Watch time prediction (WTP) has emerged as a pivotal task in short video recommendation systems, designed to quantify user engagement through continuous interaction modeling. Predicting users' watch times on videos often encounters fundamental challenges, including wide value ranges and imbalanced data distributions, which can lead to significant estimation bias when directly applying regression techniques. Recent studies have attempted to address these issues by converting the continuous watch time estimation into an ordinal regression task. While these methods demonstrate partial effectiveness, they exhibit notable limitations: (1) The discretization process frequently relies on bucket partitioning, inherently reducing prediction flexibility and accuracy. (2) The interdependencies among different partition intervals remain underutilized, missing opportunities for effective error correction.

With an exponentially growing number of graphs from disparate repositories, there is a strong need to analyze a graph database containing an extensive collection of small- or medium-sized data graphs (e.g., chemical compounds). Although subgraph enumeration and subgraph mining have been proposed to bring insights into a graph database by a set of subgraph structures, they often end up with similar or homogenous topologies, which is undesirable in many graph applications. To address this limitation, we propose the Top-k Edge-Diversified Patterns Discovery problem to retrieve a set of subgraphs that cover the maximum number of edges in a database. To efficiently process such query, we present a generic and extensible framework called Ted which achieves a guaranteed approximation ratio to the optimal result. Two optimization strategies are further developed to improve the performance. Experimental studies on real-world datasets demonstrate the superiority of Ted to traditional techniques.

With the rapid increase of mobile devices and online media, more and more people prefer posting/viewing videos online. Generally, these videos are presented on video streaming sites with image thumbnails and text titles. While facing huge amounts of videos, a viewer clicks through a certain video with high probability because of its eye-catching thumbnail. However, current video thumbnails are created manually, which is time-consuming and quality-unguaranteed. And static image thumbnails contain very limited information of the corresponding videos, which prevents users from successfully clicking what they really want to view. In this paper, we address a novel problem, namely GIF thumbnail generation, which aims to automatically generate GIF thumbnails for videos and consequently boost their Click-Through-Rate (CTR). Here, a GIF thumbnail is an animated GIF file consisting of multiple segments from the video, containing more information of the target video than a static image thumbnail. To support this study, we build the first GIF thumbnails benchmark dataset that consists of 1070 videos covering a total duration of 69.1 hours, and 5394 corresponding manually-annotated GIFs. To solve this problem, we propose a learning-based automatic GIF thumbnail generation model, which is called Generative Variational Dual-Encoder (GEVADEN). As not relying on any user interaction information (e.g. time-sync comments and real-time view counts), this model is applicable to newly-uploaded/rarely-viewed videos. Experiments on our built dataset show that GEVADEN significantly outperforms several baselines, including video-summarization and highlight-detection based ones. Furthermore, we develop a pilot application of the proposed model on an online video platform with 9814 videos covering 1231 hours, which shows that our model achieves a 37.5% CTR improvement over traditional image thumbnails. This further validates the effectiveness of the proposed model and the promising application prospect of GIF thumbnails.

Despite powerful representation ability, deep neural networks (DNNs) are prone to over-fitting, because of over-parametrization. Existing works have explored various regularization techniques to tackle the over-fitting problem. Some of them employed soft targets rather than one-hot labels to guide network training (e.g. label smoothing in classification tasks), which are called target-based regularization approaches in this paper. To alleviate the over-fitting problem, here we propose a new and general regularization framework that introduces an auxiliary network to dynamically incorporate guided semantic disturbance to the labels. We call it Network as Regularization (NaR in short). During training, the disturbance is constructed by a convex combination of the predictions of the target network and the auxiliary network. These two networks are initialized separately. And the auxiliary network is trained independently from the target network, while providing instance-level and class-level semantic information to the latter progressively. We conduct extensive experiments to validate the effectiveness of the proposed method. Experimental results show that NaR outperforms many state-of-the-art target-based regularization methods, and other regularization approaches (e.g. mixup) can also benefit from combining with NaR.

