Search for: All records

Abstract One possible solution against the accumulation of petrochemical plastics in natural environments is to develop biodegradable plastic substitutes using natural components. However, discovering all-natural alternatives that meet specific properties, such as optical transparency, fire retardancy and mechanical resilience, which have made petrochemical plastics successful, remains challenging. Current approaches still rely on iterative optimization experiments. Here we show an integrated workflow that combines robotics and machine learning to accelerate the discovery of all-natural plastic substitutes with programmable optical, thermal and mechanical properties. First, an automated pipetting robot is commanded to prepare 286 nanocomposite films with various properties to train a support-vector machine classifier. Next, through 14 active learning loops with data augmentation, 135 all-natural nanocomposites are fabricated stagewise, establishing an artificial neural network prediction model. We demonstrate that the prediction model can conduct a two-way design task: (1) predicting the physicochemical properties of an all-natural nanocomposite from its composition and (2) automating the inverse design of biodegradable plastic substitutes that fulfils various user-specific requirements. By harnessing the model’s prediction capabilities, we prepare several all-natural substitutes, that could replace non-biodegradable counterparts as exhibiting analogous properties. Our methodology integrates robot-assisted experiments, machine intelligence and simulation tools to accelerate the discovery and design of eco-friendly plastic substitutes starting from building blocks taken from the generally-recognized-as-safe database. 

One fundamental challenge in building an instance segmen- tation model for a large number of classes in complex scenes is the lack of training examples, especially for rare objects. In this paper, we ex- plore the possibility to increase the training examples without laborious data collection and annotation. We find that an abundance of instance segments can potentially be obtained freely from object-centric images, according to two insights: (i) an object-centric image usually contains one salient object in a simple background; (ii) objects from the same class often share similar appearances or similar contrasts to the background. Motivated by these insights, we propose a simple and scalable frame- work FreeSeg for extracting and leveraging these “free” object fore- ground segments to facilitate model training in long-tailed instance seg- mentation. Concretely, we investigate the similarity among object-centric images of the same class to propose candidate segments of foreground instances, followed by a novel ranking of segment quality. The resulting high-quality object segments can then be used to augment the exist- ing long-tailed datasets, e.g., by copying and pasting the segments onto the original training images. Extensive experiments show that FreeSeg yields substantial improvements on top of strong baselines and achieves state-of-the-art accuracy for segmenting rare object categories. Our code is publicly available at https://github.com/czhang0528/FreeSeg. 

In this paper, we explore the possibility to increase the training examples without laborious data collection and annotation for long-tailed instance segmentation. We find that an abundance of instance segments can potentially be obtained freely from object-centric images, according to two insights: (i) an object-centric image usually contains one salient object in a simple background; (ii) objects from the same class often share similar appearances or similar contrasts to the background. Motivated by these insights, we propose a simple and scalable framework FREESEG for extracting and leveraging these “free” object segments to facilitate model training. Concretely, we investigate the similarity among object-centric images of the same class to propose candidate segments of foreground instances, followed by a novel ranking of segment quality. The resulting high quality object segments can then be used to augment the existing long-tailed datasets, e.g., by copying and pasting the segments onto the original training images. Extensive experiments show that FREESEG yields substantial improvements on top of strong baselines and achieves state-of-the-art accuracy for segmenting rare object categories.