1: \begin{abstract}

2: Nonnegative Matrix Factorization (NMF) is the problem of approximating a given nonnegative matrix $\M$ through the conic combination of two nonnegative low-rank matrices $\W$ and $\H$. 

3: Traditionally NMF is tackled by optimizing a specific objective function evaluating the quality of the approximation. 

4: This assessment is often done based on the Frobenius norm.

5: In this study, we argue that the Frobenius norm as the ``point-to-point'' distance may not always be appropriate. 

6: Due to the nonnegative combination resulting in a polyhedral cone, this conic perspective of NMF may not naturally align with conventional point-to-point distance measures.

7: Hence, a ray-to-ray chordal distance is proposed as an alternative way of measuring the discrepancy between $\M$ and $\W\H$.

8: This measure is related to the Euclidean distance on the unit sphere, motivating us to employ nonsmooth manifold optimization approaches.

9: 

10: We apply Riemannian optimization technique to solve chordal-NMF by casting it on a manifold.

11: Unlike existing works on Riemannian optimization that require the manifold to be smooth, the nonnegativity in chordal-NMF is a non-differentiable manifold.

12: We propose a Riemannian Multiplicative Update (RMU) that preserves the convergence properties of Riemannian gradient descent without breaking the smoothness condition on the manifold.

13: 

14: We showcase the effectiveness of the Chordal-NMF on synthetic datasets as well as real-world multispectral images.

15: 

16: \keywords{Nonnegative Matrix Factorization 

17: \and Manifold 

18: \and Chordal distance 

19: \and Nonconvex Optimization

20: \and Multiplicative Update

21: \and Riemannian gradient

22: }

23: 

24: % \PACS{PACS code1 \and PACS code2 \and more}

25: \subclass{MSC 15A23 

26: \and MSC 78M50 

27: \and MSC 49Q99 

28: \and MSC 90C26

29: \and MSC 90C30

30: }

31: \end{abstract}

32: