The Entity Set Expansion (ESE) task aims to expand a handful of seed entities with new entities belonging to the same semantic class. Conventional ESE methods are based on mono-modality (i.e., literal modality), which struggle to deal with complex entities in the real world such as (1) Negative entities with fine-grained semantic differences. (2) Synonymous entities. (3) Polysemous entities. (4) Long-tailed entities. These challenges prompt us to propose novel Multi-modal Entity Set Expansion (MESE), where models integrate information from multiple modalities to represent entities. Intuitively, the benefits of multi-modal information for ESE are threefold: (1) Different modalities can provide complementary information. (2) Multi-modal information provides a unified signal via common visual properties for the same semantic class or entity. (3) Multi-modal information offers robust alignment signals for synonymous entities. To assess model performance in MESE, we constructed the MESED dataset which is the first multi-modal dataset for ESE with large-scale and elaborate manual calibration. A powerful multi-modal model MultiExpan is proposed which is pre-trained on four multimodal pre-training tasks. The extensive experiments and analyses on MESED demonstrate the high quality of the dataset and the effectiveness of our MultiExpan, as well as pointing the direction for future research. The benchmark and code are public at https://github.com/THUKElab/MESED.

Noise is a pervasive issue in recommendation systems, which can stem from user behaviors that do not align with their intentions. As a result, noise reduction has become a prominent area of research in the field of recommendation systems. However, existing noise reduction techniques in recommendation tend to compromise the performance of certain task objectives. Moreover, they require modifying the structure of the model, which introduces inference latency and additional space cost. In this paper, we propose a straightforward yet powerful approach, Multi-layer Tensor trace Norm (MTN), to address noise-related challenges. Our method achieves this by promoting information sharing across different tasks using tensor trace norms. By leveraging norms, MTN effectively reduces noise without modifying the model’s structure or incurring substantial time and space complexities. Extensive experiments on public datasets and generated noisy datasets demonstrate the effectiveness of MTN on several of the most popular multi-task models.

Multi-task ranking models have become essential for modern real-world recommendation systems. While most recommendation researches focus on designing sophisticated models for specific scenarios, achieving performance improvement for multi-task ranking models across various scenarios still remains a significant challenge. Training all tasks naively can result in in-consistent learning, highlighting the need for the development of multi-task optimization (MTO) methods to tackle this challenge. Conventional methods assume that the optimal joint gradient on shared parameters leads to optimal parameter updates. However, the actual update on model parameters may deviates significantly from gradients when using momentum based optimizers such as Adam. In this paper, we propose a novel Parameter Update Balancing algorithm for multi-task optimization, denoted as PUB. In contrast to traditional MTO method which are based on gradient level tasks fusion or loss level tasks fusion, PUB is the first work to optimize multiple tasks through parameter update balancing. Comprehensive experiments on benchmark multi-task ranking datasets demonstrate that PUB consistently improves several multi-task backbones and achieves state-of-the-art performance. Furthermore, we deployed our method for an industrial evaluation on the real-world commercial platform, HUAWEI AppGallery, where PUB significantly enhances the on-line multi-task ranking model, efficiently managing the primary traffic of a crucial channel.

Music source separation and pitch estimation are two vital tasks in music information retrieval. Typically, the input of pitch estimation is obtained from the output of music source separation. Therefore, existing methods have tried to perform these two tasks simultaneously, so as to leverage the mutually beneficial relationship between both tasks. However, these methods still face two critical challenges that limit the improvement of both tasks: the lack of labeled data and joint learning optimization. To address these challenges, we propose a Model-Agnostic Joint Learning (MAJL) framework for both tasks. MAJL is a generic framework and can use variant models for each task. It includes a two-stage training method and a dynamic weighting method named Dynamic Weights on Hard Samples (DWHS), which addresses the lack of labeled data and joint learning optimization, respectively. Experimental results on public music datasets show that MAJL outperforms state-of-the-art methods on both tasks, with significant improvements of 0.92 in Signal-to-Distortion Ratio (SDR) for music source separation and 2.71% in Raw Pitch Accuracy (RPA) for pitch estimation. Furthermore, comprehensive studies not only validate the effectiveness of each component of MAJL, but also indicate the great generality of MAJL in adapting to different model architectures.

