DeepSig, a pioneer in AI-native wireless communications, has introduced the OmniSIG Model Hub, this innovative, user-friendly platform is a central repository for models that power the OmniSIG Engine—the AI signal processing and inference software formerly known as OmniSIG Sensor—which applies artificial intelligence to radio systems, unlocking new levels of performance in radio frequency sensing and awareness. The Model Hub offers developers, researchers and operators a streamlined and accessible location to store, manage and retrieve pre-trained models and secure their own customized, proprietary models.
“By offering a central model repository, we accelerate the ability of users to leverage powerful deep learning-driven AI/ML spectrum sensing with new signals and bands. This allows people to develop spectrum-aware systems and applications more quickly, without needing to build, curate, and train their own models to get up and running for a wide variety of use cases. In the future, we envision DeepSig’s Model Hub extending to vetted OmniPHY AI models and collaboration for 5G, 6G and RAN Digital Twin applications,” says DeepSig’s CTO, Tim O’Shea.
The Model Hub includes numerous models that are developed and vetted by DeepSig. Existing and new OmniSIG customers can access capabilities beyond OmniSIG Engine’s default model. This includes models for various commonly seen signal types, such as commercial drone signals, push-to-talk radios and other communications devices. Other key features include:
- Centralized model management: Search, manage and share pre-trained and custom models in one secure location.
- Advanced search: Find models faster with detailed criteria like signal types, frequency ranges or training datasets.
- Certified models: Download and deploy tested and validated DeepSig-certified models.
- Enhanced collaboration: Accelerate development by collaborating with the OmniSIG community.
Empowering greater model collaboration among users enhances collective efforts, accelerates development and impact, reduces training costs and allows users to focus on fine-tuning and deploying models to meet specific needs rapidly.
With OmniSIG Engine, users can rapidly detect and identify a wide variety of RF emissions across large instantaneous bandwidths of spectrum in milliseconds. It enables near-real-time reporting of anomalies, changes and threats, thus enhancing operations’ efficiency and security and providing near-real-time input to a wide variety of spectrum-aware applications and systems.
The DeepSig Model Hub, available to OmniSIG Engine and OmniSIG Studio customers through the newly redesigned customer portal, ensures reproducibility, simplifies debugging and enhances model performance monitoring. It also gives OmniSIG users greater access and utility to off-the-shelf models that can quickly provide valuable spectrum insights. In addition, DeepSig’s redesigned customer portal offers a streamlined experience for performing license management, accessing training materials, and exploring relevant documentation.
Click here for more information on DeepSIG's OmniSIG Model Hub.
