The Wall Lab at Stanford University

Menu

Projects


Gamified Data Collection & Digital Intervention

The Wall Lab at Stanford is committed to developing AI-enhanced and data backed games that can improve early detection and intervention options for complex conditions that originate in childhood and perpetuate through the life course, including autism and related developmental delays.



AI Model Development

We're on a mission to pioneer early detection and enhance outcomes for children with ASD and related developmental delays. Rooted in our mission we explore all frontiers spanning from detection to improving algorithmic performance and building robust datasets. By embracing these diverse fronts, we're committed to uncovering insights that can make a tangible difference.

Overview on current research:

  • Ensemble Learning

    Ensemble learning offers a versatile framework for leveraging the strengths of multiple models. Using ensemble learning approaches we can identify subgroups within the spectrum by clustering individuals based on their characteristics or response patterns, thereby enabling more personalized interventions or treatments.

  • Crowdsourcing for NT data

    Limited data set size can compromise the generalizability and robustness of the AI models, particularly in health-critical fields like digital phenotyping that require minimal bias and optimal performance. Crowdsourcing enables one to tap into the collective wisdom and capabilities of a diverse group of individuals, fostering collaboration, and problem-solving on a global scale. It offers a flexible and scalable approach to addressing various challenges.

  • Temporal Adaptation

    We introduce Temporal consistency for Test-time adaptation (TempT), a novel method for test-time adaptation on videos through the use of temporal coherence of predictions across sequential frames as a self-supervision signal. TempT is an approach with broad potential applications in computer vision tasks, including facial expression recognition (FER) in videos.