Multimodal generative models can process instructions in various modalities and demonstrate outstanding performance across a wide range of image generation tasks. However, their robustness in complex real-world scenarios remains limited due to insufficient generalized instruction alignment. We introduces \textbf{OmniGen2}, a unified multimodal generator designed to follow complex, fine-grained instructions. Our core contribution is a two-stage design that first builds a strong, world-knowledge-grounded foundation model and then aligns it using a progressive, multi-task instruction tuning strategy. The foundation model features a streamlined architecture with decoupled decoding for versatile multimodal generation and a novel positional encoding scheme to improve learning efficiency. We ground this model in real-world knowledge using large-scale data construction pipelines. Building on this foundation, we propose a progressive, reinforcement-based alignment process. This phase carefully schedules training tasks and reward signals to foster cross-task knowledge transfer, significantly improving the model's instruction-following capabilities. Our models demonstrate competitive performance on standard benchmarks and our dedicated in-context generation benchmark, \textbf{OmniContext}. We will release our models, code, benchmark, and training datasets to catalyze future research in building more capable and instruction aligned generative models.

Cardiac magnetic resonance (CMR) is a cornerstone for diagnosing cardiovascular disease. However, it remains underutilized due to complex, time-consuming interpretation across multi-sequences, phases, quantitative measures that heavily reliant on specialized expertise. Here, we present BAAI Cardiac Agent, a multimodal intelligent system designed for end-to-end CMR interpretation. The agent integrates specialized cardiac expert models to perform automated segmentation of cardiac structures, functional quantification, tissue characterization and disease diagnosis, and generates structured clinical reports within a unified workflow. Evaluated on CMR datasets from two hospitals (2413 patients) spanning 7-types of major cardiovascular diseases, the agent achieved an area under the receiver-operating-characteristic curve exceeding 0.93 internally and 0.81 externally. In the task of estimating left ventricular function indices, the results generated by this system for core parameters such as ejection fraction, stroke volume, and left ventricular mass are highly consistent with clinical reports, with Pearson correlation coefficients all exceeding 0.90. The agent outperformed state-of-the-art models in segmentation and diagnostic tasks, and generated clinical reports showing high concordance with expert radiologists (six readers across three experience levels). By dynamically orchestrating expert models for coordinated multimodal analysis, this agent framework enables accurate, efficient CMR interpretation and highlights its potentials for complex clinical imaging workflows. Code is available at https://github.com/plantain-herb/Cardiac-Agent.

UMI-style interfaces enable scalable robot learning, but existing systems remain largely visuomotor, relying primarily on RGB observations and trajectory while providing only limited access to physical interaction signals. This becomes a fundamental limitation in contact-rich manipulation, where success depends on contact dynamics such as tactile interaction, internal grasping force, and external interaction wrench that are difficult to infer from vision alone. We present OmniUMI, a unified framework for physically grounded robot learning via human-aligned multimodal interaction. OmniUMI synchronously captures RGB, depth, trajectory, tactile sensing, internal grasping force, and external interaction wrench within a compact handheld system, while maintaining collection–deployment consistency through a shared embodiment design. To support human-aligned demonstration, OmniUMI provides dual-force feedback through bilateral gripper feedback and natural perception of external interaction wrench in the handheld embodiment. Built on this interface, we extend diffusion policy with visual, tactile, and force-related observations, and deploy the learned policy through impedance-based execution for unified regulation of motion and contact behavior. Experiments demonstrate reliable sensing and strong downstream performance on force-sensitive pick-and-place, interactive surface erasing, and tactile-informed selective release. Overall, OmniUMI combines physically grounded multimodal data acquisition with human-aligned interaction, providing a scalable foundation for learning contact-rich manipulation.

Spiking Neural Networks(SNNs) are a promising paradigm designed to emulate the brain's energy efficient by incorporating the timing of spikes. Conversion is an efficient way to obtain high-performance SNNs from Artificial Neural Networks(ANNs). Existing conversion methods often face a trade-off between accuracy and time steps, which is largely caused by the incomplete release of residual membrane potentials. To minimize the conversion error, this paper proposed a harmonious mathematical property-based neuron, called Harmony Multi-Threshold Neurons (H-MT Neuron), which utilizes multiple spikes to minimize residual membrane potentials. The proposed neuron is further enhanced with an optional effective communication mechanism to achieve more accurate conversion. In addition, we propose a threshold optimization method applicable to a broader range cases of spiking neurons to to find the optimal neuron thresholds. Experiment results demonstrate that our method achieve superior accuracy on ImageNet benchmark datasets while significantly reducing the required time steps and energy consumption.

High-speed vision tasks have long been a challenge in computer vision. Recently, the spike camera has shown great potential in these tasks due to its high temporal resolution. Unlike traditional cameras, it emits asynchronous spike signals to capture visual information. However, under low-light conditions, spike signals become highly sparse, and the sparse spike stream severely hinders the effectiveness of existing spike-based methods in high-speed scenarios. To address this challenge, we introduce SS2DS, the first deep learning framework that enhances sparse spike streams into dense spike streams. SS2DS first estimates the spike firing frequency within sparse streams. Subsequently, the spike firing frequency is enhanced by a neural network. Finally, SS2DS decodes the enhanced spike stream from the enhanced spike firing frequency sequence. SS2DS can adjust the temporal distribution of sparse spike streams and improve the performance degradation of existing methods in low-light and high-speed scenarios. To evaluate sparse spike stream enhancement, we construct both synthetic and real sparse spike stream datasets. By comparing the reconstruction results, enhanced spike streams achieve an average improvement of +0.78 MA, -18.42 BRISQUE, and -1.42 NIQE over sparse spike streams. Moreover, the enhanced spike streams also benefit other spike-based vision tasks, such as 3D reconstruction (+1.325 dB PSNR, +0.005 SSIM, and -0.01 LPIPS) and superresolution (+0.63 MA, -13.67 BRISQUE, and -1.28 NIQE).

Deep Neural Networks (DNNs) are vulnerable to elaborately designed adversarial noise, although they have achieved extraordinary success in many tasks. Compared with DNNs, the human visual system is highly robust. However, it is unclear how the human visual system defends against adversarial attacks, especially the role of the early visual system and its influence on the brain manifold. Due to retina gap junctions being crucial for the denoising function in the early visual system, we combine a retina gap junction-based filter, G-filter, with DNN as an abstract human visual system model called the biological hybrid model. We adopt this model to study the defense performance of retina gap junctions and their impact on the brain manifold. Compared with other defense methods, the biological hybrid model is more robust and can be further improved by introducing noise during training. Next, we analyze the manifold and its decision boundary of the biological hybrid model from a geometry perspective. The results show that the biological hybrid model has a unique 2D decision boundary with high nonlinearity and a lower curvature of the decision boundary of the manifold compared to other defense methods. The transforming manifold may account for the high robustness of the biological hybrid model. Finally, to dissect G-filter and clarify its internal mechanism, we borrow the Neural Ordinary Differential Equation (ODE) concept and rewrite G-filter into an equivalent recurrent neural network. The results show that the decision boundary of the model's manifold will gradually change with time and eventually reach a steady state, which is modulated by gap junction conductance, revealing the influence of retina gap junctions on the brain manifold is a gradually evolving process.

The convergence of neuroscience and artificial intelligence has positioned Spiking Neural Networks (SNNs) as one of the pivotal paradigms for future computing. However, the field faces a theoretical challenge: reconciling the mathematical clarity of static deep learning with the rich, non-equilibrium dynamics of biological circuits. We introduce the dynamical superspace, a framework that reimagines neural computing as a continuous hierarchy defined by temporal density and state-space complexity. We suggest that while current synchronous SNNs successfully optimize rate-based equilibria, they often neglect the intrinsic power of biological time. The true neuromorphic advantage emerges by ascending to asynchronous timing, where information is decoupled from clock cycles, and to complex non-equilibrium dynamics, where heterogeneity and criticality drive computation through transient trajectories. We propose a roadmap to bridge global optimization with local execution, leveraging evolutionary priors to support innate learning. By identifying native applications, from ultra-low-latency event perception to infinite-context memory for AGI, this perspective invites the community to view SNNs not merely as efficient quantization, but as dynamical systems capable of stable transience, offering a physical bridge to the next generation of intelligence.

Multi-object tracking (MOT) plays a fundamental role in visual perception, where accurate trajectory prediction is essential for reliable target association under complex motion patterns. Recent trackers have improved motion modeling with densely activated artificial neural networks, yet they largely overlook whether such dense responses are necessary for trajectory prediction. In this paper, we formulate activation sparsity preference (ASP) by tackling two key questions: 1. How can we identify a model architecture that appropriately and formally explains ASP, and 2. How can we translate this explanation into competitive tracking performance. Theoretical analysis shows that sparse gating is no worse than state-independent dropout under the same activation rate. Based on this insight, SpikingMOT is proposed as a spike-driven tracker that adaptively models sparse trajectory dynamics with spiking neural networks (SNNs). Specifically, SpikingMOT decomposes each trajectory state into pseudo-trajectory bases and uses the current prediction error to calibrate the posterior for next-frame prediction. With this brain-inspired loop, SpikingMOT achieves state-of-the-art performance in extensive experiments, 74.9 HOTA on SportsMOT and 56.5 HOTA on DanceTrack, while reducing the parameters and energy by 72

The process through which humans perceive and learn visual representations in dynamic environments is highly complex. From a structural perspective, the human eye decouples the functions of cone and rod cells: cones are primarily responsible for color perception, while rods are specialized in detecting motion, particularly variations in light intensity. These two distinct modalities of visual information are integrated and processed within the visual cortex, thereby enhancing the robustness of the human visual system. Inspired by this biological mechanism, modern hardware systems have evolved to include not only color-sensitive RGB cameras but also motion-sensitive Dynamic Visual Systems, such as spike cameras. Building upon these advancements, this study seeks to emulate the human visual system by integrating decomposed multi-modal visual inputs with modern latent-space generative frameworks. We named it SpikeGen. We evaluate its performance across various spike-RGB tasks, including conditional image and video deblurring, dense frame reconstruction from spike streams, and high-speed scene novel-view synthesis. Supported by extensive experiments, we demonstrate that leveraging the latent space manipulation capabilities of generative models enables an effective synergistic enhancement of different visual modalities, addressing spatial sparsity in spike inputs and temporal sparsity in RGB inputs.

Spiking Neural Networks (SNNs), with their temporal processing capabilities and biologically plausible dynamics, offer a natural platform for unsupervised representation learning. However, current unsupervised SNNs predominantly employ shallow architectures or localized plasticity rules, limiting their ability to model long-range temporal dependencies and maintain temporal feature consistency. This results in semantically unstable representations, thereby impeding the development of deep unsupervised SNNs for large-scale temporal video data. We propose PredNext, which explicitly models temporal relationships through cross-view future Step Prediction and Clip Prediction. This plug-and-play module seamlessly integrates with diverse self-supervised objectives. We firstly establish standard benchmarks for SNN self-supervised learning on UCF101, HMDB51, and MiniKinetics, which are substantially larger than conventional DVS datasets. PredNext delivers significant performance improvements across different tasks and self-supervised methods. PredNext achieves performance comparable to ImageNet-pretrained supervised weights through unsupervised training solely on UCF101. Additional experiments demonstrate that PredNext, distinct from forced consistency constraints, substantially improves temporal feature consistency while enhancing network generalization capabilities. This work provides a effective foundation for unsupervised deep SNNs on large-scale temporal video data.

Spiking neural networks (SNNs) have recently shown strong potential in unimodal visual and textual tasks, yet building a directly trained, low-energy, and high-performance SNN for multimodal applications such as image-text retrieval (ITR) remains highly challenging. Existing artificial neural network (ANN)-based methods often pursue richer unimodal semantics using deeper and more complex architectures, while overlooking cross-modal interaction, retrieval latency, and energy efficiency. To address these limitations, we present a brain-inspired Cross-Modal Spike Fusion network (CMSF) and apply it to ITR for the first time. The proposed spike fusion mechanism integrates unimodal features at the spike level, generating enhanced multimodal representations that act as soft supervisory signals to refine unimodal spike embeddings, effectively mitigating semantic loss within CMSF. Despite requiring only two time steps, CMSF achieves top-tier retrieval accuracy, surpassing state-of-the-art ANN counterparts while maintaining exceptionally low energy consumption and high retrieval speed. This work marks a significant step toward multimodal SNNs, offering a brain-inspired framework that unifies temporal dynamics with cross-modal alignment and provides new insights for future spiking-based multimodal research. The code is available at https://github.com/zxt6174/CMSF.

Spiking Neural Networks (SNNs) offer superior energy efficiency over Artificial Neural Networks (ANNs). However, they encounter significant deficiencies in training and inference metrics when applied to Spiking Vision Transformers (S-ViTs). Existing paradigms including ANN-SNN Conversion and Spatial-Temporal Backpropagation (STBP) suffer from inherent limitations, precluding concurrent optimization of memory, accuracy and energy consumption. To address these issues, we propose Ge^2mS-T, a novel architecture implementing grouped computation across temporal, spatial and network structure dimensions. Specifically, we introduce the Grouped-Exponential-Coding-based IF (ExpG-IF) model, enabling lossless conversion with constant training overhead and precise regulation for spike patterns. Additionally, we develop Group-wise Spiking Self-Attention (GW-SSA) to reduce computational complexity via multi-scale token grouping and multiplication-free operations within a hybrid attention-convolution framework. Experiments confirm that our method can achieve superior performance with ultra-high energy efficiency on challenging benchmarks. To our best knowledge, this is the first work to systematically establish multi-dimensional grouped computation for resolving the triad of memory overhead, learning capability and energy budget in S-ViTs.

As a retina-inspired sensor with ultra-high temporal resolution, spike camera can continuously capture dynamic scenes with high-speed motion. It is a key task to restore clear images from spike streams. The quantization effects in spike readout bring degradation to the visual quality of restored images. To tackle the degradation without introducing motion blur, existing methods often employ a short-term temporal window to infer the light intensity at a certain time point. However, these methods only focus on the spike signals within the current window, which limits their performance. Motivated by the human-like memory mechanism for visual signals from the retina, we explore Spike Stream Memory Transfer (SSMT) to restore the dynamic scenes, considering spike signals beyond the window. Specifically, we design a framework that leverages temporal memory by transferring previously inferred light intensity and motion to enhance current reconstruction. The framework enables a long-term temporal perception of spike streams to handle the spike quantization effects. Besides, we utilize the estimated motion to suppress the potential blur from inter-stream clips, considering the underlying motion of spike streams. We also develop a spike interval-guided alignment module to tackle the blur from intra-stream clips. Experimental results on both synthetic and real-captured data demonstrate that our method can restore high-quality images from spike streams.

Developing a unified algorithm that can learn from and generate across modalities such as text, images and video has been a fundamental challenge in artificial intelligence. Although next-token prediction has driven major advances in large language models1, its extension to multimodal domains has remained limited, and diffusion models for image and video synthesis2,3 and compositional frameworks that integrate vision encoders with language models4 still dominate. Here we introduce Emu3, a family of multimodal models trained solely with next-token prediction. Emu3 equals the performance of well-established task-specific models across both perception and generation, matching flagship systems while removing the need for diffusion or compositional architectures. It further demonstrates coherent, high-fidelity video generation, interleaved vision-language generation and vision-language-action modelling for robotic manipulation. By reducing multimodal learning to unified token prediction, Emu3 establishes a robust foundation for large-scale multimodal modelling and offers a promising route towards unified multimodal intelligence.

3D Gaussian Splatting (3DGS) has emerged as a powerful representation for 3D scenes, widely adopted due to its exceptional efficiency and high-fidelity visual quality. Given the significant value of 3DGS assets, recent works have introduced specialized watermarking schemes to ensure copyright protection and ownership verification. However, can existing 3D Gaussian watermarking approaches genuinely guarantee robust protection of the 3D assets? In this paper, for the first time, we systematically explore and validate possible vulnerabilities of 3DGS watermarking frameworks. We demonstrate that conventional watermark removal techniques designed for 2D images do not effectively generalize to the 3DGS scenario due to the specialized rendering pipeline and unique attributes of each gaussian primitives. Motivated by this insight, we propose GSPure, the first watermark purification framework specifically for 3DGS watermarking representations. By analyzing view-dependent rendering contributions and exploiting geometrically accurate feature clustering, GSPure precisely isolates and effectively removes watermark-related Gaussian primitives while preserving scene integrity. Extensive experiments demonstrate that our GSPure achieves the best watermark purification performance, reducing watermark PSNR by up to 16.34dB while minimizing degradation to original scene fidelity with less than 1dB PSNR loss. Moreover, it consistently outperforms existing methods in both effectiveness and generalization.

Biophysically detailed neural networks represent a promising frontier for brain-inspired AI, offering intrinsic spatio-temporal dynamics to enhance the expressivity and computational density of deep learning systems. However, general-purpose deep learning frameworks suffer from a fundamental mismatch between their dense parallel optimizations and the irregular, tree-structured complexity of biological mechanisms. In this work, we propose , a framework designed to unify high-performance simulation with scalable training. Unlike approaches that adapt biology to existing deep learning tools, HelioX adopts a "GPU-to-Biophysics" paradigm. We tailor the underlying GPU parallelism to biological structures by implementing custom-fused CUDA kernels for both the Dendritic Hierarchical Scheduling (DHS) algorithm and its gradient propagation. This design eliminates the runtime overhead of generic automatic differentiation and enables multi-stream concurrency for spike generation and equation assembly. Experimental results demonstrate that HelioX outperforms standard simulators (NEURON) by orders of magnitude and surpasses prior GPU-based solvers in both speed and scalability. We successfully train deep biophysical MLPs and organism-scale biophysical neural networks (e.g., the BAAIWorm C. elegans model) on a single consumer-grade GPU. HelioX establishes a new standard for computational efficiency, enabling the training of biophysically detailed models at scales previously unattainable.

We introduce Orca, an initial instantiation of a general world foundation model. Orca learns a unified world latent space from multimodal world signals and exposes it through multimodal readout interfaces. Rather than optimizing isolated next-token, next-frame, or next-action prediction, we are centered on Next-State-Prediction modeling, offering a unified state-transition modeling route toward understanding, predicting, and acting upon the world. Orca learns through two complementary paradigms: unconscious learning captures dense natural state transitions from continuous videos, and conscious learning models sparse meaningful state transitions by language-described events and VQA supervision. For pre-training, we construct a large-scale world-learning inventory data, including 125K hours of video data and 160M event annotations. After pre-training, Orca learns a unified world latent space. To examine whether the learned latent supports downstream, we evaluate it by three representative downstream readouts: text generation, image prediction, and embodied action generation. Orca's backbone is frozen, and only the lightweight modality-specific decoders are trainable. Experiments show the scalability of the proposed paradigm and verify that stronger world latent enables stronger downstream readouts. Orca outperforms similar-sized specialized baselines. These results show that Orca, as a general world foundation model, presents a promising approach to understanding, predicting, and acting upon the world. Finally, we discuss the current limitations, aiming to provide useful insights and inspiration for the community.