Shuai Yao

Haifeng Wang, Hua Wu, Tian Wu et al.

In this report, we introduce ERNIE 5.0, a natively autoregressive foundation model desinged for unified multimodal understanding and generation across text, image, video, and audio. All modalities are trained from scratch under a unified next-group-of-tokens prediction objective, based on an ultra-sparse mixture-of-experts (MoE) architecture with modality-agnostic expert routing. To address practical challenges in large-scale deployment under diverse resource constraints, ERNIE 5.0 adopts a novel elastic training paradigm. Within a single pre-training run, the model learns a family of sub-models with varying depths, expert capacities, and routing sparsity, enabling flexible trade-offs among performance, model size, and inference latency in memory- or time-constrained scenarios. Moreover, we systematically address the challenges of scaling reinforcement learning to unified foundation models, thereby guaranteeing efficient and stable post-training under ultra-sparse MoE architectures and diverse multimodal settings. Extensive experiments demonstrate that ERNIE 5.0 achieves strong and balanced performance across multiple modalities. To the best of our knowledge, among publicly disclosed models, ERNIE 5.0 represents the first production-scale realization of a trillion-parameter unified autoregressive model that supports both multimodal understanding and generation. To facilitate further research, we present detailed visualizations of modality-agnostic expert routing in the unified model, alongside comprehensive empirical analysis of elastic training, aiming to offer profound insights to the community.

Yong Wang, Qifan Shen, Bao Zhang et al.

Millimeter-wave (mmWave) radar enables contactless respiratory sensing,yet fine-grained monitoring is often degraded by nonstationary interference from body micromotions.To achieve micromotion interference removal,we propose mmWave-Diffusion,an observation-anchored conditional diffusion framework that directly models the residual between radar phase observations and the respiratory ground truth,and initializes sampling within an observation-consistent neighborhood rather than from Gaussian noise-thereby aligning the generative process with the measurement physics and reducing inference overhead. The accompanying Radar Diffusion Transformer (RDT) is explicitly conditioned on phase observations, enforces strict one-to-one temporal alignment via patch-level dual positional encodings, and injects local physical priors through banded-mask multi-head cross-attention, enabling robust denoising and interference removal in just 20 reverse steps. Evaluated on 13.25 hours of synchronized radar-respiration data, mmWave-Diffusion achieves state-of-the-art waveform reconstruction and respiratory-rate estimation with strong generalization. Code repository:https://github.com/goodluckyongw/mmWave-Diffusion.

Yijun Wang, Yong Wang, Chendong xu et al.

Human Activity Recognition (HAR) such as fall detection has become increasingly critical due to the aging population, necessitating effective monitoring systems to prevent serious injuries and fatalities associated with falls. This study focuses on fine-tuning the Vision Transformer (ViT) model specifically for HAR using radar-based Time-Doppler signatures. Unlike traditional image datasets, these signals present unique challenges due to their non-visual nature and the high degree of similarity among various activities. Directly fine-tuning the ViT with all parameters proves suboptimal for this application. To address this challenge, we propose a novel approach that employs Low-Rank Adaptation (LoRA) fine-tuning in the weight space to facilitate knowledge transfer from pre-trained ViT models. Additionally, to extract fine-grained features, we enhance feature representation through the integration of a serial-parallel adapter in the feature space. Our innovative joint fine-tuning method, tailored for radar-based Time-Doppler signatures, significantly improves HAR accuracy, surpassing existing state-of-the-art methodologies in this domain. Our code is released at https://github.com/wangyijunlyy/SelaFD.

Yang Ye, Jingyuan Tan, Tianyue Jiang et al.

Training effective software engineering agents requires large volumes of task-specific trajectories, incurring substantial data construction costs. Inspired by the "Less-Is-More" hypothesis in mathematical reasoning, we investigate its extension to agentic scenarios and propose an end-to-end training framework that achieves superior agentic capabilities with fewer but higher-quality training trajectories. This is achieved via STITCH (Sliding-memory Trajectory Inference and Task Chunking Heuristic), a coarse-to-fine mechanism that filters low-value noise and retains decision-critical tokens to maximize training signal quality. We conduct experiments across multiple agent frameworks (e.g., mini-SWE-agent, MSWE-agent), model scales (30B to 355B), and multilingual settings (Python, Java, and ArkTS). On SWE-bench Verified, models trained with STITCH achieve up to 63.16% relative improvement over base models. On Multi-SWE-bench (Java), MiniMax-M2.5-STITCH achieves 43.75% with our CodeArts Agent scaffold (+16.67%). On HarmonyOS (ArkTS), GLM-4.7-STITCH improves the compilation pass rate to 61.31% (+43.34%) with less than 1K training trajectories. Our results confirm that the "Less-Is-More" paradigm generalizes effectively to complex agentic tasks across diverse languages and model scales.