Living Matter Lab
welcome to the living matter lab!
we integrate physics-based modeling with machine learning and create interactive simulation tools to understand, explore, and predict the dynamics of living systems
AI for food?
traditional attempts to transform the global food system are too slow to drive innovation at scale. can we accelerate this transition to a resilient global food systems that meet the urgent challenges of food security, climate change, and planetary health by leveraging AI to democratize food innovation?
what about fungi?
we observe a strong correlation between mechanical tests and sensory surveys, and perceive fungi steak as more moist and more fibrous than animal and plant-based meats. fungi steak is an attractive, structurally equivalent, and sensorially superior alternative protein source that is healthier for people and for the planet.
tasting stiffness?
eating less meat is associated with a healthier body and planet, but plant-based meat substitutes fail to deliver the same sensory experience as animal meat. but can we really taste stiffness differences? and, if so, how can we design plant-based alternatives that match the stiffness of animal meat?
testing stiffness?
to create adequate mimics of animal meats, plant-based meats must match in mouthfeel, taste, and texture. we characterize the texture of five plant-based and three animal meats using texture profile analysis and rheology and report each product's elasticity, viscosity, and loss of integrity.
inspired by nature?
with more than 90,000 muscle fascicles, the elephant trunk is a complex biological structure that is fascinating scientists across all disciplines. we design reduced-order models of the elephant trunk that can, within a fraction of a second, predict the trunk's motion as a result of its muscular activity.
weft or warp?
textile fabrics have unique mechanical properties, which make them ideal candidates for many engineering and medical applications. we integrate mechanical testing and machine learning to automatically discover the best models for ultra-anisotropic knitted polypropylene fabrics.
mimicking mechanics?
we ask how well today's plant-based meat products mimc the texture of animal meat, review the current literature, and identify the urgent need for data sharing and more standardized rheological testing beyond the classical texture profile analysis
discovering uncertainty
understanding uncertainty is critical, especially when data are sparse and variations are large. we integrate concepts of Bayesian learning and constitutive neural networks to discovery interpretable models, parameters, and uncertainties that best explain soft matter systems
best-in-class modeling
we uncover distinct and unexpected features of rubber, brain, artificial meat, skin, and arteries using best-in-class modeling, a new efficient, robust, and easy-to-use strategy to discover the mechanical signatures of traditional and unconventional soft materials
discovering the mechanics of artificial meat
we explore artificial meat using tension, compression, and shear tests and discover the model and parameters that best explain its behavior to inform the design of more authentic meat substitutes that are sustainable and environmentally friendly while still offering the familiar taste and texture of traditional meat
a universal material subroutine
we design a universal material subroutine that automatically integrates novel constitutive models of varying complexity into non-linear finite element programs, and empowers all users--not just domain experts--to perform reliable engineering analyses of soft matter systems
automated model discovery for soft matter
we challenge the general belief that neural networks can teach us nothing about the physics of a material. we reverse engineer a new family of constitutive artificial neural networks that have the potential to induce a paradigm shift in constitutive modeling, from user-defined model selection to automated model discovery
soft robotics inspired by the elephant trunk
inspired by nature, we model, simulate, and build soft slender actuators for programmable multimodal deformation using the theory of active filaments, machine learning, and liquid crystal elastomers that activate a soft matter system by joule heating to guide the design and control of soft robotic systems
democratizing simulation through automation
we democratize constitutive modeling through automated model discovery, embedded in a universal material subroutine, to make scientific simulations accessible to a more inclusive and diverse community, and accelerate the design of new functional materials with tailored properties
discovering knowledge from massive data
we discover interpretable and predictive models that provide simple relationships among scientific variables by integrating Lp regularization for subset selection with constitutive neural networks, and induce sparsity by a hybrid approach that combines controlled regularization and physical constraints
automated model discovery for human brain
we propose a new strategy to simultaneously discover both model and parameters that best characterize the behavior of human gray and white matter tissue. our discovered model outperforms existing models in tension, compression, and shear tests
integrating bayesian inference, neural networks, and physics
we integrate data, physics, and uncertainties by combining neural networks, physics informed modeling, and bayesian inference to improve the predictive potential of traditional neural network models.
amyloid-beta drives tau pathology
we personalize a network diffusion model using longitudinal tau pet data of 76 subjects and apply bayesian inference with hierarchical priors to identify correlations between amyloid beta and tau and infer personalized tau diffusion and production rates
sex matters!
we integrate multiscale modeling and machine learning to gain mechanistic insight into the sex-specific origin of drug-induced cardiac arrhythmias and show that sex differences in ion channel activity, tissue conductivity, and heart dimensions put females at higher arrhythmogenic risk than males
can we reverse engineer an elephant trunk?
we explore the active filament theory that efficiently correlates fiber stretch and orientation to the intrinsic curvature of slender structures to robustly solve inverse problems in soft robotics inspired by natural manipulators such as the elephant trunk
precision medicine in human heart modeling
with a view towards precision medicine, we integrate human heart electrophysiology, solid mechanics, and fluid dynamics and explore clinical applications in drug development, pacing lead failure, heart failure, ventricular assist devices, and mitral valve repair
in the news
- improving plant-based meats national science foundation news
- can AI improve plant-based meats? stanford report
- the mechanical and sensory signature of plant-based and animal meat behind the paper
- the telltale heart washington post feature about the living heart project
- women’s heart disease is underdiagnosed, but new AI models can help remedy this lab manager
- new models improve heart disease risk prediction, especially for women news medical
- AI offers paradigm shift in study of brain injury stanford | news
- this week in AI TechCrunch
- can computational modeling help us understand Alzheimer's disease? the future of everything
- college campuses are COVID-19 superspreaders? US news & world report
- students develop computer models to test return-to-campus strategies stanford engineering
- the newfoundland story atlantic ctv news cbc radio canada
- how effective are travel bans? swiss public radio welt
- passionate scientist and triathlete interview
- stanford-led team simulates how alzheimer’s disease spreads through the brain stanford report