Single-cell RNA sequencing (scRNA-seq), especially temporally resolved datasets, enables genome-wide profiling of gene expression dynamics at single-cell resolution across discrete time points. However, current technologies provide only sparse, static snapshots of cell states and are inherently influenced by technical noise, complicating the inference and representation of continuous transcriptional dynamics. Although embedding methods can reduce dimensionality and mitigate technical noise, the majority of existing approaches typically treat trajectory inference separately from embedding construction, often neglecting temporal structure. To address this challenge, here we introduce CellStream, a novel deep learning framework that jointly learns embedding and cellular dynamics from single-cell snapshots data by integrating an autoencoder with unbalanced dynamical optimal transport. Compared to existing methods, CellStream generates dynamics-informed embeddings that robustly capture temporal developmental processes while maintaining high consistency with the underlying data manifold. We demonstrate CellStream’s effectiveness on both simulated datasets and real scRNA-seq data, including spatial transcriptomics. Our experiments indicate significant quantitative improvements over state-of-the-art methods in representing cellular trajectories with enhanced temporal coherence and reduced noise sensitivity. Overall, CellStream provides a new tool for learning and representing continuous streams from the noisy, static snapshots of single-cell gene expression.

Multi-vehicle collaborative sensing has emerged as a new paradigm to boost the safety of autonomous vehicles. The cornerstone of this vision is the real-time and accurate registration of mmWave radar point clouds among multiple vehicles. To accomplish this, we designmmReg, an innovative system capable of achieving centimeter-level and real-time sensing fusion between vehicles.mmRegconsists of three major components: (i) aSAR imaging-driven point cloud generationcomponent leverages SAR imaging to image sparse and disordered radar point clouds to generate high-quality point clouds; (ii) amotion-aware frame synchronizationcomponent can achieve the spatio-temporal alignment of point clouds between vehicles for effectively mitigating the impact of asynchronous radar frames; (iii) ashared object-based registrationcomponent can capture and understand the unique global position of shared objects, supporting real-time and accurate registration. We implement and evaluatemmRegon CARLA and real-world campus datasets. The results demonstrate thatmmRegcan improve the vehicle’s sensing range by 117% in an average of 99.91 ms, achieving a 4.82x improvement in accuracy.

Millimeter wave radar has recently emerged as a promising modality for enabling pervasive gesture recognition while protecting user privacy. However, personalized user behaviors, interference from unexpected actions, and long-term variability in user gestures significantly degrade the accuracy of gesture and user identity recognition, thereby compromising the quality of user experiences. To this end, we designmmGeswith four key modules: (i) afine-grained feature extractorextracts micro-level features from the time-series radar data to identify users' personalized behaviors; (ii) auser-specific feature classifierextracts coarse-grained features from a global perspective, followed by analyzing the micro-details of gesture features to recognize the user; (iii) avoting-based multi-user recognizerretrieves all pre-trained models from the user model database, followed by obtaining the probability indicators of each to return recognition results; (iv) alifelong learning modelretains previous knowledge while adjusting its feature selection capabilities using newly collected data to adapt to gesture changes. We implement and evaluatemmGesusing three self-collected real-world radar datasets, demonstrating its superior performance compared to other state-of-the-art gesture recognition methods.

Millimeter-wave (mmWave) radar enables privacy-preserving gesture recognition but suffers from limited training data, particularly for lying postures. Existing mmWave radar data generation methods are ineffective due to insufficient 2D video data. To this end, we design a novel system named Venus to generate realistic radar data for lying postures using few 2D videos, which addresses two key challenges including i) the simulation of diverse reflected signals and ii) few real-world data leading to low data fidelity. Venus consists of two key components: (i) a gesture sequence generation and signal simulation network, which combines several key modules, movement information extractor, spatio-temporal latent diffusion model, and mmWave signal simulator, to generate diverse gesture vertex sequences under certain conditions and simulate signal propagation characteristics to obtain coarse radar data; (ii) a meta-learning domain adaption network generates realistic radar data with few real-world data via ''meta-learning'' strategy. Extensive experiments on both generated and self-collected datasets demonstrate that Venus significantly outperforms state-of-the-art methods in recognizing gestures performed in lying postures.

