AI Should Sense Better, Not Just Scale Bigger: Adaptive Sensing as a Paradigm Shift
Eunsu Baek, Keondo Park, Jeonggil Ko, Min-hwan Oh, Taesik Gong, Hyung-Sin Kim
TL;DR
The paper argues that the traditional model-centric AI paradigm is unsustainable and inequitable, proposing adaptive sensing as a foundational shift to modulate inputs at the sensor level. It introduces a closed-loop, multimodal sensing framework that integrates sensing, perception, and action, supported by theoretical formulations and the Lens prototype demonstrating large gains for small models. Key contributions include a formal RL/non-RL framework for embodied AI with sensing policies, identified research directions, and open challenges related to benchmarks, metrics, and privacy. The work emphasizes that refining how we sense—before learning—can yield greener, more robust, and more accessible AI systems across real-world domains. Overall, adaptive sensing is presented as a practical, impactful avenue to complement existing scaling approaches and broaden AI’s societal benefits.
Abstract
Current AI advances largely rely on scaling neural models and expanding training datasets to achieve generalization and robustness. Despite notable successes, this paradigm incurs significant environmental, economic, and ethical costs, limiting sustainability and equitable access. Inspired by biological sensory systems, where adaptation occurs dynamically at the input (e.g., adjusting pupil size, refocusing vision)--we advocate for adaptive sensing as a necessary and foundational shift. Adaptive sensing proactively modulates sensor parameters (e.g., exposure, sensitivity, multimodal configurations) at the input level, significantly mitigating covariate shifts and improving efficiency. Empirical evidence from recent studies demonstrates that adaptive sensing enables small models (e.g., EfficientNet-B0) to surpass substantially larger models (e.g., OpenCLIP-H) trained with significantly more data and compute. We (i) outline a roadmap for broadly integrating adaptive sensing into real-world applications spanning humanoid, healthcare, autonomous systems, agriculture, and environmental monitoring, (ii) critically assess technical and ethical integration challenges, and (iii) propose targeted research directions, such as standardized benchmarks, real-time adaptive algorithms, multimodal integration, and privacy-preserving methods. Collectively, these efforts aim to transition the AI community toward sustainable, robust, and equitable artificial intelligence systems.