Capturing user interests across extensive watch histories is critical for short-form video recommendation, yet scaling sequence length is limited by two bottlenecks: the semantic sparsity of atomic Video IDs and the quadratic computational complexity of Transformers. Traditional orthogonal Video IDs fail to capture content relationships and demand large embedding tables, while the quadratic complexity of self-attention restricts the maximum sequence length under strict industrial latency and resource constraints. In this work, we present a production-deployed framework for modeling ultra-long user behavior sequences at a billion-user scale. We first address the representation bottleneck by adopting content-native Semantic IDs. By utilizing depth-truncated, coarse-grained Semantic IDs, we shrink the embedding table size from corpus cardinality. This compact representation naturally generalizes to cold-start content through shared semantic prefixes. Second, to overcome the sequence scaling barrier, we introduce a Global-Aware Compression Transformer that leverages non-parametric temporal folding and unified global query integration to effectively condense the sequence, alleviating both the memory and computational bottlenecks of standard self-attention. Offline profiling on our computing infrastructure demonstrates an order-of-magnitude reduction in peak memory footprint and a drastic decrease in computational overhead. This efficiency gain enables supporting longer sequence lengths at an affordable cost in production, yielding substantial online gains in satisfied user engagement and satisfied content consumption in large-scale online A/B tests.

Recently, quantization has been widely used for the compression and acceleration of large language models (LLMs). Due to the outliers in LLMs, it is crucial to flatten weights and activations to minimize quantization error with equally spaced quantization points. Prior research explores various pre-quantization transformations to suppress outliers, such as per-channel scaling and Hadamard transformation. However, we observe that these transformed weights and activations can still exhibit steep and dispersed distributions. In this paper, we propose FlatQuant (Fast and Learnable Affine Transformation), a new post-training quantization approach that enhances the flatness of weights and activations. Our approach identifies optimal affine transformations for each linear layer, calibrated in hours via a lightweight objective. To reduce runtime overhead of affine transformation, we apply Kronecker product with two lightweight matrices, and fuse all operations in FlatQuant into a single kernel. Extensive experiments demonstrate that FlatQuant establishes a new state-of-the-art benchmark for quantization. For example, it achieves less than 1\% accuracy drop for W4A4 quantization on the LLaMA-3-70B model, surpassing SpinQuant by 7.5\%. Additionally, it provides up to 2.3x prefill speedup and 1.7x decoding speedup compared to the FP16 model. Code is available at: https://github.com/ruikangliu/FlatQuant.

Simplex component analysis (SCA) aims to estimate the vertices of the convex hull where data samples reside in. SCA finds various applications in signal processing, e.g., hyperspectral unmixing and noisy label learning. Recent works proposed to tackle SCA from a probabilistic viewpoint using variational inference (VI) tools, which fends against noise more effectively relative to the deterministic counterparts. However, the computational efficiency of VI for SCA hinges on the use of the Dirichlet variational posterior. Such variational posterior appears to lack expressiveness—making the SCA performance limited if the true posterior is complex. This work proposes to employ a logistic-normal variational posterior, which exhibits enhanced expressive power. To circumvent the computational bottleneck, a neural representation-based inference algorithm is proposed—which exploits a connection between the logistic-normal distribution and variational auto-encoding. Numerical experiments using simulated and semi-real data are conducted to showcase the effectiveness of our algorithm design.

Exposure bias and its induced feedback loop effect are well-known problems in recommender systems. Exploration is believed to be the key to break such feedback loops. While classical contextual bandit algorithms such as Upper-Confidence-Bound and Thompson Sampling have been successful in addressing the exploration-exploitation trade-off in the single-task settings with one clear reward signal, modern recommender systems often leverage multiple rich sources of feedback such as clicks, likes, dislikes, shares, satisfaction survey responses, and employ multi-task learning in practice. It is unclear how one can incorporate exploration in the multi-task setup with different objectives. In this paper, we study an efficient bandit algorithm tailored to multi-task recommender systems, named Multi-task Neural Linear Bandit (mtNLB). In particular, we investigate efficient feature embeddings in the multi-task setups that could be used as contextual features in the Neural Linear Bandit, a contextual bandit algorithm that nicely combines the representation power from DNN and simplicity in uncertainty calculation from linear models. We further study cost-effective approximations of the uncertainty estimate and principled ways to incorporate uncertainty into the multi-task scoring of items. To showcase the efficacy of our proposed method, we conduct live experiments on a large-scale commercial recommendation platform that serves billions of users. We evaluate the quality of the uncertainty estimate and demonstrate its ability to improve exploration across the different dimensions of the reward signals in comparison to baseline approaches.

