Learning to Assist: Physics-Grounded Human-Human Control via Multi-Agent Reinforcement Learning

Humanoid robotics has strong potential to transform daily service and caregiving applications. Although recent advances in general motion tracking within physics engines (GMT) have enabled virtual characters and humanoid robots to reproduce a broad range of human motions, these behaviors are primarily limited to contact-less social interactions or isolated movements. Assistive scenarios, by contrast, require continuous awareness of a human partner and rapid adaptation to their evolving posture and dynamics. In this paper, we formulate the imitation of closely interacting, force-exchanging human-human motion sequences as a multi-agent reinforcement learning problem. We jointly train partner-aware policies for both the supporter (assistant) agent and the recipient agent in a physics simulator to track assistive motion references. To make this problem tractable, we introduce a partner policies initialization scheme that transfers priors from single-human motion-tracking controllers, greatly improving exploration. We further propose dynamic reference retargeting and contact-promoting reward, which adapt the assistant's reference motion to the recipient's real-time pose and encourage physically meaningful support. We show that AssistMimic is the first method capable of successfully tracking assistive interaction motions on established benchmarks, demonstrating the benefits of a multi-agent RL formulation for physically grounded and socially aware humanoid control.

Formula-Supervised Sound Event Detection: Pre-Training Without Real Data

In this paper, we propose a novel formula-driven supervised learning (FDSL) framework for pre-training an environmental sound analysis model by leveraging acoustic signals parametrically synthesized through formula-driven methods. Specifically, we outline detailed procedures and evaluate their effectiveness for sound event detection (SED). The SED task, which involves estimating the types and timings of sound events, is particularly challenged by the difficulty of acquiring a sufficient quantity of accurately labeled training data. Moreover, it is well known that manually annotated labels often contain noises and are significantly influenced by the subjective judgment of annotators. To address these challenges, we propose a novel pre-training method that utilizes a synthetic dataset, Formula-SED, where acoustic data are generated solely based on mathematical formulas. The proposed method enables large-scale pre-training by using the synthesis parameters applied at each time step as ground truth labels, thereby eliminating label noise and bias. We demonstrate that large-scale pre-training with Formula-SED significantly enhances model accuracy and accelerates training, as evidenced by our results in the DESED dataset used for DCASE2023 Challenge Task 4. The project page is at https://yutoshibata07.github.io/Formula-SED/

Acoustic-based 3D Human Pose Estimation Robust to Human Position

This paper explores the problem of 3D human pose estimation from only low-level acoustic signals. The existing active acoustic sensing-based approach for 3D human pose estimation implicitly assumes that the target user is positioned along a line between loudspeakers and a microphone. Because reflection and diffraction of sound by the human body cause subtle acoustic signal changes compared to sound obstruction, the existing model degrades its accuracy significantly when subjects deviate from this line, limiting its practicality in real-world scenarios. To overcome this limitation, we propose a novel method composed of a position discriminator and reverberation-resistant model. The former predicts the standing positions of subjects and applies adversarial learning to extract subject position-invariant features. The latter utilizes acoustic signals before the estimation target time as references to enhance robustness against the variations in sound arrival times due to diffraction and reflection. We construct an acoustic pose estimation dataset that covers diverse human locations and demonstrate through experiments that our proposed method outperforms existing approaches.

Anatomical 3D Style Transfer Enabling Efficient Federated Learning with Extremely Low Communication Costs

In this study, we propose a novel federated learning (FL) approach that utilizes 3D style transfer for the multi-organ segmentation task. The multi-organ dataset, obtained by integrating multiple datasets, has high scalability and can improve generalization performance as the data volume increases. However, the heterogeneity of data owing to different clients with diverse imaging conditions and target organs can lead to severe overfitting of local models. To align models that overfit to different local datasets, existing methods require frequent communication with the central server, resulting in higher communication costs and risk of privacy leakage. To achieve an efficient and safe FL, we propose an Anatomical 3D Frequency Domain Generalization (A3DFDG) method for FL. A3DFDG utilizes structural information of human organs and clusters the 3D styles based on the location of organs. By mixing styles based on these clusters, it preserves the anatomical information and leads models to learn intra-organ diversity, while aligning the optimization of each local model. Experiments indicate that our method can maintain its accuracy even in cases where the communication cost is highly limited (=1.25% of the original cost) while achieving a significant difference compared to baselines, with a higher global dice similarity coefficient score of 4.3%. Despite its simplicity and minimal computational overhead, these results demonstrate that our method has high practicality in real-world scenarios where low communication costs and a simple pipeline are required. The code used in this project will be publicly available.