Time series forecasting has witnessed significant advancements through deep learning techniques. However, most existing methods struggle in non-stationary environments, where data distributions evolve over time due to concept drift. To address the challenge of non-stationarity in time series, various stabilization techniques have been proposed to mitigate temporal variations. Nonetheless, these methods operate at the instance level, assuming a homogeneous distribution across all time steps within an instance and relying on fixed statistical normalization. This limits their ability to effectively capture fine-grained distributional shifts.

Story visualization aims to create visual content, such as images and videos, that is consistent, coherent, and complete with a given story. Despite significant advances in the application of diffusion models for general text-to-image generation tasks, they still encounter difficulties when directly used to produce consistent visual content that accurately aligns with the narrative text. In this paper, we propose a novel training-free automated story visualization framework called AgentStory that can generate image illustrations based on a story synopsis provided by users. Specifically, the framework employs multiple agents empowered by Large Language Models (LLMs) to create detailed descriptions of each subject and scene in the entire story. Then, it integrates a masking mechanism with a fine-grained consistency refinement adapter to incorporate different subjects in a scene. Furthermore, it utilizes the visual understanding capabilities of multimodal LLMs to include detailed features of different subjects in the refinement adapter, thus improving the consistency of each subject across multiple scenes. Finally, we compare the AgentStory framework with state-of-the-art baselines for story visualization on the DS-500 dataset and demonstrate its superior performance in terms of subject consistency, text-image alignment, and aesthetic quality. Our code is publicly available at https://github.com/tc2000731/AgentStory.

Recent advancements in diffusion models have significantly advanced text-to-image generation, yet global text prompts alone remain insufficient for achieving fine-grained control over individual entities within an image. To address this limitation, we present EliGen, a novel framework for Entity-level controlled image Generation. Firstly, we put forward regional attention, a mechanism for diffusion transformers that requires no additional structures, seamlessly integrating entity prompts and arbitrary-shaped spatial masks. By contributing a high-quality dataset with fine-grained spatial and semantic entity-level annotations, we train EliGen to achieve robust and accurate entity-level manipulation, surpassing existing methods in both spatial precision and image quality. Additionally, we propose an inpainting fusion pipeline, extending EliGen’s capabilities to multi-entity image inpainting tasks. We further demonstrate EliGen’s flexibility by integrating it with other open-source models such as IP-Adapter, In-Context LoRA and MLLM, unlocking new creative possibilities. The source code, model, and dataset will be published.

In recent years, diffusion models have become the most popular and powerful methods in the field of image synthesis, even rivaling human artists in artistic creativity. However, the key issue currently limiting the application of diffusion models is its extremely slow generation process. Although several methods were proposed to speed up the generation process, there still exists a trade-off between efficiency and quality. In this paper, we first provide a detailed theoretical and empirical analysis of the generation process of the diffusion models based on schedulers. We transform the designing problem of schedulers into the determination of several parameters, and further transform the accelerated generation process into an expansion process of the linear subspace. Based on these analyses, we consequently propose a novel method called Optimal Linear Subspace Search (OLSS), which accelerates the generation process by searching for the optimal approximation process of the complete generation process in the linear subspaces spanned by latent variables. OLSS is able to generate high-quality images with a very small number of steps. To demonstrate the effectiveness of our method, we conduct extensive comparative experiments on open-source diffusion models. Experimental results show that with a given number of steps, OLSS can significantly improve the quality of generated images. Using an NVIDIA A100 GPU, we make it possible to generate a high-quality image by Stable Diffusion within only one second without other optimization techniques.

With the explosive growth of video and text data on the web, text-video retrieval has become a vital task for online video platforms. Recently, text-video retrieval methods based on pre-trained models have attracted a lot of attention. However, existing methods cannot effectively capture the fine-grained information in videos, and typically suffer from the hubness problem where a collection of similar videos are retrieved by a large number of different queries. In this paper, we propose Match4Match, a new text-video retrieval method based on CLIP (Contrastive Language-Image Pretraining) and graph optimization theories. To balance calculation efficiency and model accuracy, Match4Match seamlessly supports three inference modes for different application scenarios. In fast vector retrieval mode, we embed texts and videos in the same space and employ a vector retrieval engine to obtain the top K videos. In fine-grained alignment mode, our method fully utilizes the pre-trained knowledge of the CLIP model to align words with corresponding video frames, and uses the fine-grained information to compute text-video similarity more accurately. In flow-style matching mode, to alleviate the detrimental impact of the hubness problem, we model the retrieval problem as a combinatorial optimization problem and solve it using maximum flow with minimum cost algorithm. To demonstrate the effectiveness of our method, we conduct experiments on five public text-video datasets. The overall performance of our proposed method outperforms state-of-the-art methods. Additionally, we evaluate the computational efficiency of Match4Match. Benefiting from the three flexible inference modes, Match4Match can respond to a large number of query requests with low latency or achieve high recall with acceptable time consumption.

Algorithmic trading is a crucial yet challenging task in the financial domain, where trading decisions are made sequentially from milliseconds to days based on the historical price movements and trading frequency. To model such a sequential decision making process in the dynamic financial markets, Deep Reinforcement Learning (DRL) based methods have been applied and demonstrated their success in finding trading strategies that achieve profitable returns. However, the financial markets are complex imperfect information games with high-level of noise and uncertainties which usually make the exploration policy of DRL less effective. In this paper, we propose an end-to-end DRL method that explores solutions on the whole graph via a probabilistic dynamic programming algorithm. Specifically, we separate the state into environment state and position state, and model the position state transition as a directed acyclic graph. To obtain reliable gradients for model training, we adopt a probabilistic dynamic programming algorithm to explore solutions over the whole graph instead of sampling a path. By avoiding the sampling procedure, we propose an efficient training algorithm and overcome the efficiency problem in most existing DRL methods. Furthermore, our method is compatible with most recurrent neural network architecture, which makes our method easy to implement and very effective in practice. Extensive experiments have been conducted on two real-world stock datasets. Experimental results demonstrate that our method can generate stable trading strategies for both high-frequency and low-frequency trading, significantly outperforming the baseline DRL methods on annualized return and Sharpe ratio.