ReTronics

Call For Papers

The On-Device Intelligence for Smart Interactions Workshop aims to explore the deployment of AI models directly onto embedded processors found in devices like microcontrollers, smartphones, and smartwatches. This on-device processing facilitates real-time data handling, reduces latency, and bolsters user privacy by eliminating the need for cloud-based computations. The workshop seeks to investigate novel input modalities that can revolutionize user interactions with mobile devices.

Key topics include but not limited to:

• Deployment Techniques:
  • Strategies for optimizing and implementing AI models on resource-constrained devices
  • AI inference on microcontrollers
  • Hardware acceleration for low-power, real-time inference on wearables and IoT devices
  • Energy-efficient AI architectures tailored for mobile and wearable devices
• Interaction Paradigms
  • Integration of sensors in textiles for interaction (e.g. capacitive sensing)
  • Gesture and motion sensing through e-textiles and flexible electronics
  • Muscle activity, and physiological sensing (e.g. through Bio-Impedance) for wearable AI applications
  • Adaptive and context-aware textile interfaces for mobile devices
  • Tangible interaction for sensing and actuation (e.g. haptics)
• Design Innovations:
  • User-centric design approaches for wearables and on-device AI
  • Evaluating the usability, ergonomics, and interaction quality of AI-driven smart inputs
  • Human factors and the cognitive load of wearable and textile-based interfaces
• Efficient Pipelines:
  • Optimizing the pipeline from sensor input to AI-based predictions with a focus on latency and usability
  • Low-power, event-driven AI models for continuous sensing and real-time feedback
• Ethical and Privacy Considerations:
  • Ethical considerations of deploying AI at the edge, including bias mitigation and fairness.
  • Privacy-preserving AI techniques for wearable applications.
  • Security risks and opportunities for on-device AI models and sensor data.
  • User trust in AI-driven feedback from wearables

Submissions should be concise, in the form of full or short papers:

• Full papers (up to 6 pages) should report reasonably mature work within the scope of ODISI and are expected to demonstrate concrete and reproducible results.
• Short papers (up to 3 pages) are encouraged to report novel and creative ideas that are yet to produce concrete research results but are at a stage where community feedback would be useful.

Submission Instructions

Deadline for submissions: July 21, 2025 

Notifications of acceptance: August 8, 2025

Submission Link: EasyChair

Submissions should follow the ACM Master Article Templates documentclass[sigconf]{acmart}.