Whole-body multimodal motion generation, controlled by text, speech, or music, has numerous applications including video generation and character animation. However, employing a unified model to process different condition modalities presents two main challenges: motion distribution drifts across different tasks (e.g., co-speech gestures and text-driven daily actions) and the complex optimization of mixed conditions with varying granularities (e.g., text and audio). In this paper, we propose MotionCraft, a unified diffusion transformer that crafts whole-body motion with plug-and-play multimodal control. Our framework employs a coarse-to-fine training strategy, starting with the text-to-motion semantic pre-training, followed by the multimodal low-level control adaptation. To effectively learn and transfer motion knowledge across different distributions, we design MC-Attn for parallel modeling of static and dynamic human topology graphs. To overcome the motion format inconsistency of existing benchmarks, we introduce MC-Bench, the first available multimodal whole-body motion generation benchmark based on the unified SMPL-X format. Extensive experiments show that MotionCraft achieves state-of-the-art performance on various standard motion generation tasks.

Portfolio optimization is a significant concern in finance. Existing research on portfolio optimization fails to adequately learn from the long and short-term relationships among equities, which inevitably leads to suboptimal performance. In this paper, we propose a Dynamic Graph-based Deep Reinforcement Learning (DGDRL) for optimal portfolio decisions. We achieve this goal by devising two mechanisms for naturally modeling the financial market. Firstly, we utilize the static and dynamic graphs to represent the long and short-term relations, which are then naturally represented by the proposed multi-channel graph attention neural network. Secondly, compared with the traditional two-phase approach, forecasting equity's trend and then weighting them by combinatorial optimization, we naturally optimize the portfolio decisions, which could directly guide the model to converge to optimal rewards. Through extensive experiments on three real-world datasets, we have demonstrated that our method significantly outperforms state-of-the-art benchmark methods in portfolio management. Furthermore, the evaluation of the industrial trading system has shown the applicability of our model to real-world financial markets.

Text-guided diffusion models have revolutionized image generation and editing, offering exceptional realism and diversity. Specifically, in the context of diffusion-based editing, where a source image is edited according to a target prompt, the process commences by acquiring a noisy latent vector corresponding to the source image via the diffusion model. This vector is subsequently fed into separate source and target diffusion branches for editing. The accuracy of this inversion process significantly impacts the final editing outcome, influencing both essential content preservation of the source image and edit fidelity according to the target prompt. Prior inversion techniques aimed at finding a unified solution in both the source and target diffusion branches. However, our theoretical and empirical analyses reveal that disentangling these branches leads to a distinct separation of responsibilities for preserving essential content and ensuring edit fidelity. Building on this insight, we introduce “PnP Inversion,” a novel technique achieving optimal performance of both branches with just three lines of code. To assess image editing performance, we present PIE-Bench, an editing benchmark with 700 images showcasing diverse scenes and editing types, accompanied by versatile annotations and comprehensive evaluation metrics. Compared to state-of-the-art optimization-based inversion techniques, our solution not only yields superior performance across 8 editing methods but also achieves nearly an order of speed-up.

In this study, we present $\text{aL\small{LM}4T\small{S}}$, an innovative framework that adapts Large Language Models (LLMs) for time-series representation learning. Central to our approach is that we reconceive time-series forecasting as a self-supervised, multi-patch prediction task, which, compared to traditional mask-and-reconstruction methods, captures temporal dynamics in patch representations more effectively. Our strategy encompasses two-stage training: (i). a causal continual pre-training phase on various time-series datasets, anchored on next patch prediction, effectively syncing LLM capabilities with the intricacies of time-series data; (ii). fine-tuning for multi-patch prediction in the targeted time-series context. A distinctive element of our framework is the patch-wise decoding layer, which departs from previous methods reliant on sequence-level decoding. Such a design directly transposes individual patches into temporal sequences, thereby significantly bolstering the model’s proficiency in mastering temporal patch-based representations. $\text{aL\small{LM}4T\small{S}}$ demonstrates superior performance in several downstream tasks, proving its effectiveness in deriving temporal representations with enhanced transferability and marking a pivotal advancement in the adaptation of LLMs for time-series analysis.

The distributions of cloud phases play an important role in influencing the weather and climate system. The characteristics of clouds above the Tibetan Plateau (TP) can profoundly affect regional and global atmospheric circulation. To research the distributions of cloud phases in the TP region, a retrieval algorithm was developed based on the combination of polarization lidar and millimeter cloud radar measurements and applied to the data from a comprehensive field campaign on the central TP in the summer of 2014. The structure and phase of four different types of clouds were retrieved accordingly, which validates the reliability of the algorithm. The result shows that the occurrence frequency of low clouds remains around 50%, which is very high throughout the whole day in Nagqu, Tibetan in summer. The liquid and mixed cloud frequencies are higher in the morning and afternoon, while ice cloud mainly occurs from the afternoon to midnight. Liquid and ice phase distributions show an inverse relationship in the atmospheric layer from 2 to 8 km in height. Meanwhile, the proportion of the liquid phase to the cloud top is significantly higher than that to the cloud body, which indicates that the supercooled water is more likely to appear at the cloud top than in the cloud. The fractional probabilities of the ice phase and liquid phase in the total cloud top phase intersect at about$-26.7\,\,^{\circ }\text{C}$.