This paper describes a system developed for the GENEA (Generation and Evaluation of Non-verbal Behaviour for Embodied Agents) Challenge 2023. Our solution builds on an existing difusion-based motion synthesis model. We propose a contrastive speech and motion pretraining (CSMP) module, which learns a joint embedding for speech and gesture with the aim to learn a semantic coupling between these modalities. The output of the CSMP module is used as a conditioning signal in the difusion-based gesture synthesis model in order to achieve semantically-aware co-speech gesture generation. Our entry achieved highest human-likeness and highest speech appropriateness rating among the submitted entries. This indicates that our system is a promising approach to achieve human-like co-speech gestures in agents that carry semantic meaning.

One of the main goals of robotics and intelligent agent research is to enable them to communicate with humans in physically situated settings. Human communication consists of both verbal and non-verbal modes. Recent studies in enabling communication for intelligent agents have focused on verbal modes, i.e., language and speech. However, in a situated setting the non-verbal mode is crucial for an agent to adapt flexible communication strategies. In this work, we focus on learning to generate non-verbal communicative expressions in situated embodied interactive agents. Specifically, we show that an agent can learn pointing gestures in a physically simulated environment through a combination of imitation and reinforcement learning that achieves high motion naturalness and high referential accuracy. We compared our proposed system against several baselines in both subjective and objective evaluations. The subjective evaluation is done in a virtual reality setting where an embodied referential game is played between the user and the agent in a shared 3D space, a setup that fully assesses the communicative capabilities of the generated gestures. The evaluations show that our model achieves a higher level of referential accuracy and motion naturalness compared to a state-of-the-art supervised learning motion synthesis model, showing the promise of our proposed system that combines imitation and reinforcement learning for generating communicative gestures. Additionally, our system is robust in a physically-simulated environment thus has the potential of being applied to robots.

This paper focuses on enhancing human-agent communication by integrating spatial context into virtual agents’ non-verbal behaviors, specifically gestures. Recent advances in co-speech gesture generation have primarily utilized data-driven methods, which create natural motion but limit the scope of gestures to those performed in a void. Our work aims to extend these methods by enabling generative models to incorporate scene information into speech-driven gesture synthesis. We introduce a novel synthetic gesture dataset tailored for this purpose. This development represents a critical step toward creating embodied conversational agents that interact more naturally with their environment and users.

Human motion generation has advanced rapidly in recent years, yet the critical problem of creating spatially grounded, context-aware gestures has been largely overlooked. Existing models typically specialize either in descriptive motion generation, such as locomotion and object interaction, or in isolated co-speech gesture synthesis aligned with utterance semantics. However, both lines of work often treat motion and environmental grounding separately, limiting advances toward embodied, communicative agents. To address this gap, our work introduces a multimodal dataset and framework for grounded gesture generation, combining two key resources: (1) a synthetic dataset of spatially grounded referential gestures, and (2) MM-Conv, a VR-based dataset capturing two-party dialogues. Together, they provide over 7.7 hours of synchronized motion, speech, and 3D scene information, standardized in the HumanML3D format. Our framework further connects to a physics-based simulator, enabling synthetic data generation and situated evaluation. By bridging gesture modeling and spatial grounding, our contribution establishes a foundation for advancing research in situated gesture generation and grounded multimodal interaction.

Referential gestures, pointing, indicating, and orienting the body toward objects in shared space, are fundamental to embodied com- munication. For virtual agents and physical robots operating in human environments, the ability to generate spatially grounded gestures is essential for disambiguation, instruction, and collabora- tive interaction. Yet, research on communicative gesture generation has largely focused on co-speech beat and iconic gestures, trained on corpora in which spatial grounding is absent or incidental. This lack of active research on referential gestures can be attributed to three key factors: datasets that pair gestures with 3D scene context are scarce, referential gesture generation lacks task formu- lation, and metrics for evaluating spatial grounding do not exist. In this work, we address all three gaps by introducing the MM- Conv Referential Gesture Generation Challenge. Specifically, the benchmark consists of three components: (i) a paired data release of 3,000 pointing-annotated clips from MM-Conv and SGS-HSI, with pointing-quality annotations and scene-disjoint splits; (ii) a task formulation that requires systems to produce spatially grounded reference gestures aligned with speech, without oracle apex timing or motion templates; (iii) a spatio-temporal evaluation protocol decomposing referential gesture quality into temporal alignment, spatial accuracy, and referent recall. We present a modular baseline based on OmniControl and position the benchmark as the founda- tion for the scene-aware gesture generation challenge at the 1st Workshop on Human–Scene Interaction at ECCV 2026. We envision this challenge as a testbed for the next generation of referential gesture synthesis works. 

Grounding language in the physical world requires AI systems to interpret references that emerge dynamically during conversation. While current vision-language models (VLMs) excel at static image tasks, they struggle to resolve ambiguous expressions in spontaneous, multi-turn dialogue. We address this gap by introducing MM-Conv—speak, point, look—a benchmark for referential communication in dynamic 3D environments, built from 6.7 hours of egocentric VR interaction with synchronized speech, motion, gaze, and 3D scene geometry. The benchmark includes over 4,200 manually verified referring expressions spanning full, partitive, and pronominal types, enabling systematic evaluation of multimodal reference resolution.

We introduce Look and Tell, a multimodal dataset for studying referential communication across egocentric and exocentric perspectives. Using Meta Project Aria smart glasses and stationary cameras, we recorded synchronized gaze, speech, and video as 25 participants instructed a partner to identify ingredients in a kitchen. Combined with 3D scene reconstructions, this setup provides a benchmark for evaluating how different spatial representations (2D vs. 3D; ego vs. exo) affect multimodal grounding. The dataset contains 3.67 hours of recordings, including 2,707 richly annotated referential expressions, and is designed to advance the development of embodied agents that can understand and engage in situated dialogue.