Sequential models are invaluable for powering personalized recommendation systems. In the context of short-form video (SFV) feeds, where user behavior history is typically longer, systems must be able to understand users’ long-term interests. However, deploying large sequence models to extensive web-scale applications faces challenges due to high serving cost. To address this, we propose an industrial framework designed for efficiently serving large user sequence models. Specifically, the proposed infrastructure decouples serving of the user sequence model and the main recommendation model, with the user sequence model being served offline (asynchronously) with periodical refresh. The proposed infrastructure is also model-agnostic; thus, it can be used to support any type of user sequence models (even LLMs) with controllable costs. Empirical results show that large user models deployed with our framework significantly and consistently enhance the quality of the main recommendation model with minimal serving costs increase.

Multi-task prediction models and value models are the de-facto standard ranking components in modern large-scale content recommendation systems. However, they are typically optimized to model users’ passive consumption behaviors, and rank content in a way to grow only consumption-centric values. In this talk, we discuss the key insight that it is possible to model sparse participatory content-generation actions as well and grow ecosystem value through a new ranking system. We made the following key technical contributions in this system: (1) introducing ranking for content generation based on a categorization of user participation actions of different sparsity, including proxy intent action or access point clicks. (2) improving sparse task prediction quality and stability by causal task relationship modeling, conditional loss modeling and ResNet based shared bottom network. (3) personalizing the value model to minimize conflicts between different values, through e.g. ranking inspiring content higher for users who actively generate content. (4) conducting systematic evaluation of proposed approach in a large short-form video UGC (User-Generated Content) platform.

Effective exploration is believed to positively influence the long-term user experience on recommendation platforms. Determining its exact benefits, however, has been challenging. Regular A/B tests on exploration often measure neutral or even negative engagement metrics while failing to capture its long-term benefits. We here introduce new experiment designs to formally quantify the long-term value of exploration by examining its effects on content corpus, and connecting content corpus growth to the long-term user experience from real-world experiments. Once established the values of exploration, we investigate the Neural Linear Bandit algorithm as a general framework to introduce exploration into any deep learning based ranking systems. We conduct live experiments on one of the largest short-form video recommendation platforms that serves billions of users to validate the new experiment designs, quantify the long-term values of exploration, and to verify the effectiveness of the adopted neural linear bandit algorithm for exploration.

In recent years, social media users spend significant amount of time on Short-Form Video (SFV) platforms. Its success in creating an immersive viewership experience is not only from the content, but also due to its unique UI innovation: instead of providing choices for users to click, SFV platforms actively recommend content to users to watch one at a time. In this paper, we highlight unique challenges rooted from such UI changes for SFV recommendation system design. Firstly, there is yet much unexplored for sources of system biases under the new UI, as there are no clicks nor the common click-based position biases. Additionally, when training multiple types of user activities, positive labels for activities like sharing and commenting can be much sparser and more skewed than traditional click-based recommendation systems, as the latter can filter non-click impressions when generating "post-click" activities.

In this paper we develop a simple scheme for a coupled matrix factorization problem arising in the topic of hyperspectral super-resolution (HSR). HSR considers the problem of recovering a super-resolution image from a multispectral image and a hyperspectral image, which have lower spectral and spatial resolutions, respectively, and it is a motivated topic in the domain of remote sensing. The challenge with coupled factorization (COFAC) is that we are required to simultaneously factorize two data matrices, with their factors being interrelated. We adopt a separable COFAC strategy, in which we first factorize one data matrix, and then use the retrieved factors and the coupled factor relationship to help us factorize another matrix; the merit is that it may lead to simple COFAC schemes. Our scheme is based on the simplex-structured factorization model, which is commonly used in HSR, and a sparse factor assumption. In particular, we leverage the coupled factor structure to exploit sparsity in an easy way; we solve simple constrained least squares problems, and we sidestep the need to do sparse optimization. Numerical results show that our proposed scheme works reasonably on both semi-real data and synthetic data, and it runs much faster than some state-of-the-art COFAC schemes.

Promising progress has been made toward learning efficient time series representations in recent years, but the learned representations often lack interpretability and do not encode semantic meanings by the complex interactions of many latent factors. Learning representations that disentangle these latent factors can bring semantic-rich representations of time series and further enhance interpretability. However, directly adopting the sequential models, such as Long Short-Term Memory Variational AutoEncoder (LSTM-VAE), would encounter a Kullback?Leibler (KL) vanishing problem: the LSTM decoder often generates sequential data without efficiently using latent representations, and the latent spaces sometimes could even be independent of the observation space. And traditional disentanglement methods may intensify the trend of KL vanishing along with the disentanglement process, because they tend to penalize the mutual information between the latent space and the observations. In this paper, we propose Disentangle Time-Series, a novel disentanglement enhancement framework for time series data. Our framework achieves multi-level disentanglement by covering both individual latent factors and group semantic segments. We propose augmenting the original VAE objective by decomposing the evidence lower-bound and extracting evidence linking factorial representations to disentanglement. Additionally, we introduce a mutual information maximization term between the observation space to the latent space to alleviate the KL vanishing problem while preserving the disentanglement property. Experimental results on five real-world IoT datasets demonstrate that the representations learned by DTS achieve superior performance in various tasks with better interpretability.

Anomaly detection is an important data mining task with numerous applications, such as intrusion detection, credit card fraud detection, and video surveillance. However, given a specific complicated task with complicated data, the process of building an effective deep learning-based system for anomaly detection still highly relies on human expertise and laboring trials. Also, while neural architecture search (NAS) has shown its promise in discovering effective deep architectures in various domains, such as image classification, object detection, and semantic segmentation, contemporary NAS methods are not suitable for anomaly detection due to the lack of intrinsic search space, unstable search process, and low sample efficiency. To bridge the gap, in this article, we propose AutoAD, an automated anomaly detection framework, which aims to search for an optimal neural network model within a predefined search space. Specifically, we first design a curiosity-guided search strategy to overcome the curse of local optimality. A controller, which acts as a search agent, is encouraged to take actions to maximize the information gain about the controller's internal belief. We further introduce an experience replay mechanism based on self-imitation learning to improve the sample efficiency. Experimental results on various real-world benchmark datasets demonstrate that the deep model identified by AutoAD achieves the best performance, comparing with existing handcrafted models and traditional search methods.

Outlier detection is an important data mining task with numerous applications such as intrusion detection, credit card fraud detection, and video surveillance. However, given a specific task with complex data, the process of building an effective deep learning based system for outlier detection still highly relies on human expertise and laboring trials. Moreover, while Neural Architecture Search (NAS) has shown its promise in discovering effective deep architectures in various domains, such as image classification, object detection and semantic segmentation, contemporary NAS methods are not suitable for outlier detection due to the lack of intrinsic search space and low sample efficiency. To bridge the gap, in this paper, we propose AutoOD, an automated outlier detection framework, which aims to search for an optimal neural network model within a predefined search space. Specifically, we introduce an experience replay mechanism based on self-imitation learning to improve the sample efficiency. Experimental results on various real-world benchmark datasets demonstrate that the deep model identified by AutoOD achieves the best performance, comparing with existing handcrafted models and traditional search methods.

