The field of remote inference and edge intelligence is moving towards more efficient and accurate methods of data collection and processing. Researchers are exploring new approaches to optimize task-oriented Age of Information (AoI) functions, which is critical for real-time inference tasks. Another area of focus is the development of scalable and energy-efficient frameworks for Wireless Sensor Networks, leveraging techniques such as constructive interference to enable low-latency and low-energy data collection. Noteworthy papers include:

• A study on a two-modality scheduling problem to minimize ML model's inference error, which developed an index-based threshold policy and proved its optimality.
• A framework for Wireless Sensor Networks that leverages constructive interference, which has been extensively validated on a real-world testbed deployment.
• A novel approach to Hierarchical Inference Learning, which introduced two policies that achieve order-optimal regret and have low computational complexity.