Melody extraction is a core task in music information retrieval, and the estimation of pitch, onset and offset are key sub-tasks in melody extraction. Existing methods have limited accuracy, and work for only one type of data, either single-pitch or multi-pitch. In this paper, we propose a highly accurate method for joint estimation of pitch, onset and offset, named JEPOO. We address the challenges of joint learning optimization and handling both single-pitch and multi-pitch data through novel model design and a new optimization technique named Pareto modulated loss with loss weight regularization. This is the first method that can accurately handle both single-pitch and multi-pitch music data, and even a mix of them. A comprehensive experimental study on a wide range of real datasets shows that JEPOO outperforms state-of-the-art methods by up to 10.6\%, 8.3\% and 10.3\% for the prediction of Pitch, Onset and Offset, respectively, and JEPOO is robust for various types of data and instruments. The ablation study validates the effectiveness of each component of JEPOO.

The goal of sequential recommendation (SR) is to predict a user's potential interested items based on her/his historical interaction sequences. Most existing sequential recommenders are developed based on ID features, which, despite their widespread use, often underperform with sparse IDs and struggle with the cold-start problem. Besides, inconsistent ID mappings hinder the model's transferability, isolating similar recommendation domains that could have been co-optimized. This paper aims to address these issues by exploring the potential of multi-modal information in learning robust and generalizable sequence representations. We propose MISSRec, a multi-modal pre-training and transfer learning framework for SR. On the user side, we design a Transformer-based encoder-decoder model, where the contextual encoder learns to capture the sequence-level multi-modal synergy while a novel interest-aware decoder is developed to grasp item-modality-interest relations for better sequence representation. On the candidate item side, we adopt a dynamic fusion module to produce user-adaptive item representation, providing more precise matching between users and items. We pre-train the model with contrastive learning objectives and fine-tune it in an efficient manner. Extensive experiments demonstrate the effectiveness and flexibility of MISSRec, promising an practical solution for real-world recommendation scenarios.

Embedding knowledge graphs (KGs) into continuous vector space is an essential problem in knowledge extraction. Current models continue to improve embedding by focusing on discriminating relation-specific information from entities with increasingly complex feature engineering. We noted that they ignored the inherent relevance between relations and tried to learn unique discriminate parameter set for each relation. Thus, these models potentially suffer from high time complexity and large parameters, preventing them from efficiently applying on real-world KGs. In this paper, we follow the thought of parameter sharing to simultaneously learn more expressive features, reduce parameters and avoid complex feature engineering. Based on gate structure from LSTM, we propose a novel model TransGate and develop shared discriminate mechanism, resulting in almost same space complexity as indiscriminate models. Furthermore, to develop a more effective and scalable model, we reconstruct the gate with weight vectors making our method has comparative time complexity against indiscriminate model. We conduct extensive experiments on link prediction and triplets classification. Experiments show that TransGate not only outperforms state-of-art baselines, but also reduces parameters greatly. For example, TransGate outperforms ConvE and RGCN with 6x and 17x fewer parameters, respectively. These results indicate that parameter sharing is a superior way to further optimize embedding and TransGate finds a better trade-off between complexity and expressivity.

Anomalies detection is to spot the objects whose patterns singularly differ from the reference majority. Attributed networks often contain node attributes and network structure, which are widely used for real-life applications. Meanwhile, how to detect anomalies on attributed networks has caused a lot of attention. Most existing works on anomaly detection attempt to incorporate node attributes with the network structure. However, there may exist structurally irrelevant attributes in the networks, which may have adverse effects on the detection results. Besides, the heterogeneity of node attributes and network structure may further make the detection of anomalies difficult. In order to overcome the above challenges, in this paper, we propose a Robust Embedding Method for Anomaly Detection on Attributed Networks, called REMAD. Methodologically, the proposed REMAD combines network embedding and residual analysis together. By performing network embedding on the network, REMAD obtains the representative attributes that are closely coherent with the network structure. Simultaneously, by adopting the residual analysis, REMAD characterizes and analyzes the residuals of attribute information to discover anomalies. Experimental results on both synthetic and real-world datasets demonstrate the advantages of our proposed method REMAD against the state-of-the-art anomaly detection methods.