Modeling/simulation people
--Curse of dimensionality (it's been lifted)
--you should not be afraid of musculoskeletal modeling at scale
Cost function people
--you need to test your algorithms in a more visible space
--consider the cost gradients across the fitness landscape (cost functions distort the landscape)
--refer to optimization algorithms in terms of robustness
Synergy people
--adding synergies will only constrain this space further
--you can model synergies as basis functions, but how does the CNS stay in the space
--in the future, 'synergists' should focus on exploitation of families of solutions
--synergies are exploiting these landscapes (see Racz & FVC)
Learning people
--Now they have a way to view how these landscapes change.
--Motor learning is (at its essence) focused on finding and exploring these subspaces- now we've opened the ability for them to track and visualize them.
Stochastic Motor Control
--Histograms offer a joint distribution of solutions that serve as the hard priors for bayesian motor control.
Data mining/ML people
--cost functions are now multidimensional
--parallel coordinates offer a quick way to make decisions
--now that we can enter these spaces, how do we visualize them? that remains a challenge. fruitful avenue for research.
----------Broader impacts ___
effects on each subgroup, and how it will inspire