Collaborative Metric Learning (CML) has recently emerged as a popular method in recommendation systems (RS), closing the gap between metric learning and Collaborative Filtering. Following the convention of RS, existin...
ISBN:
(纸本)9781713871088
Collaborative Metric Learning (CML) has recently emerged as a popular method in recommendation systems (RS), closing the gap between metric learning and Collaborative Filtering. Following the convention of RS, existing methods exploit unique user representation in their model design. This paper focuses on a challenging scenario where a user has multiple categories of interests. Under this setting, we argue that the unique user representation might induce preference bias, especially when the item category distribution is imbalanced. To address this issue, we propose a novel method called Diversity-Promoting Collaborative Metric Learning (DPCML), with the hope of considering the commonly ignored minority interest of the user. The key idea behind DPCML is to include a multiple set of representations for each user in the system. Based on this embedding paradigm, user preference toward an item is aggregated from different embeddings by taking the minimum item-user distance among the user embedding set. Furthermore, we observe that the diversity of the embeddings for the same user also plays an essential role in the model. To this end, we propose a Diversity Control Regularization Scheme (DCRS) to accommodate the multi-vector representation strategy better. Theoretically, we show that DPCML could generalize well to unseen test data by tackling the challenge of the annoying operation that comes from the minimum value. Experiments over a range of benchmark datasets speak to the efficacy of DPCML.
Viewport prediction is a crucial aspect of tile-based 360◦ video streaming system. However, existing trajectory based methods lack of robustness, also oversimplify the process of information construction and fusion be...
详细信息
This paper proposes a configurable Sigmoid and Tanh activation function circuit design by second-order approximation and deviation compensation. The proposed circuit is synthesized in TSMC 180nm technology within a 72...
详细信息
The automatic generation of music comments is of great significance for increasing the popularity of music and the music platform’s activity. In human music comments, there exists high distinction and diverse perspec...
详细信息
Neutron resonance imaging (NRI) has recently emerged as an appealing technique for neutron radiography. Its complexity surpasses that of conventional transmission imaging, as it requires a high demand for both a neutr...
Neutron resonance imaging (NRI) has recently emerged as an appealing technique for neutron radiography. Its complexity surpasses that of conventional transmission imaging, as it requires a high demand for both a neutron source and detector. Consequently, the progression of NRI technology has been sluggish since its inception in the 1980s, particularly considering the limited studies analyzing the neutron energy range above keV. The white neutron source (Back-n) at the China Spallation Neutron Source (CSNS) provides favorable beam conditions for the development of the NRI technique over a wide neutron energy range from eV to MeV. Neutron-sensitive microchannel plates (MCP) have emerged as a cutting-edge tool in the field of neutron detection owing to their high temporal and spatial resolutions, high detection efficiency, and low noise. In this study, we report the development of a $$^{10}$$ B-doped MCP detector, along with its associated electronics, data processing system, and NRI experiments at the Back-n. Individual heavy elements such as gold, silver, tungsten, and indium can be easily identified in the transmission images by their characteristic resonance peaks in the 1–100 eV energy range; the more difficult medium-weight elements such as iron, copper, and aluminum with resonance peaks in the 1–100 keV energy range can also be identified. In particular, results in the neutron energy range of dozens of keV (Aluminum) are reported here for the first time.
Besides entity-centric knowledge, usually organized as Knowledge Graph (KG), events are also an essential kind of knowledge in the world, which trigger the spring up of event-centric knowledge representation form like...
详细信息
In this paper, we investigate the recommendation task in the most common scenario with implicit feedback (e.g., clicks, purchases). State-of-the-art methods in this direction usually cast the problem as to learn a per...
详细信息
Password generation model based on generative adversarial network usually has the problem of high duplicate rate, which further leads to low cover rate. In this regard, we propose PGGAN model. It sets up an additional...
详细信息
The neutron capture cross section of $$^{232}Th$$ has been measured with the time-of-flight technique in the energy range from 10 to 200 keV at the back-streaming white neutron beam-line (Back-n) of China Spallation N...
The neutron capture cross section of $$^{232}Th$$ has been measured with the time-of-flight technique in the energy range from 10 to 200 keV at the back-streaming white neutron beam-line (Back-n) of China Spallation Neutron Source (CSNS). The pulse height weighting technique (PHWT) was applied with four C $$_{6}$$ D $$_{6}$$ liquid scintillators to measure the prompt gamma-ray energy release following neutron capture. The measurement data, corrected with the PHWT, have been normalized to the saturated resonances at 21.8 eV. The background was determined by a lead sample measurement and detailed Monte Carlo simulations. The $$^{232}Th(n,\gamma )$$ average cross sections have been determined relative to the $$^{197}Au(n,\gamma )$$ reaction cross sections. The results are consistent with the evaluation values of CENDL-3.2 and JENDL-5. The total uncertainties, including the PHWT, normalization, background subtraction, corrections, and relative measurement, are in the range of 4.5–4.8%.
The speech denoising model based on adversarial generative network has achieved better results than the traditional machine learning model. In this paper, for the short cut connection in the generator, we discuss its ...
详细信息
暂无评论