Ex-Omni: New Model Enables 3D Facial Animation Generation for Omni-modal Language Models
Researchers have developed Ex-Omni, an open-source model that enables large language models to generate synchronized speech and 3D facial animations for more natural human-computer interaction. The model addresses the technical challenge of bridging discrete semantic reasoning in language models with the continuous temporal dynamics required for realistic facial motion. This advancement could improve conversational AI systems by making them more expressive and lifelike in their interactions.
Ex-Omni is a new framework that extends omni-modal large language models (OLLMs) to jointly produce speech and 3D facial animation, addressing a gap in multimodal AI capabilities. The core innovation involves decoupling semantic reasoning from temporal generation through a blendshape-aware speech unit generator and decoder, where speech units provide temporal scaffolding while hidden speech representations carry facial animation cues. The researchers introduced a unified token-as-query gated fusion mechanism for controlled semantic injection and created InstructS2SF-1200K, a dataset of 1.2 million samples for model pre-training. Experimental results demonstrate that Ex-Omni maintains competitive speech understanding and generation abilities while achieving better audio-visual synchronization and lower face-generation latency compared to cascaded pipeline approaches. The open-source release of the model makes it available for further research and development.
What's missing
The paper does not discuss potential limitations regarding diversity of facial expressions across different ethnicities, age groups, or emotional states in the training dataset. Additionally, computational requirements for deployment and real-time performance metrics on consumer hardware are not detailed in the abstract.
What different sources said
- arXiv cs.AICenter
Ex-Omni: Enabling 3D Facial Animation Generation for Omni-modal Large Language Models
Related
Topology-Aware Thermodynamics Improves DNA Probe Specificity Design
Researchers developed a new framework for designing DNA probes that accounts for the spatial organization of matched sequences, not just overall thermodynamic stability. Traditional methods rely on scalar measures like melting temperature and free energy, which miss how mismatches are distributed along the probe. The approach could improve diagnostic accuracy in applications like HPV detection and gene expression profiling.
Study Identifies Optimal Thermal Dose for Combining Focused Ultrasound with Immunotherapy in Tumors
Researchers used multimodal PET imaging to identify an optimal thermal dose range for focused ultrasound ablation that destroys tumor tissue while preserving conditions for immunotherapy delivery. The study found that excessive heating collapses blood vessels needed for antibody access, while insufficient heating fails to adequately reduce tumor burden. The findings could guide clinical design of combination treatments pairing thermal ablation with immunotherapies.
Plant MSH1 Protein Functions as Mismatch-Directed Nuclease for Organelle Genome Maintenance
Researchers have identified the precise mechanism by which the AtMSH1 protein in Arabidopsis plants recognizes and cleaves DNA mismatches and lesions, preventing mutations in organellar genomes. The protein combines a DNA mismatch recognition module with a nuclease domain that makes staggered cuts at specific positions relative to DNA damage. This discovery explains how plants maintain unusually low mutation rates in their mitochondrial and chloroplast DNA compared to other eukaryotes.