Understanding Natural Language Instructions for Fetching Daily Objects Using GAN-Based Multimodal Target-Source Classification
This work addresses the challenge of multimodal language understanding for unconstrained fetching instructions in domestic service robots, representing an incremental improvement.
The paper tackles the problem of interpreting natural language instructions for domestic robots to fetch objects by predicting target and source regions in a scene, and reports that their proposed model outperforms the state-of-the-art method on a standard dataset.
In this paper, we address multimodal language understanding for unconstrained fetching instruction in domestic service robots context. A typical fetching instruction such as "Bring me the yellow toy from the white shelf" requires to infer the user intention, that is what object (target) to fetch and from where (source). To solve the task, we propose a Multimodal Target-source Classifier Model (MTCM), which predicts the region-wise likelihood of target and source candidates in the scene. