With the rapid evolution of Large Language Models (LLMs), generative recommendation is gradually reshaping the paradigm of recommender systems. However, most existing methods remain confined to the interaction-driven next-item prediction paradigm, struggling to keep pace with the latest evolving trends or address the diverse recommendation tasks along with business-specific requirements in real-world scenarios. To this end, we present SIGMA, a Semantic-Grounded Instruction-Driven Generative Multi-Task Recommender deployed at AliExpress. Specifically, we first ground item entities in a unified latent space capturing both general semantics and collaborative signals. Building upon this, we introduce a hybrid item tokenization method for both precise modeling and efficient generation. Moreover, we construct a large-scale multi-task supervised fine-tuning dataset empowering SIGMA to fulfill various recommendation demands via instruction-following. Finally, we design a three-step item generation procedure integrated with an adaptive probabilistic fusion mechanism to calibrate the output distributions based on task-specific requirements for recommendation accuracy and diversity. Extensive offline experiments and online A/B tests demonstrate the effectiveness of SIGMA across various real-world recommendation tasks.

Yuan Liu, Zhongyin Zhao, Le Tian, Haicheng Wang, Xubing Ye, Yangxiu You, Zilin Yu, Chuhan Wu, Zhou Xiao, Yang Yu, Jie Zhou. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Researchers have successfully adapted the privacy-preserving Federated Learning (FL) to news recommendation tasks to better protect users' privacy, although typically at the cost of performance degradation due to the data heterogeneity issue. To address this issue, Personalized Federated Learning (PFL) has emerged, among which model interpolation is a promising approach that interpolates the local personalized models with the global model. However, the existing model interpolation method may not work well for news recommendation tasks for some reasons. First, it neglects the fine-grained personalization needs at both the temporal and spatial levels in news recommendation tasks. Second, due to the cold-user problem in real-world news recommendation tasks, the local personalized models may perform poorly, thus limiting the performance gain from model interpolation. To this end, we propose FINDING (Federated News Recommendation with Fine-grained Interpolation and Dynamic Clustering ), a novel personalized federated learning framework based on model interpolation. Specifically, we first propose the fine-grained model interpolation strategy which interpolates the local personalized models with the global model in a time-aware and layer-aware way. Then, to address the cold-user problem in news recommendation tasks, we adopt the group-level personalization approach where users are dynamically clustered into groups and the group-level personalized models are used for interpolation. Extensive experiments on two real-world datasets show that our method can effectively handle the above limitations of the current model interpolation method and alleviate the heterogeneity issue faced by traditional FL.

Federated learning (FL) is important for privacy-preserving services by training models without collecting raw user data. Most FL algorithms assume all data is annotated, which is impractical due to the high cost of labeling data in real applications. To alleviate the reliance on labeled data, semi-supervised federated learning (SSFL) has been proposed to utilize unlabeled data on clients to improve model performance. However, most existing methods either have privacy issues which share models trained on other clients, or generate pseudo-labels for unlabeled local datasets with the global model, which is usually biased towards the global data distribution. The latter may lead to sub-optimal accuracy of pseudo-labels, due to the gap between the local data distribution and the global model, especially in non-IID settings. In this paper, we propose a semi-supervised heterogeneous federated learning method with local knowledge enhancement, called FedLoKe, which aims to train an accurate global model from both labeled and unlabeled local data with non-IID distributions. Specifically, in FedLoKe, the server maintains a global model to capture global data distribution, and each client learns a local model to capture local data distribution. Since the distribution captured by the local model is aligned with the local data distribution, we utilize it to generate high-accuracy pseudo-labels of the unlabeled dataset for global model training. To prevent the local model from severely overfitting the small number of local labeled data, we further use the exponential moving average and apply the global model to generate pseudo-labels for local modeling training. Experiments on four datasets show the effectiveness of FedLoKe. Our code is available at: https://github.com/zcfinal/FedLoKe.

The pre-trained model (PTM) is revolutionizing Artificial Intelligence (AI) technology. However, the hardware requirement of PTM training is prohibitively high, making it a game for a small proportion of people. Therefore, we proposed PatrickStar system to lower the hardware requirements of PTMs and make them accessible to everyone. PatrickStar uses the CPU-GPU heterogeneous memory space to store the model data. Different from existing works, we organize the model data in memory chunks and dynamically distribute them in the heterogeneous memory. Guided by the runtime memory statistics collected in a warm-up iteration, chunks are orchestrated efficiently in heterogeneous memory and generate lower CPU-GPU data transmission volume and higher bandwidth utilization. Symbiosis with the Zero Redundancy Optimizer, PatrickStar scales to multiple GPUs on multiple nodes. The system can train tasks on bigger models and larger batch sizes, which cannot be accomplished by existing works. Experimental results show that PatrickStar extends model scales 2.27 and 2.5 times of DeepSpeed, and exhibits significantly higher execution speed. PatricStar also successfully runs the 175B GPT3 training task on a 32 GPU cluster. Our code is available athttps://github.com/Tencent/PatrickStar.

News recommendation is a widely adopted technique to provide personalized news feeds for the user.Recently, pre-trained language models (PLMs) have demonstrated the great capability of natural language understanding and benefited news recommendation via improving news modeling.However, most existing works simply finetune the PLM with the news recommendation task, which may suffer from the known domain shift problem between the pre-training corpus and downstream news texts.Moreover, PLMs usually contain a large volume of parameters and have high computational overhead, which imposes a great burden on low-latency online services.In this paper, we propose Tiny-NewsRec, which can improve both the effectiveness and the efficiency of PLM-based news recommendation.We first design a self-supervised domain-specific posttraining method to better adapt the general PLM to the news domain with a contrastive matching task between news titles and news bodies.We further propose a two-stage knowledge distillation method to improve the efficiency of the large PLM-based news recommendation model while maintaining its performance.Multiple teacher models originated from different time steps of our post-training procedure are used to transfer comprehensive knowledge to the student model in both its posttraining stage and finetuning stage.Extensive experiments on two real-world datasets validate the effectiveness and efficiency of our method.

Tao Qi, Fangzhao Wu, Chuhan Wu, Peiru Yang, Yang Yu, Xing Xie, Yongfeng Huang. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

The transformer is the most critical algorithm innovation of the Nature Language Processing (NLP) field in recent years. Unlike the Recurrent Neural Network (RNN) models, transformers are able to process on dimensions of sequence lengths in parallel, therefore leads to better accuracy on long sequences. However, efficient deployments of them for online services in data centers equipped with GPUs are not easy. First, more computation introduced by transformer structures makes it more challenging to meet the latency and throughput constraints of serving. Second, NLP tasks take in sentences of variable length. The variability of input dimensions brings a severe problem to efficient memory management and serving optimization.