Search

State-of-the-art sequential recommendation models heavily rely on transformer's attention mechanism. However, the quadratic computational and memory complexities of self attention have limited its scalability for modeling users' long range behaviour sequences. To address this problem, we propose ELASTIC, an Efficient Linear Attention for SequenTial Interest Compression, requiring only linear time complexity and decoupling model capacity from computational cost. Specifically, ELASTIC introduces a fixed length interest experts with linear dispatcher attention mechanism which compresses the long-term behaviour sequences to a significantly more compact representation which reduces up to 90% GPU memory usage with x2.7 inference speed up. The proposed linear dispatcher attention mechanism significantly reduces the quadratic complexity and makes the model feasible for adequately modeling extremely long sequences. Moreover, in order to retain the capacity for modeling various user interests, ELASTIC initializes a vast learnable interest memory bank and sparsely retrieves compressed user's interests from the memory with a negligible computational overhead. The proposed interest memory retrieval technique significantly expands the cardinality of available interest space while keeping the same computational cost, thereby striking a trade-off between recommendation accuracy and efficiency. To validate the effectiveness of our proposed ELASTIC, we conduct extensive experiments on various public datasets and compare it with several strong sequential recommenders. Experimental results demonstrate that ELASTIC consistently outperforms baselines by a significant margin and also highlight the computational efficiency of ELASTIC when modeling long sequences. We will make our implementation code publicly available., Comment: We hereby withdraw this paper from arXiv due to incomplete experiments. Upon further review, we have determined that additional experimental work is necessary to fully validate our findings and conclusions

Classical logics for strategic reasoning, such as Coalition Logic and Alternating-time Temporal Logic, formalize absolute strategic reasoning about the unconditional strategic abilities of agents to achieve their goals. Goranko and Ju introduced a logic ConStR for strategic reasoning about conditional strategic abilities. However, its completeness is still an open problem. ConStR has three featured operators, and one of them has the following reading: For some action of A that guarantees the achievement of her goal, B has an action to guarantee the achievement of his goal. The logic about this operator is called CConStR. In this paper, we prove completeness for two fragments of CConStR. The key notions of our proof approach include downward validity lemma, grafted models, and upward derivability lemma. The proof approach has good potential to be applied to the completeness of ConStR and other logics.

Coalition logic is a central logic in strategic reasoning studies. In this paper, we first argue that Coalition Logic models, concurrent game models, have three too-strong assumptions. The first one is the independence of agents; that is, the merge of two available joint actions of two disjoint coalitions is always available for the union of the two coalitions. The second one is seriality; that is, coalitions always have available joint actions. The third one is determinism, that is, the grand coalition's joint actions always have a unique outcome. Second, we present a coalition logic based on general concurrent game models, which do not have the three assumptions. We show the completeness of this logic and compare it with Coalition Logic in detail. This logic seems minimal in the context of strategic reasoning.

Modern personalized recommendation services often rely on user feedback, either explicit or implicit, to improve the quality of services. Explicit feedback refers to behaviors like ratings, while implicit feedback refers to behaviors like user clicks. However, in the scenario of full-screen video viewing experiences like Tiktok and Reels, the click action is absent, resulting in unclear feedback from users, hence introducing noises in modeling training. Existing approaches on de-noising recommendation mainly focus on positive instances while ignoring the noise in a large amount of sampled negative feedback. In this paper, we propose a meta-learning method to annotate the unlabeled data from loss and gradient perspectives, which considers the noises in both positive and negative instances. Specifically, we first propose an Inverse Dual Loss (IDL) to boost the true label learning and prevent the false label learning. Then we further propose an Inverse Gradient (IG) method to explore the correct updating gradient and adjust the updating based on meta-learning. Finally, we conduct extensive experiments on both benchmark and industrial datasets where our proposed method can significantly improve AUC by 9.25% against state-of-the-art methods. Further analysis verifies the proposed inverse learning framework is model-agnostic and can improve a variety of recommendation backbones. The source code, along with the best hyper-parameter settings, is available at this link: https://github.com/Guanyu-Lin/InverseLearning., Comment: WSDM 2024