Most of the existing methods for anomaly detection use only positive data to learn the data distribution, thus they usually need a pre-defined threshold at the detection stage to determine whether a test instance is an outlier. Unfortunately, a good threshold is vital for the performance and it is really hard to find an optimal one. In this paper, we take the discriminative information implied in unlabeled data into consideration and propose a new method for anomaly detection that can learn the labels of unlabelled data directly. Our proposed method has an end-to-end architecture with one encoder and two decoders that are trained to model inliers and outliers’ data distributions in a competitive way. This architecture works in a discriminative manner without suffering from overfitting, and the training algorithm of our model is adopted from SGD, thus it is efficient and scalable even for large-scale datasets. Empirical studies on 7 datasets including KDD99, MNIST, Caltech-256, and ImageNet etc. show that our model outperforms the state-of-the-art methods.

Most existing ensemble methods aim to train the underlying embedded models independently and simply aggregate their final outputs via averaging or weighted voting. As many prediction tasks contain uncertainty, most of these ensemble methods just reduce variance of the predictions without considering the collaborations among the ensembles. Different from these ensemble methods, multiple choice learning (MCL) methods exploit the cooperation among all the embedded models to generate multiple diverse hypotheses. In this paper, a new MCL method, called vMCL (the abbreviation of versatile Multiple Choice Learning), is developed to extend the application scenarios of MCL methods by ensembling deep neural networks. Our vMCL method keeps the advantage of existing MCL methods while overcoming their major drawback, thus achieves better performance. The novelty of our vMCL lies in three aspects: (1) a choice network is designed to learn the confidence level of each specialist which can provide the best prediction base on multiple hypotheses; (2) a hinge loss is introduced to alleviate the overconfidence issue in MCL settings; (3) Easy to be implemented and can be trained in an end-to-end manner, which is a very attractive feature for many real-world applications. Experiments on image classification and image segmentation task show that vMCL outperforms the existing state-of-the-art MCL methods.

Video compression artifact reduction aims to recover high-quality videos from low-quality compressed videos. Most existing approaches use a single neighboring frame or a pair of neighboring frames (preceding and/or following the target frame) for this task. Furthermore, as frames of high quality overall may contain low-quality patches, and high-quality patches may exist in frames of low quality overall, current methods focusing on nearby peak-quality frames (PQFs) may miss high-quality details in low-quality frames. To remedy these shortcomings, in this paper we propose a novel end-to-end deep neural network called non-local ConvLSTM (NL-ConvLSTM in short) that exploits multiple consecutive frames. An approximate non-local strategy is introduced in NL-ConvLSTM to capture global motion patterns and trace the spatiotemporal dependency in a video sequence. This approximate strategy makes the non-local module work in a fast and low space-cost way. Our method uses the preceding and following frames of the target frame to generate a residual, from which a higher quality frame is reconstructed. Experiments on two datasets show that NL-ConvLSTM outperforms the existing methods.

BACKGROUND: Long non-coding RNA (lncRNA) plays important roles in many biological and pathological processes, including transcriptional regulation and gene regulation. As lncRNA interacts with multiple proteins, predicting lncRNA-protein interactions (lncRPIs) is an important way to study the functions of lncRNA. Up to now, there have been a few works that exploit protein-protein interactions (PPIs) to help the prediction of new lncRPIs. RESULTS: In this paper, we propose to boost the prediction of lncRPIs by fusing multiple protein-protein similarity networks (PPSNs). Concretely, we first construct four PPSNs based on protein sequences, protein domains, protein GO terms and the STRING database respectively, then build a more informative PPSN by fusing these four constructed PPSNs. Finally, we predict new lncRPIs by a random walk method with the fused PPSN and known lncRPIs. Our experimental results show that the new approach outperforms the existing methods. CONCLUSION: Fusing multiple protein-protein similarity networks can effectively boost the performance of predicting lncRPIs.

The identification of interactions between compounds and proteins plays an important role in network pharmacology and drug discovery. However, experimentally identifying compound-protein interactions (CPIs) is generally expensive and time-consuming, computational approaches are thus introduced. Among these, machine-learning based methods have achieved a considerable success. However, due to the nonlinear and imbalanced nature of biological data, many machine learning approaches have their own limitations. Recently, deep learning techniques show advantages over many state-of-the-art machine learning methods in many applications. In this study, we aim at improving the performance of CPI prediction based on deep learning, and propose a method called DL-CPI (the abbreviation of Deep Learning for Compound-Protein Interactions prediction), which employs deep neural network (DNN) to effectively learn the representations of compound-protein pairs. Extensive experiments show that DL-CPI can learn useful features of compound-protein pairs by a layerwise abstraction, and thus achieves better prediction performance than existing methods on both balanced and imbalanced datasets.