Infrastructure-assisted autonomous driving has become a new paradigm that enables autonomous vehicles to fuse sensor data and improve driving safety, where a key enabling technology for achieving this vision is to real-time and accurate registering 3-D mmWave radar point clouds between the infrastructure and the vehicle. To this end, we proposeArtemis, a novel lightweight system capable of achieving real-time registration with decimeter-level localization.Artemisconsists of three components: 1) a modal association-based salient object extraction component leverages the complementary advantages of cameras and radars to extract semantics and areas of salient objects for radar point clouds; 2) a salient object shape construction component extracts the shape contour of salient objects based on their inherent geometries; and 3) a contour-guided 3-D point cloud registration component combines two key strategies, keypoint matching strategy and early exit strategy, to quickly select keypoints and transformation directions for achieving accurate registration in real-time. We implement and evaluateArtemiswith two multiview datasets collected in the CARLA platform and campus. The experiment results show thatArtemisachieves an average registration error of 0.33 m within 32.26 ms.

Millimeter wave radar is gaining traction recently as a promising modality for enabling pervasive and privacy-preserving gesture recognition. However, the lack of rich and fine-grained radar datasets hinders progress in developing generalized deep learning models for gesture recognition across various user postures (e.g., standing, sitting), positions, and scenes. To remedy this, we resort to designing a software pipeline that exploits wealthy 2D videos to generate realistic radar data, but it needs to address the challenge of simulating diversified and fine-grained reflection properties of user gestures. To this end, we designG3Rwith three key components: i) agesture reflection point generatorexpands the arm's skeleton points to form human reflection points; ii) asignal simulation modelsimulates the multipath reflection and attenuation of radar signals to output the human intensity map; iii) anencoder-decoder modelcombines asampling moduleand afitting moduleto address the differences in number and distribution of points between generated and real-world radar data for generating realistic radar data. We implement and evaluateG3Rusing 2D videos from public data sources and self-collected real-world radar data, demonstrating its superiority over other state-of-the-art approaches for gesture recognition.

Knowledge Graph (KG) always suffers from incompleteness. Knowledge Graph Reasoning (KGR) aims to predict the unknown entity or find reasoning paths for relations over incomplete KG. However, multi-hop reasoning still facing challenges, because the process of reasoning usually experiences the neighbor information issue. Prior works just use path efficiency as a part of reward function and do not utilize the neighbor information effectively. In order to deal with the situation, we propose the length reward in the Reinforcement Learning (RL) framework to represent the length of reasoning paths and use the neighbor information effectively. Our model utilizes the position information to design reward function. To solve this problem of save the semantic information of entity neighbors and historical trajectory information, we propose a new GRU-GAT framework to capture neighbor feature of the current entity and the target entity. Experimental results on NELL-995 and FB15K237 demonstrate the effectiveness of our model and our model can identify a more balanced route for every relation.

The existing deep learning fine-grained recognition models have low recognition accuracy (Acc) in complex natural environments. We propose a fine-grained ship recognition model [global to local progressive learning module (GLPM)] based on global to local and progressive learning for complex backgrounds to address this problem. GLPM improves fine-grained ship targets’ recognition Acc in natural complex backgrounds by progressively guiding high-level global features to low-level local features for learning and enhancing local fine-grained features expression ability. First, the global features are obtained by adding the feature pyramid attention (FPA) module to the deepest level of the backbone network. Second, the decoding and reconstruction of low-level local feature information at different levels are guided, and the global features are used to weight the key pixel values of local features of the target. Finally, the output of the backbone network is combined with the output vector of the global to local progressive module for classification, and the whole model can be trained end-to-end. We experiment on naturally collected complex background images containing 59 ship classes, and the public maritime (MAR)-ships contain pictures of 23 types of ships. The results show that the proposed method outperforms the methods compared in this letter while providing valuable ideas for the research of fine-grained ship recognition models in complex natural environments.