Image captioning has made substantial progress with huge supporting image collections sourced from the web. However, recent studies have pointed out that captioning datasets, such as COCO, contain gender bias found in web corpora. As a result, learning models could heavily rely on the learned priors and image context for gender identification, leading to incorrect or even offensive errors. To encourage models to learn correct gender features, we reorganize the COCO dataset and present two new splits COCO-GB V1 and V2 datasets where the train and test sets have different gender-context joint distribution. Models relying on contextual cues will suffer from huge gender prediction errors on the anti-stereotypical test data. Benchmarking experiments reveal that most captioning models learn gender bias, leading to high gender prediction errors, especially for women. To alleviate the unwanted bias, we propose a new Guided Attention Image Captioning model (GAIC) which provides self-guidance on visual attention to encourage the model to capture correct gender visual evidence. Experimental results validate that GAIC can significantly reduce gender prediction errors with a competitive caption quality. Our codes and the designed benchmark datasets are available at https://github.com/datamllab/Mitigating_Gender_Bias_In_Captioning_System.

With advancements of deep learning techniques, it is now possible to generate super-realistic images and videos, i.e., deepfakes. These deepfakes could reach mass audience and result in adverse impacts on our society. Although lots of efforts have been devoted to detect deepfakes, their performance drops significantly on previously unseen but related manipulations and the detection generalization capability remains a problem. Motivated by the fine-grained nature and spatial locality characteristics of deepfakes, we propose Locality-Aware AutoEncoder (LAE) to bridge the generalization gap. In the training process, we use a pixel-wise mask to regularize local interpretation of LAE to enforce the model to learn intrinsic representation from the forgery region, instead of capturing artifacts in the training set and learning superficial correlations to perform detection. We further propose an active learning framework to select the challenging candidates for labeling, which requires human masks for less than 3% of the training data, dramatically reducing the annotation efforts to regularize interpretations. Experimental results on three deepfake detection tasks indicate that LAE could focus on the forgery regions to make decisions. The analysis further shows that LAE outperforms the state-of-the-arts by 6.52%, 12.03%, and 3.08% respectively on three deepfake detection tasks in terms of generalization accuracy on previously unseen manipulations.

Hyperspectral unmixing (HU) is a problem of blindly identifying the underlying materials, in form of spectral signatures, in the captured hyperspectral image. HU has received tremendous interest in remote sensing, and fundamentally the problem can be regarded as solving a simplex-structured matrix factorization problem. The majority of HU research has been focused on the linear mixture model. On the other hand, remote sensing research has long indicated that real-life hyperspectral images can exhibit spatial variant and nonlinear effects. In this study we introduce a probabilistic approach for HU under two different models, namely, the normal composition model for modeling spatial endmember variability and the generalized bilinear model for modeling multi-path nonlinear effects. Our approach formulates HU as a maximum-likelihood (ML) model parameter estimation problem, and we demonstrate that the ML formulation can flexibly handle the two models. The ML problem is technical challenging, and we apply a combination of techniques, namely, sample average approximation, block coordinate descent and majorization minimization, to tackle the problem. Our empirical study suggests that the ML method gives competitive recovery performance.

Policy distillation, which transfers a teacher policy to a student policy has achieved great success in challenging tasks of deep reinforcement learning. This teacher-student framework requires a well-trained teacher model which is computationally expensive. Moreover, the performance of the student model could be limited by the teacher model if the teacher model is not optimal. In the light of collaborative learning, we study the feasibility of involving joint intellectual efforts from diverse perspectives of student models. In this work, we introduce dual policy distillation (DPD), a student-student framework in which two learners operate on the same environment to explore different perspectives of the environment and extract knowledge from each other to enhance their learning. The key challenge in developing this dual learning framework is to identify the beneficial knowledge from the peer learner for contemporary learning-based reinforcement learning algorithms, since it is unclear whether the knowledge distilled from an imperfect and noisy peer learner would be helpful. To address the challenge, we theoretically justify that distilling knowledge from a peer learner will lead to policy improvement and propose a disadvantageous distillation strategy based on the theoretical results. The conducted experiments on several continuous control tasks show that the proposed framework achieves superior performance with a learning-based agent and function approximation without the use of expensive teacher models.

Graph neural networks (GNNs), which learn the representation of a node by aggregating its neighbors, have become an effective computational tool in downstream applications. Over-smoothing is one of the key issues which limit the performance of GNNs as the number of layers increases. It is because the stacked aggregators would make node representations converge to indistinguishable vectors. Several attempts have been made to tackle the issue by bringing linked node pairs close and unlinked pairs distinct. However, they often ignore the intrinsic community structures and would result in sub-optimal performance. The representations of nodes within the same community/class need be similar to facilitate the classification, while different classes are expected to be separated in embedding space. To bridge the gap, we introduce two over-smoothing metrics and a novel technique, i.e., differentiable group normalization (DGN). It normalizes nodes within the same group independently to increase their smoothness, and separates node distributions among different groups to significantly alleviate the over-smoothing issue. Experiments on real-world datasets demonstrate that DGN makes GNN models more robust to over-smoothing and achieves better performance with deeper GNNs.

Anomaly detection in attributed networks (instance-to-instance dependencies and interactions are available) has various applications such as monitoring suspicious accounts in social media and financial fraud in transaction networks. However, it remains a challenging task since the definition of anomaly becomes more complicated and topological structures are heterogeneous with nodal attributes. In this paper, we propose a spectral convolution and deconvolution based framework - SpecAE, to project the attributed network into a tailored space to detect global and community anomalies. SpecAE leverages Laplacian sharpening to amplify the distances between representations of anomalies and the ones of the majority. The learned representations along with reconstruction errors are combined with a density estimation model to perform the detection. Experiments on real-world datasets demonstrate the effectiveness of the proposed SpecAE.

Anomaly detection is a fundamental problem in data mining field with many real-world applications. A vast majority of existing anomaly detection methods predominately focused on data collected from a single source. In real-world applications, instances often have multiple types of features, such as images (ID photos, finger prints) and texts (bank transaction histories, user online social media posts), resulting in the so-called multi-modal data. In this paper, we focus on identifying anomalies whose patterns are disparate across different modalities, i.e., cross-modal anomalies. Some of the data instances within a multi-modal context are often not anomalous when they are viewed separately in each individual modality, but contains inconsistent patterns when multiple sources are jointly considered. The existence of multi-modal data in many real-world scenarios brings both opportunities and challenges to the canonical task of anomaly detection. On the one hand, in multimodal data, information of different modalities may complement each other in improving the detection performance. On the other hand, complicated distributions across different modalities call for a principled framework to characterize their inherent and complex correlations, which is often difficult to capture with conventional linear models. To this end, we propose a novel deep structured anomaly detection framework to identify the cross-modal anomalies embedded in the data. Experiments on real-world datasets demonstrate the effectiveness of the proposed framework comparing with the state-of-the-art.

Random walks are widely adopted in various network analysis tasks ranging from network embedding to label propagation. It could capture and convert geometric structures into structured sequences while alleviating the issues of sparsity and curse of dimensionality. Though random walks on plain networks have been intensively studied, in real-world systems, nodes are often not pure vertices, but own different characteristics, described by the rich set of data associated with them. These node attributes contain plentiful information that often complements the network, and bring opportunities to the random-walk-based analysis. However, it is unclear how random walks could be developed for attributed networks towards an effective joint information extraction. Node attributes make the node interactions more complicated and are heterogeneous with respect to topological structures.

Networks have been widely used as the data structure for abstracting real-world systems as well as organizing the relations among entities. Network embedding models are powerful tools in mapping nodes in a network into continuous vector-space representations in order to facilitate subsequent tasks such as classification and link prediction. Existing network embedding models comprehensively integrate all information of each node, such as links and attributes, towards a single embedding vector to represent the node's general role in the network. However, a real-world entity could be multifaceted, where it connects to different neighborhoods due to different motives or self-characteristics that are not necessarily correlated. For example, in a movie recommender system, a user may love comedies or horror movies simultaneously, but it is not likely that these two types of movies are mutually close in the embedding space, nor the user embedding vector could be sufficiently close to them at the same time. In this paper, we propose a polysemous embedding approach for modeling multiple facets of nodes, as motivated by the phenomenon of word polysemy in language modeling. Each facet of a node is mapped to an embedding vector, while we also maintain an association degree between each pair of node and facet. The proposed method is adaptive to various existing embedding models, without significantly complicating the optimization process. We also discuss how to engage embedding vectors of different facets for inference tasks including classification and link prediction. Experiments on real-world datasets help comprehensively evaluate the performance of the proposed method.