Explicit Uncertainty Modeling for Video Watch Time Prediction

In video recommendation, a critical component that determines the system's recommendation accuracy is the watch-time prediction module, since how long a user watches a video directly reflects personalized preferences. One of the key challenges of this problem is the user's stochastic watch-time behavior. To improve the prediction accuracy for such an uncertain behavior, existing approaches show that one can either reduce the noise through duration bias modeling or formulate a distribution modeling task to capture the uncertainty. However, the uncontrolled uncertainty is not always equally distributed across users and videos, inducing a balancing paradox between the model accuracy and the ability to capture out-of-distribution samples. In practice, we find that the uncertainty of the watch-time prediction model also provides key information about user behavior, which, in turn, could benefit the prediction task itself. Following this notion, we derive an explicit uncertainty modeling strategy for the prediction model and propose an adversarial optimization framework that can better exploit the user watch-time behavior. This framework has been deployed online on an industrial video sharing platform that serves hundreds of millions of daily active users, which obtains a significant increase in users' video watch time by 0.31% through the online A/B test. Furthermore, extended offline experiments on two public datasets verify the effectiveness of the proposed framework across various watch-time prediction backbones.

Value Function Decomposition in Markov Recommendation Process

Recent advances in recommender systems have shown that user-system interaction essentially formulates long-term optimization problems, and online reinforcement learning can be adopted to improve recommendation performance. The general solution framework incorporates a value function that estimates the user's expected cumulative rewards in the future and guides the training of the recommendation policy. To avoid local maxima, the policy may explore potential high-quality actions during inference to increase the chance of finding better future rewards. To accommodate the stepwise recommendation process, one widely adopted approach to learning the value function is learning from the difference between the values of two consecutive states of a user. However, we argue that this paradigm involves an incorrect approximation in the stochastic process. Specifically, between the current state and the next state in each training sample, there exist two separate random factors from the stochastic policy and the uncertain user environment. Original temporal difference (TD) learning under these mixed random factors may result in a suboptimal estimation of the long-term rewards. As a solution, we show that these two factors can be separately approximated by decomposing the original temporal difference loss. The disentangled learning framework can achieve a more accurate estimation with faster learning and improved robustness against action exploration. As empirical verification of our proposed method, we conduct offline experiments with online simulated environments built based on public datasets.

Coarse-to-fine Dynamic Uplift Modeling for Real-time Video Recommendation

With the rise of short video platforms, video recommendation technology faces more complex challenges. Currently, there are multiple non-personalized modules in the video recommendation pipeline that urgently need personalized modeling techniques for improvement. Inspired by the success of uplift modeling in online marketing, we attempt to implement uplift modeling in the video recommendation scenario. However, we face two main challenges: 1) Design and utilization of treatments, and 2) Capture of user real-time interest. To address them, we design adjusting the distribution of videos with varying durations as the treatment and propose Coarse-to-fine Dynamic Uplift Modeling (CDUM) for real-time video recommendation. CDUM consists of two modules, CPM and FIC. The former module fully utilizes the offline features of users to model their long-term preferences, while the latter module leverages online real-time contextual features and request-level candidates to model users' real-time interests. These two modules work together to dynamically identify and targeting specific user groups and applying treatments effectively. Further, we conduct comprehensive experiments on the offline public and industrial datasets and online A/B test, demonstrating the superiority and effectiveness of our proposed CDUM. Our proposed CDUM is eventually fully deployed on the Kuaishou platform, serving hundreds of millions of users every day. The source code will be provided after the paper is accepted.

Incorporating Group Prior into Variational Inference for Tail-User Behavior Modeling in CTR Prediction

User behavior modeling -- which aims to extract user interests from behavioral data -- has shown great power in Click-through rate (CTR) prediction, a key component in recommendation systems. Recently, attention-based algorithms have become a promising direction, as attention mechanisms emphasize the relevant interactions from rich behaviors. However, the methods struggle to capture the preferences of tail users with sparse interaction histories. To address the problem, we propose a novel variational inference approach, namely Group Prior Sampler Variational Inference (GPSVI), which introduces group preferences as priors to refine latent user interests for tail users. In GPSVI, the extent of adjustments depends on the estimated uncertainty of individual preference modeling. In addition, We further enhance the expressive power of variational inference by a volume-preserving flow. An appealing property of the GPSVI method is its ability to revert to traditional attention for head users with rich behavioral data while consistently enhancing performance for long-tail users with sparse behaviors. Rigorous analysis and extensive experiments demonstrate that GPSVI consistently improves the performance of tail users. Moreover, online A/B testing on a large-scale real-world recommender system further confirms the effectiveness of our proposed approach.

End-to-End Learnable Item Tokenization for Generative Recommendation

Recently, generative recommendation has emerged as a promising new paradigm that directly generates item identifiers for recommendation. However, a key challenge lies in how to effectively construct item identifiers that are suitable for recommender systems. Existing methods typically decouple item tokenization from subsequent generative recommendation training, likely resulting in suboptimal performance. To address this limitation, we propose ETEGRec, a novel End-To-End Generative Recommender by seamlessly integrating item tokenization and generative recommendation. Our framework is developed based on the dual encoder-decoder architecture, which consists of an item tokenizer and a generative recommender. In order to achieve mutual enhancement between the two components, we propose a recommendation-oriented alignment approach by devising two specific optimization objectives: sequence-item alignment and preference-semantic alignment. These two alignment objectives can effectively couple the learning of item tokenizer and generative recommender, thereby fostering the mutual enhancement between the two components. Finally, we further devise an alternating optimization method, to facilitate stable and effective end-to-end learning of the entire framework. Extensive experiments demonstrate the effectiveness of our proposed framework compared to a series of traditional sequential recommendation models and generative recommendation baselines.

