1: \begin{abstract}

2: Electrical Impedance Tomography (EIT) is widely applied in medical diagnosis, industrial inspection, and environmental monitoring. 

3: Combining the physical principles of the imaging system with the advantages of data-driven deep learning networks, physics-embedded deep unrolling networks have recently emerged as a promising solution in computational imaging. 

4: However, the inherent nonlinear and ill-posed properties of EIT image reconstruction still present challenges to existing methods in terms of accuracy and stability.

5: To tackle this challenge, we propose the learned half-quadratic splitting (HQSNet) algorithm for incorporating physics into learning-based EIT imaging. We then apply Anderson acceleration (AA) to the HQSNet algorithm, denoted as AA-HQSNet, which can be interpreted as AA applied to the Gauss-Newton step and the learned proximal gradient descent step of the HQSNet, respectively. 

6: AA is a widely-used technique for accelerating the convergence of fixed-point iterative algorithms and has gained significant interest in numerical optimization and machine learning. However, the technique has received little attention in the inverse problems community thus far. 

7: Employing AA enhances the convergence rate compared to the standard HQSNet

8: while simultaneously avoiding artifacts in the reconstructions.

9: Lastly, we conduct rigorous numerical and visual experiments to show that the AA module strengthens the HQSNet, leading to robust, accurate, and considerably superior reconstructions compared to state-of-the-art methods.

10: Our Anderson acceleration scheme to enhance HQSNet is generic and can be applied to improve the performance of various physics-embedded deep learning methods.

11: \keywords{Nonlinear inverse problems \and Algorithm unrolling \and Anderson acceleration \and Half-quadratic splitting \and Electrical impedance tomography.}

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

13: \subclass{78A46 \and 68U10 \and 68T07}

14: \end{abstract}

15: