Qijun Huang

Maiyue Chen, Kaihui Wang, Bo Zhang et al.

In this report, we present HoloMotion-1, a humanoid motion foundation model for zero-shot whole-body motion tracking. A key innovation of HoloMotion-1 is to scale control-policy training with a large-scale hybrid motion corpus, where video-reconstructed motions from in-the-wild videos provide the dominant source of motion diversity, while curated motion-capture and in-house motion data provide higher-fidelity supervision and deployment-oriented coverage. This data regime enables HoloMotion-1 to move beyond conventional MoCap-only training and exposes the policy to substantially broader behaviors, capture conditions, and motion styles. Learning from such heterogeneous data introduces new challenges, including reconstruction noise, source-domain mismatch, uneven motion quality, and the need for temporal modeling under large behavioral variation. To address these challenges, HoloMotion-1 integrates large-capacity temporal modeling, a sparsely activated Mixture-of-Experts Transformer with KV-cache inference for real-time control, and a sequence-level training strategy that improves learning efficiency on extended motion sequences. Extensive experiments on multiple unseen motion benchmarks show that HoloMotion-1 generalizes robustly across diverse motion types and capture conditions, significantly improves tracking accuracy over prior methods, and transfers directly to a real humanoid robot without task-specific fine-tuning.

Huaicheng Zhang, Wei Tan, Guangzheng Li et al.

Recent advances in audio-based generative language models have accelerated AI-driven lyric-to-song generation. However, these models frequently suffer from content hallucination, producing outputs misaligned with the input lyrics and undermining musical coherence. Current supervised fine-tuning (SFT) approaches, limited by passive label-fitting, exhibit constrained self-improvement and poor hallucination mitigation. To address this core challenge, we propose a novel reinforcement learning (RL) framework leveraging preference optimization for hallucination control. Our key contributions include: (1) Developing a robust hallucination preference dataset constructed via phoneme error rate (PER) computation and rule-based filtering to capture alignment with human expectations; (2) Implementing and evaluating three distinct preference optimization strategies within the RL framework: Direct Preference Optimization (DPO), Proximal Policy Optimization (PPO), and Group Relative Policy Optimization (GRPO). DPO operates off-policy to enhance positive token likelihood, achieving a significant 7.4% PER reduction. PPO and GRPO employ an on-policy approach, training a PER-based reward model to iteratively optimize sequences via reward maximization and KL-regularization, yielding PER reductions of 4.9% and 4.7%, respectively. Comprehensive objective and subjective evaluations confirm that our methods effectively suppress hallucinations while preserving musical quality. Crucially, this work presents a systematic, RL-based solution to hallucination control in lyric-to-song generation. The framework's transferability also unlocks potential for music style adherence and musicality enhancement, opening new avenues for future generative song research.

Jundong Xu, Hao Fei, Yuhui Zhang et al.

Multimodal symbolic logical reasoning, which aims to deduce new facts from multimodal input via formal logic, is critical in high-stakes applications such as autonomous driving and medical diagnosis, as its rigorous, deterministic reasoning helps prevent serious consequences. To evaluate such capabilities of current state-of-the-art vision language models (VLMs), we introduce the first benchmark MuSLR for multimodal symbolic logical reasoning grounded in formal logical rules. MuSLR comprises 1,093 instances across 7 domains, including 35 atomic symbolic logic and 976 logical combinations, with reasoning depths ranging from 2 to 9. We evaluate 7 state-of-the-art VLMs on MuSLR and find that they all struggle with multimodal symbolic reasoning, with the best model, GPT-4.1, achieving only 46.8%. Thus, we propose LogiCAM, a modular framework that applies formal logical rules to multimodal inputs, boosting GPT-4.1's Chain-of-Thought performance by 14.13%, and delivering even larger gains on complex logics such as first-order logic. We also conduct a comprehensive error analysis, showing that around 70% of failures stem from logical misalignment between modalities, offering key insights to guide future improvements. All data and code are publicly available at https://llm-symbol.github.io/MuSLR.

Huaicheng Zhang, Ruoxin Wang, Chenlian Zhou et al.

As one of the most effective methods for cardiovascular disease (CVD) diagnosis, multi-lead Electrocardiogram (ECG) signals present a characteristic multi-sensor information fusion challenge that has been continuously researched in deep learning domains. Despite the numerous algorithms proposed with different DL architectures, maintaining a balance among performance, computational complexity, and multi-source ECG feature fusion remains challenging. Recently, state space models (SSMs), particularly Mamba, have demonstrated remarkable effectiveness across various fields. Their inherent design for high-efficiency computation and linear complexity makes them particularly suitable for low-dimensional data like ECGs. This work proposes S2M2ECG, an SSM architecture featuring three-level fusion mechanisms: (1) Spatio-temporal bi-directional SSMs with segment tokenization for low-level signal fusion, (2) Intra-lead temporal information fusion with bi-directional scanning to enhance recognition accuracy in both forward and backward directions, (3) Cross-lead feature interaction modules for spatial information fusion. To fully leverage the ECG-specific multi-lead mechanisms inherent in ECG signals, a multi-branch design and lead fusion modules are incorporated, enabling individual analysis of each lead while ensuring seamless integration with others. Experimental results reveal that S2M2ECG achieves superior performance in the rhythmic, morphological, and clinical scenarios. Moreover, its lightweight architecture ensures it has nearly the fewest parameters among existing models, making it highly suitable for efficient inference and convenient deployment. Collectively, S2M2ECG offers a promising alternative that strikes an excellent balance among performance, computational complexity, and ECG-specific characteristics, paving the way for high-performance, lightweight computations in CVD diagnosis.

Ruihan Hu, Qijun Huang, Sheng Chang et al.

Machine learning algorithms have been effectively applied into various real world tasks. However, it is difficult to provide high-quality machine learning solutions to accommodate an unknown distribution of input datasets; this difficulty is called the uncertainty prediction problems. In this paper, a margin-based Pareto deep ensemble pruning (MBPEP) model is proposed. It achieves the high-quality uncertainty estimation with a small value of the prediction interval width (MPIW) and a high confidence of prediction interval coverage probability (PICP) by using deep ensemble networks. In addition to these networks, unique loss functions are proposed, and these functions make the sub-learners available for standard gradient descent learning. Furthermore, the margin criterion fine-tuning-based Pareto pruning method is introduced to optimize the ensembles. Several experiments including predicting uncertainties of classification and regression are conducted to analyze the performance of MBPEP. The experimental results show that MBPEP achieves a small interval width and a low learning error with an optimal number of ensembles. For the real-world problems, MBPEP performs well on input datasets with unknown distributions datasets incomings and improves learning performance on a multi task problem when compared to that of each single model.

Zhiri Tang, Ruohua Zhu, Peng Lin et al.

Memristive neural networks (MNNs), which use memristors as neurons or synapses, have become a hot research topic recently. However, most memristors are not compatible with mainstream integrated circuit technology and their stabilities in large-scale are not very well so far. In this paper, a hardware friendly MNN circuit is introduced, in which the memristive characteristics are implemented by digital integrated circuit. Through this method, spike timing dependent plasticity (STDP) and unsupervised learning are realized. A weight sharing mechanism is proposed to bridge the gap of network scale and hardware resource. Experiment results show the hardware resource is significantly saved with it, maintaining good recognition accuracy and high speed. Moreover, the tendency of resource increase is slower than the expansion of network scale, which infers our method's potential on large scale neuromorphic network's realization.