Towards Robust Recommendation via Decision Boundary-aware Graph Contrastive Learning

In recent years, graph contrastive learning (GCL) has received increasing attention in recommender systems due to its effectiveness in reducing bias caused by data sparsity. However, most existing GCL models rely on heuristic approaches and usually assume entity independence when constructing contrastive views. We argue that these methods struggle to strike a balance between semantic invariance and view hardness across the dynamic training process, both of which are critical factors in graph contrastive learning. To address the above issues, we propose a novel GCL-based recommendation framework RGCL, which effectively maintains the semantic invariance of contrastive pairs and dynamically adapts as the model capability evolves through the training process. Specifically, RGCL first introduces decision boundary-aware adversarial perturbations to constrain the exploration space of contrastive augmented views, avoiding the decrease of task-specific information. Furthermore, to incorporate global user-user and item-item collaboration relationships for guiding on the generation of hard contrastive views, we propose an adversarial-contrastive learning objective to construct a relation-aware view-generator. Besides, considering that unsupervised GCL could potentially narrower margins between data points and the decision boundary, resulting in decreased model robustness, we introduce the adversarial examples based on maximum perturbations to achieve margin maximization. We also provide theoretical analyses on the effectiveness of our designs. Through extensive experiments on five public datasets, we demonstrate the superiority of RGCL compared against twelve baseline models.

IFA: Interaction Fidelity Attention for Entire Lifelong Behaviour Sequence Modeling

The lifelong user behavior sequence provides abundant information of user preference and gains impressive improvement in the recommendation task, however increases computational consumption significantly. To meet the severe latency requirement in online service, a short sub-sequence is sampled based on similarity to the target item. Unfortunately, items not in the sub-sequence are abandoned, leading to serious information loss. In this paper, we propose a new efficient paradigm to model the full lifelong sequence, which is named as \textbf{I}nteraction \textbf{F}idelity \textbf{A}ttention (\textbf{IFA}). In IFA, we input all target items in the candidate set into the model at once, and leverage linear transformer to reduce the time complexity of the cross attention between the candidate set and the sequence without any interaction information loss. We also additionally model the relationship of all target items for optimal set generation, and design loss function for better consistency of training and inference. We demonstrate the effectiveness and efficiency of our model by off-line and online experiments in the recommender system of Kuaishou.

Modeling User Retention through Generative Flow Networks

Recommender systems aim to fulfill the user's daily demands. While most existing research focuses on maximizing the user's engagement with the system, it has recently been pointed out that how frequently the users come back for the service also reflects the quality and stability of recommendations. However, optimizing this user retention behavior is non-trivial and poses several challenges including the intractable leave-and-return user activities, the sparse and delayed signal, and the uncertain relations between users' retention and their immediate feedback towards each item in the recommendation list. In this work, we regard the retention signal as an overall estimation of the user's end-of-session satisfaction and propose to estimate this signal through a probabilistic flow. This flow-based modeling technique can back-propagate the retention reward towards each recommended item in the user session, and we show that the flow combined with traditional learning-to-rank objectives eventually optimizes a non-discounted cumulative reward for both immediate user feedback and user retention. We verify the effectiveness of our method through both offline empirical studies on two public datasets and online A/B tests in an industrial platform.

Modeling User Fatigue for Sequential Recommendation

Recommender systems filter out information that meets user interests. However, users may be tired of the recommendations that are too similar to the content they have been exposed to in a short historical period, which is the so-called user fatigue. Despite the significance for a better user experience, user fatigue is seldom explored by existing recommenders. In fact, there are three main challenges to be addressed for modeling user fatigue, including what features support it, how it influences user interests, and how its explicit signals are obtained. In this paper, we propose to model user Fatigue in interest learning for sequential Recommendations (FRec). To address the first challenge, based on a multi-interest framework, we connect the target item with historical items and construct an interest-aware similarity matrix as features to support fatigue modeling. Regarding the second challenge, built upon feature cross, we propose a fatigue-enhanced multi-interest fusion to capture long-term interest. In addition, we develop a fatigue-gated recurrent unit for short-term interest learning, with temporal fatigue representations as important inputs for constructing update and reset gates. For the last challenge, we propose a novel sequence augmentation to obtain explicit fatigue signals for contrastive learning. We conduct extensive experiments on real-world datasets, including two public datasets and one large-scale industrial dataset. Experimental results show that FRec can improve AUC and GAUC up to 0.026 and 0.019 compared with state-of-the-art models, respectively. Moreover, large-scale online experiments demonstrate the effectiveness of FRec for fatigue reduction. Our codes are released at https://github.com/tsinghua-fib-lab/SIGIR24-FRec.

A Model-based Multi-Agent Personalized Short-Video Recommender System

Recommender selects and presents top-K items to the user at each online request, and a recommendation session consists of several sequential requests. Formulating a recommendation session as a Markov decision process and solving it by reinforcement learning (RL) framework has attracted increasing attention from both academic and industry communities. In this paper, we propose a RL-based industrial short-video recommender ranking framework, which models and maximizes user watch-time in an environment of user multi-aspect preferences by a collaborative multi-agent formulization. Moreover, our proposed framework adopts a model-based learning approach to alleviate the sample selection bias which is a crucial but intractable problem in industrial recommender system. Extensive offline evaluations and live experiments confirm the effectiveness of our proposed method over alternatives. Our proposed approach has been deployed in our real large-scale short-video sharing platform, successfully serving over hundreds of millions users.