Scaling the effective context length is essential for advancing large language models (LLMs) toward artificial general intelligence (AGI). However, the quadratic increase in computational complexity inherent in traditional attention mechanisms presents a prohibitive overhead. Existing approaches either impose strongly biased structures, such as sink or window attention which are task-specific, or radically modify the attention mechanism into linear approximations, whose performance in complex reasoning tasks remains inadequately explored. In this work, we propose a solution that adheres to the ``less structure'' principle, allowing the model to determine where to attend autonomously, rather than introducing predefined biases. We introduce Mixture of Block Attention (MoBA), an innovative approach that applies the principles of Mixture of Experts (MoE) to the attention mechanism. This novel architecture demonstrates superior performance on long-context tasks while offering a key advantage: the ability to seamlessly transition between full and sparse attention, enhancing efficiency without the risk of compromising performance. MoBA has already been deployed to handle actual production workloads with long-context requirements, demonstrating significant advancements in efficient attention computation for LLMs. Our code is available at https://github.com/MoonshotAI/MoBA.

Multi-types of user behavior data (e.g., clicking, adding to cart, and purchasing) are recorded in most real-world recommendation scenarios, which can help to learn users’ multi-faceted preferences. However, it is challenging to explore multi-behavior data due to the unbalanced data distribution and sparse target behavior, which lead to the inadequate modeling of high-order relations when treating multi-behavior data “as features” and gradient conflict in multi-task learning when treating multi-behavior data “as labels”. In this paper, we propose CIGF, a Compressed Interaction Graph based Framework, to overcome the above limitations. Specifically, we design a novel Compressed Interaction Graph Convolution Network (CIGCN) to model instance-level high-order relations explicitly. To alleviate the potential gradient conflict when treating multi-behavior data “as labels”, we propose a Multi-Expert with Separate Input (MESI) network with separate input on the top of CIGCN for multi-task learning. Comprehensive experiments on three large-scale real-world datasets demonstrate the superiority of CIGF.

Sequential recommendation (SR) plays an important role in personalized recommender systems because it captures dynamic and diverse preferences from users' real-time increasing behaviors. Unlike the standard autoregressive training strategy, future data (also available during training) has been used to facilitate model training as it provides richer signals about users' current interests and can be used to improve the recommendation quality. However, existing methods suffer from a severe training-inference gap, i.e., both past and future contexts are modeled by the same encoder when training, while only historical behaviors are available during inference. This discrepancy leads to potential performance degradation. To alleviate the training-inference gap, we propose a new framework DualRec, which achieves past-future disentanglement and past-future mutual enhancement by a novel dual network. Specifically, a dual network structure is exploited to model the past and future context separately.And a bi-directional knowledge transferring mechanism enhances the knowledge learnt by the dual network. Extensive experiments on four real-world datasets demonstrate the superiority of our approach over baseline methods. Besides, we demonstrate the compatibility of DualRec by instantiating using different backbones. Further empirical analysis verifies the high utility of modeling future contexts under our DualRec framework.

In most real-world recommender systems, users interact with items in a sequential and multi-behavioral manner. Exploring the fine-grained relationship of items behind the users' multi-behavior interactions is critical in improving the performance of recommender systems. Despite the great successes, existing methods seem to have limitations on modelling heterogeneous item-level multi-behavior dependencies, capturing diverse multi-behavior sequential dynamics, or alleviating data sparsity problems. In this paper, we show it is possible to derive a framework to address all the above three limitations. The proposed framework MB-STR, a Multi-Behavior Sequential Transformer Recommender, is equipped with the multi-behavior transformer layer (MB-Trans), the multi-behavior sequential pattern generator (MB-SPG) and the behavior-aware prediction module (BA-Pred). Compared with a typical transformer, we design MB-Trans to capture multi-behavior heterogeneous dependencies as well as behavior-specific semantics, propose MB-SPG to encode the diverse sequential patterns among multiple behaviors, and incorporate BA-Pred to better leverage multi-behavior supervision. Comprehensive experiments on three real-world datasets show the effectiveness of MB-STR by significantly boosting the recommendation performance compared with various competitive baselines. Further ablation studies demonstrate the superiority of different modules of MB-STR.