Published Aug 4, 2020

Episode 46 - Qualcomm Technologies

Explore how Qualcomm Technologies is advancing AI and 5G integration to revolutionize device performance and consumer experiences, with innovations in Snapdragon enhancing smartphone, home tech, and network efficiency while reducing power consumption.
Episode Highlights
Eye on AI logo

Popular Clips

Episode Highlights

  • AI & 5G Synergy

    explains the symbiotic relationship between AI and 5G, emphasizing how each enhances the other. AI improves 5G by optimizing reception and throughput through contextual information, while 5G's low latency and high speed enable distributed intelligence from devices to the cloud 1. Snapdragon, Qualcomm's product line, integrates AI and 5G technologies into a single silicon piece, extending beyond handsets to automotive applications 2.

    5G makes AI better and AI makes 5G better. We are going to continue to add more and more intelligence to the device at the edge.

    ---

    This integration allows for seamless communication and enhanced user experiences, making devices smarter and more efficient.

       

    Enhancing 5G

    Qualcomm is at the forefront of integrating AI into 5G technologies to enhance network performance and user experience. highlights the importance of power efficiency in AI applications, particularly in optimizing antenna performance to conserve battery life 3. This involves AI algorithms that adapt to user habits, ensuring efficient power transmission and reception.

    We create an AI algorithm to understand when we can move to another antenna so it can transmit without any blocking.

    ---

    These advancements not only improve connectivity but also extend the battery life of devices, making them more reliable and user-friendly 4.

       

    Federated Learning

    Federated learning is a key aspect of Qualcomm's AI strategy, allowing for personalized user experiences without compromising privacy. explains how this technology enables on-device learning, such as improving typing predictions on Google Keyboard, while only sharing model weights with the cloud 5. This approach ensures that personal data remains secure while continuously enhancing the model's accuracy.

    It's really a group learning exercise which makes the experience continually better over time.

    ---

    Federated learning also finds applications in photo galleries, where it can recognize family members without sending images to the cloud, thus maintaining user privacy.

Related Episodes