Recently, some works have tried to combine diffusion and Generative Adversarial Networks (GANs) to alleviate the computational cost of the iterative denoising inference in Diffusion Models (DMs). However, existing works in this line suffer from either training instability and mode collapse or subpar one-step generation learning efficiency. To address these issues, we introduce YOSO, a novel generative model designed for rapid, scalable, and high-fidelity one-step image synthesis with high training stability and mode coverage. Specifically, we smooth the adversarial divergence by the denoising generator itself, performing self-cooperative learning. We show that our method can serve as a one-step generation model training from scratch with competitive performance. Moreover, we extend our YOSO to one-step text-to-image generation based on pre-trained models by several effective training techniques (i.e., latent perceptual loss and latent discriminator for efficient training along with the latent DMs; the informative prior initialization (IPI), and the quick adaption stage for fixing the flawed noise scheduler). Experimental results show that YOSO achieves the state-of-the-art one-step generation performance even with Low-Rank Adaptation (LoRA) fine-tuning. In particular, we show that the YOSO-PixArt-$\alpha$ can generate images in one step trained on 512 resolution, with the capability of adapting to 1024 resolution without extra explicit training, requiring only \textasciitilde10 A800 days for fine-tuning. Our code is available at: [https://github.com/Luo-Yihong/YOSO](https://github.com/Luo-Yihong/YOSO)

The rise of link recommendation systems in online social networks has sparked significant research interest in strategically adding links to enhance social influence. This paper delves into the influence maximization with augmentation (IMA) problem that aims to add k edges connecting seed nodes and ordinary nodes to boost the influence propagation of the given seed set. IMA is a monotone submodular maximization problem so that the greedy algorithm provides a (1-1/e-ε)-approximate solution, where ε is an error term caused by the intractable nature of influence spread computation. Previous work often utilizes an unbiased estimator that relies on the chosen edges for influence estimation, resulting in non-submodular estimate with respect to edge selection. To ensure the overall error being bounded by ε, such an estimator requires Θ(ε/k) multiplicative error for each estimation, incurring prohibitive overhead. Meanwhile, some other work approximates IMA via conventional influence maximization (IM) on an augmented graph by adding a new node for every edge candidate, leading to heavy extra sampling due to a significant increase in graph size. To address these challenges, we design a novel unbiased estimator on the original graph that is independent of the chosen edges by leveraging the tractability of one-hop influence computation. We show that the estimate via our estimator is submodular so that it enables the estimate of all k edges in a whole with a bounded estimation error of Θ(ε), saving O(k2) time compared to the chosen-edge-dependent estimator while retaining the same graph size. Moreover, we propose several techniques based on the properties of our estimator to further speed up the greedy selection. Putting it together, we develop a scalable algorithm for the IMA problem, namely ScaLIM. Finally, extensive experiments are conducted to validate the effectiveness and efficiency of our proposed approach, e.g., ScaLIM is faster than baselines by nearly two orders of magnitude.

Link recommendation systems in online social networks (OSNs), such as Facebook's "People You May Know", Twitter's "Who to Follow", and Instagram's "Suggested Accounts", facilitate the formation of new connections among users. This paper addresses the challenge of link recommendation for the purpose of social influence maximization. In particular, given a graph G and the seed set S, our objective is to select k edges that connect seed nodes and ordinary nodes to optimize the influence dissemination of the seed set. This problem, referred to as influence maximization with augmentation (IMA), has been proven to be NP-hard.

Graph embedding aims at mapping each node to a low-dimensional vector, beneficial for various applications like pattern matching, retrieval augmented generation and recommendation. In this paper, we study the large-scale temporal graph embedding problem. Different from simple graphs, each edge has a timestamp in temporal graphs, which requires the embeddings to encode the temporal biases. Factorizing similarity matrix is a common approach for generating simple graph embeddings where similarity can be well characterized by some conventional metrics like personalized PageRank. However, how to construct a similarity that can encode interactions with temporal biases is a critical problem for large scale temporal graphs. To address this, we introduce the concept of temporal-based bipartite graph (TBG) and develop the temporal preferential attachment similarity (TPASim) that reflects concurrent node activity over time. Directly factorizing the TPASim matrix, which contains nearly n 2 non-zeros, is not feasible for large graphs with n nodes. Instead, we present LTGE, which constructs and factorizes a temporal matrix with at most 2 m non-zeros, where m is the number of edges. Our theoretical analysis shows that LTGE achieves the same embeddings as factorizing the TPASim matrix but significantly reduces complexity by a factor of n 2 / m. On the other hand, when graphs evolve over time, to avoid recomputing, we further propose LTGEInc that utilizes a novel incremental singular value decomposition (SVD) algorithm with provable guarantee for updating the embeddings. Extensive experiments on several datasets with up to 17 million nodes and 1.3 billion edges demonstrate that LTGE outperforms the state of the art significantly and is orders of magnitude faster than the baselines specially designed for temporal graphs. For embeddings update, LTGEInc retains the performance with small computational overhead.