Hunyuan-TurboS: Advancing Large Language Models through Mamba-Transformer Synergy and Adaptive Chain-of-Thought

As Large Language Models (LLMs) rapidly advance, we introduce Hunyuan-TurboS, a novel large hybrid Transformer-Mamba Mixture of Experts (MoE) model. It synergistically combines Mamba's long-sequence processing efficiency with Transformer's superior contextual understanding. Hunyuan-TurboS features an adaptive long-short chain-of-thought (CoT) mechanism, dynamically switching between rapid responses for simple queries and deep "thinking" modes for complex problems, optimizing computational resources. Architecturally, this 56B activated (560B total) parameter model employs 128 layers (Mamba2, Attention, FFN) with an innovative AMF/MF block pattern. Faster Mamba2 ensures linear complexity, Grouped-Query Attention minimizes KV cache, and FFNs use an MoE structure. Pre-trained on 16T high-quality tokens, it supports a 256K context length and is the first industry-deployed large-scale Mamba model. Our comprehensive post-training strategy enhances capabilities via Supervised Fine-Tuning (3M instructions), a novel Adaptive Long-short CoT Fusion method, Multi-round Deliberation Learning for iterative improvement, and a two-stage Large-scale Reinforcement Learning process targeting STEM and general instruction-following. Evaluations show strong performance: overall top 7 rank on LMSYS Chatbot Arena with a score of 1356, outperforming leading models like Gemini-2.0-Flash-001 (1352) and o4-mini-2025-04-16 (1345). TurboS also achieves an average of 77.9% across 23 automated benchmarks. Hunyuan-TurboS balances high performance and efficiency, offering substantial capabilities at lower inference costs than many reasoning models, establishing a new paradigm for efficient large-scale pre-trained models.

mmMirror: Device Free mmWave Indoor NLoS Localization Using Van-Atta-Array IRS

Industry 4.0 is transforming manufacturing and logistics by integrating robots into shared human environments, such as factories, warehouses, and healthcare facilities. However, the risk of human-robot collisions, especially in Non-Line-of-Sight (NLoS) scenarios like around corners, remains a critical challenge. Existing solutions, such as vision-based and LiDAR systems, often fail under occlusion, lighting constraints, or privacy concerns, while RF-based systems are limited by range and accuracy. To address these limitations, we propose mmMirror, a novel system leveraging a Van Atta Array-based millimeter-wave (mmWave) reconfigurable intelligent reflecting surface (IRS) for precise, device-free NLoS localization. mmMirror integrates seamlessly with existing frequency-modulated continuous-wave (FMCW) radars and offers: (i) robust NLoS localization with centimeter-level accuracy at ranges up to 3 m, (ii) seamless uplink and downlink communication between radar and IRS, (iii) support for multi-radar and multi-target scenarios via dynamic beam steering, and (iv) reduced scanning latency through adaptive time slot allocation. Implemented using commodity 24 GHz radars and a PCB-based IRS prototype, mmMirror demonstrates its potential in enabling safe human-robot interactions in dynamic and complex environments.

A Learnable Multi-views Contrastive Framework with Reconstruction Discrepancy for Medical Time-Series

In medical time series disease diagnosis, two key challenges are identified.First, the high annotation cost of medical data leads to overfitting in models trained on label-limited, single-center datasets. To address this, we propose incorporating external data from related tasks and leveraging AE-GAN to extract prior knowledge,providing valuable references for downstream tasks. Second, many existing studies employ contrastive learning to derive more generalized medical sequence representations for diagnostic tasks, usually relying on manually designed diverse positive and negative sample pairs.However, these approaches are complex, lack generalizability, and fail to adaptively capture disease-specific features across different conditions.To overcome this, we introduce LMCF (Learnable Multi-views Contrastive Framework), a framework that integrates a multi-head attention mechanism and adaptively learns representations from different views through inter-view and intra-view contrastive learning strategies.Additionally, the pre-trained AE-GAN is used to reconstruct discrepancies in the target data as disease probabilities, which are then integrated into the contrastive learning process.Experiments on three target datasets demonstrate that our method consistently outperforms seven other baselines, highlighting its significant impact on healthcare applications such as the diagnosis of myocardial infarction, Alzheimer's disease, and Parkinson's disease.

Application of Structured State Space Models to High energy physics with locality-sensitive hashing

Modern high-energy physics (HEP) experiments are increasingly challenged by the vast size and complexity of their datasets, particularly regarding large-scale point cloud processing and long sequences. In this study, to address these challenges, we explore the application of structured state space models (SSMs), proposing one of the first trials to integrate local-sensitive hashing into either a hybrid or pure Mamba Model. Our results demonstrate that pure SSMs could serve as powerful backbones for HEP problems involving tasks for long sequence data with local inductive bias. By integrating locality-sensitive hashing into Mamba blocks, we achieve significant improvements over traditional backbones in key HEP tasks, surpassing them in inference speed and physics metrics while reducing computational overhead. In key tests, our approach demonstrated promising results, presenting a viable alternative to traditional transformer backbones by significantly reducing FLOPS while maintaining robust performance.

Robust Load Prediction of Power Network Clusters Based on Cloud-Model-Improved Transformer

Load data from power network clusters indicates economic development in each area, crucial for predicting regional trends and guiding power enterprise decisions. The Transformer model, a leading method for load prediction, faces challenges modeling historical data due to variables like weather, events, festivals, and data volatility. To tackle this, the cloud model's fuzzy feature is utilized to manage uncertainties effectively. Presenting an innovative approach, the Cloud Model Improved Transformer (CMIT) method integrates the Transformer model with the cloud model utilizing the particle swarm optimization algorithm, with the aim of achieving robust and precise power load predictions. Through comparative experiments conducted on 31 real datasets within a power network cluster, it is demonstrated that CMIT significantly surpasses the Transformer model in terms of prediction accuracy, thereby highlighting its effectiveness in enhancing forecasting capabilities within the power network cluster sector.

BUFF: Boosted Decision Tree based Ultra-Fast Flow matching

Tabular data stands out as one of the most frequently encountered types in high energy physics. Unlike commonly homogeneous data such as pixelated images, simulating high-dimensional tabular data and accurately capturing their correlations are often quite challenging, even with the most advanced architectures. Based on the findings that tree-based models surpass the performance of deep learning models for tasks specific to tabular data, we adopt the very recent generative modeling class named conditional flow matching and employ different techniques to integrate the usage of Gradient Boosted Trees. The performances are evaluated for various tasks on different analysis level with several public datasets. We demonstrate the training and inference time of most high-level simulation tasks can achieve speedup by orders of magnitude. The application can be extended to low-level feature simulation and conditioned generations with competitive performance.

Super-resolution of biomedical volumes with 2D supervision

Volumetric biomedical microscopy has the potential to increase the diagnostic information extracted from clinical tissue specimens and improve the diagnostic accuracy of both human pathologists and computational pathology models. Unfortunately, barriers to integrating 3-dimensional (3D) volumetric microscopy into clinical medicine include long imaging times, poor depth / z-axis resolution, and an insufficient amount of high-quality volumetric data. Leveraging the abundance of high-resolution 2D microscopy data, we introduce masked slice diffusion for super-resolution (MSDSR), which exploits the inherent equivalence in the data-generating distribution across all spatial dimensions of biological specimens. This intrinsic characteristic allows for super-resolution models trained on high-resolution images from one plane (e.g., XY) to effectively generalize to others (XZ, YZ), overcoming the traditional dependency on orientation. We focus on the application of MSDSR to stimulated Raman histology (SRH), an optical imaging modality for biological specimen analysis and intraoperative diagnosis, characterized by its rapid acquisition of high-resolution 2D images but slow and costly optical z-sectioning. To evaluate MSDSR's efficacy, we introduce a new performance metric, SliceFID, and demonstrate MSDSR's superior performance over baseline models through extensive evaluations. Our findings reveal that MSDSR not only significantly enhances the quality and resolution of 3D volumetric data, but also addresses major obstacles hindering the broader application of 3D volumetric microscopy in clinical diagnostics and biomedical research.

Step-Calibrated Diffusion for Biomedical Optical Image Restoration

High-quality, high-resolution medical imaging is essential for clinical care. Raman-based biomedical optical imaging uses non-ionizing infrared radiation to evaluate human tissues in real time and is used for early cancer detection, brain tumor diagnosis, and intraoperative tissue analysis. Unfortunately, optical imaging is vulnerable to image degradation due to laser scattering and absorption, which can result in diagnostic errors and misguided treatment. Restoration of optical images is a challenging computer vision task because the sources of image degradation are multi-factorial, stochastic, and tissue-dependent, preventing a straightforward method to obtain paired low-quality/high-quality data. Here, we present Restorative Step-Calibrated Diffusion (RSCD), an unpaired image restoration method that views the image restoration problem as completing the finishing steps of a diffusion-based image generation task. RSCD uses a step calibrator model to dynamically determine the severity of image degradation and the number of steps required to complete the reverse diffusion process for image restoration. RSCD outperforms other widely used unpaired image restoration methods on both image quality and perceptual evaluation metrics for restoring optical images. Medical imaging experts consistently prefer images restored using RSCD in blinded comparison experiments and report minimal to no hallucinations. Finally, we show that RSCD improves performance on downstream clinical imaging tasks, including automated brain tumor diagnosis and deep tissue imaging. Our code is available at https://github.com/MLNeurosurg/restorative_step-calibrated_diffusion.

Cross-Resolution Land Cover Classification Using Outdated Products and Transformers

Large-scale high-resolution land cover classification is a prerequisite for constructing Earth system models and addressing ecological and resource issues. Advancements in satellite sensor technology have led to an improvement in spatial resolution and wider coverage areas. Nevertheless, the lack of high-resolution labeled data is still a challenge, hindering the largescale application of land cover classification methods. In this paper, we propose a Transformerbased weakly supervised method for cross-resolution land cover classification using outdated data. First, to capture long-range dependencies without missing the fine-grained details of objects, we propose a U-Net-like Transformer based on a reverse difference mechanism (RDM) using dynamic sparse attention. Second, we propose an anti-noise loss calculation (ANLC) module based on optimal transport (OT). Anti-noise loss calculation identifies confident areas (CA) and vague areas (VA) based on the OT matrix, which relieves the impact of noises in outdated land cover products. By introducing a weakly supervised loss with weights and employing unsupervised loss, the RDM-based U-Net-like Transformer was trained. Remote sensing images with 1 m resolution and the corresponding ground-truths of six states in the United States were employed to validate the performance of the proposed method. The experiments utilized outdated land cover products with 30 m resolution from 2013 as training labels, and produced land cover maps with 1 m resolution from 2017. The results show the superiority of the proposed method compared to state-of-the-art methods. The code is available at https://github.com/yu-ni1989/ANLC-Former.

A self-supervised framework for learning whole slide representations

Whole slide imaging is fundamental to biomedical microscopy and computational pathology. However, whole slide images (WSIs) present a complex computer vision challenge due to their gigapixel size, diverse histopathologic features, spatial heterogeneity, and limited/absent data annotations. These challenges highlight that supervised training alone can result in suboptimal whole slide representations. Self-supervised representation learning can achieve high-quality WSI visual feature learning for downstream diagnostic tasks, such as cancer diagnosis or molecular genetic prediction. Here, we present a general self-supervised whole slide learning (S3L) framework for gigapixel-scale self-supervision of WSIs. S3L combines data transformation strategies from transformer-based vision and language modeling into a single unified framework to generate paired views for self-supervision. S3L leverages the inherent regional heterogeneity, histologic feature variability, and information redundancy within WSIs to learn high-quality whole-slide representations. We benchmark S3L visual representations on two diagnostic tasks for two biomedical microscopy modalities. S3L significantly outperforms WSI baselines for cancer diagnosis and genetic mutation prediction. Additionally, S3L achieves good performance using both in-domain and out-of-distribution patch encoders, demonstrating good flexibility and generalizability.