AutoDispNet: Improving Disparity Estimation With AutoML

By Tonmoy Saikia, Yassine Marrakchi, Arber Zela, Frank Hutter, Thomas Brox. 2019

85 citations recorded by the source corpus.

Much research work in computer vision is being spent on optimizing existing network architectures to obtain a few more percentage points on benchmarks. Recent AutoML approaches promise to relieve us from this effort. However, they are mainly designed for comparatively small-scale classification tasks. In this work, we show how to use and extend existing AutoML techniques to efficiently optimize large-scale U-Net-like encoder-decoder architectures. In particular, we leverage gradient-based neural architecture search and Bayesian optimization for hyperparameter search. The resulting optimization does not require a large-scale compute cluster. We show results on disparity estimation that clearly outperform the manually optimized baseline and reach state-of-the-art performance.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A Comprehensive Survey of Neural Architecture Search (2020, methodology): “AutoDispNet [37] proposes to apply the automatic architecture search technology to optimize large-scale U-Net-like encoder-decoder architectures.”
Understanding and Robustifying Differentiable Architecture Search (2019, methodology): “We base this search space on AutoDispNet (Saikia et al., 2019), which used DARTS for a space containing normal, downsampling and upsampling cells.”
Best Practices for Scientific Research on Neural Architecture Search (2019, background): “, 2019a), disparity estimation (Saikia et al., 2019), machine translation (So et al.”
Hierarchical Neural Architecture Search for Deep Stereo Matching (2020, methodology): “[25], not only in accuracy but also in inference time and the size of the resulting network (as shown in Figure 1).”
NAS-Bench-301 and the Case for Surrogate Benchmarks for Neural Architecture Search (2020, background): “…the ﬁeld by removing human bias from architecture design (see, e.g., Elsken et al. (2019b)), achieve new state of the art on many tasks (Real et al., 2019; Chenxi et al., 2019; Saikia et al., 2019) and create resource-aware architectures (Tan et al., 2018; Elsken et al., 2019a; Cai et al., 2020).”
Optimized Generic Feature Learning for Few-shot Classification across Domains (2020, methodology): “In computer vision, BOHB was used to optimize training a disparity network [47].”