with scanning electron microscopy (SEM). The test results showed that the quantity of each metal element released from TNZS alloy into fluoridated solutions was much higher than the solutions without fluoride ions. It was highest in fluoridated acidified saliva and lowest in artificial saliva (p < 0.01). The total elemental release from TNZS alloy was lower than commercially pure titanium and Ti-6Al-4V alloy in the same solution (p < 0.01). SEM micrographs indicated that TNZS alloy possessed better corrosion resistant performance. It can be concluded that fluoridated solutions have a negative influence on the corrosion behavior of TNZS alloy. Compared with commercially pure titanium and Ti-6Al-4V alloy, TNZS alloy demonstrates better corrosion resistance in various simulated physiological solutions, so it has greater potential for dental implant applications.