With the outbreak of today's streaming data, the sequential recommendation is a promising solution to achieve time-aware personalized modeling. It aims to infer the next interacted item of a given user based on the historical item sequence. Some recent works tend to improve the sequential recommendation via random masking on the historical item so as to generate self-supervised signals. But such approaches will indeed result in sparser item sequence and unreliable signals. Besides, the existing sequential recommendation models are only user-centric, i.e., based on the historical items by chronological order to predict the probability of candidate items, which ignores whether the items from a provider can be successfully recommended. Such user-centric recommendation will make it impossible for the provider to expose their new items, failing to consider the accordant interactions between user and item dimensions. In this paper, we propose a novel Dual Contrastive Network (DCN) to achieve mutual harmony between user and item provider, generating ground-truth self-supervised signals for sequential recommendation by auxiliary user-sequence from an item-centric dimension. Specifically, we propose dual representation contrastive learning to refine the representation learning by minimizing the Euclidean distance between the representations of a given user/item and historical items/users of them. Before the second contrastive learning module, we perform the next user prediction to capture the trends of items preferred by certain types of users and provide personalized exploration opportunities for item providers. Finally, we further propose dual interest contrastive learning to self-supervise the dynamic interest from the next item/user prediction and static interest of matching probability. Experiments on four benchmark datasets verify the effectiveness of our proposed method., Comment: 25 pages

Spatiotemporal activity prediction, aiming to predict user activities at a specific location and time, is crucial for applications like urban planning and mobile advertising. Existing solutions based on tensor decomposition or graph embedding suffer from the following two major limitations: 1) ignoring the fine-grained similarities of user preferences; 2) user's modeling is entangled. In this work, we propose a hypergraph neural network model called DisenHCN to bridge the above gaps. In particular, we first unify the fine-grained user similarity and the complex matching between user preferences and spatiotemporal activity into a heterogeneous hypergraph. We then disentangle the user representations into different aspects (location-aware, time-aware, and activity-aware) and aggregate corresponding aspect's features on the constructed hypergraph, capturing high-order relations from different aspects and disentangles the impact of each aspect for final prediction. Extensive experiments show that our DisenHCN outperforms the state-of-the-art methods by 14.23% to 18.10% on four real-world datasets. Further studies also convincingly verify the rationality of each component in our DisenHCN., Comment: Accepted as a research track paper in ICDE 2022

Recommender system is one of the most important information services on today's Internet. Recently, graph neural networks have become the new state-of-the-art approach to recommender systems. In this survey, we conduct a comprehensive review of the literature on graph neural network-based recommender systems. We first introduce the background and the history of the development of both recommender systems and graph neural networks. For recommender systems, in general, there are four aspects for categorizing existing works: stage, scenario, objective, and application. For graph neural networks, the existing methods consist of two categories, spectral models and spatial ones. We then discuss the motivation of applying graph neural networks into recommender systems, mainly consisting of the high-order connectivity, the structural property of data, and the enhanced supervision signal. We then systematically analyze the challenges in graph construction, embedding propagation/aggregation, model optimization, and computation efficiency. Afterward and primarily, we provide a comprehensive overview of a multitude of existing works of graph neural network-based recommender systems, following the taxonomy above. Finally, we raise discussions on the open problems and promising future directions in this area. We summarize the representative papers along with their code repositories in \url{https://github.com/tsinghua-fib-lab/GNN-Recommender-Systems}., Comment: accepted by ACM Transactions on Recommender Systems

Molecular dynamics (MD) simulations have been performed to investigate the mechanical properties of hydrogen functionalized graphene allotropes (GAs) for H-coverage spanning the entire range (0-100%). Four allotropes (graphyne, cyclic graphene, octagonal graphene, and biphenylene) with larger unit lattice size than graphene are considered. The effect of the degree of functionalization and molecular structure on the Young's modulus and strength are investigated, and the failure processes of some new GAs are reported for the first time. We show that the mechanical properties of the hydrogenated GAs deteriorate drastically with increasing H-coverage within the sensitive threshold, beyond which the mechanical properties remain insensitive to the increase in H-coverage. This drastic deterioration arises both from the conversion of sp2 to sp3 bonding and easy rotation of unsupported sp3 bonds. Allotropes with different lattice structures correspond to different sensitive thresholds. The Young's moduli deterioration of fully hydrogenated allotropes can be up to 70% smaller than that of the corresponding pristine structure. Moreover the tensile strength shows an even larger drop of about 90% and higher sensitivity to H-coverage even if it is small. Our results suggest that the unique coverage-dependent deterioration of the mechanical properties must be taken into account when analyzing the performance characteristics of nanodevices fabricated from functionalized GAs., Comment: 11 pages, 4 figures