1: \begin{abstract}

2:  %

3: The data-driven computing paradigm initially introduced by Kirchdoerfer \& Ortiz (2016)  \citep{kirchdoerfer2016data} is extended by incorporating locally linear tangent spaces into the data set.

4:  %

5: These tangent spaces are constructed by means of the tensor voting method introduced by  Mordohai \& Medioni (2010) \citep{mordohai2010dimensionality} which improves the learning of the underlying structure of a data set.

6:  %

7: Tensor voting is an instance-based machine learning technique which accumulates votes from the nearest neighbors to build up second-order tensors encoding tangents and normals to the underlying data structure.

8:  %

9: The here proposed second-order data-driven paradigm is a plug-in method for distance-minimizing as well as entropy-maximizing data-driven schemes.

10:  %

11: Like its predecessor  \citep{kirchdoerfer2016data}, the resulting method aims to minimize a suitably defined free energy over phase space subject to compatibility and equilibrium constraints. 

12:  %

13: The method's implementation is straightforward and numerically efficient since the data structure analysis is performed in an offline step.

14:  %

15: Selected numerical examples are presented that establish the higher-order convergence properties of the data-driven solvers enhanced by tensor voting for ideal and noisy data sets. 

16: \end{abstract}

17: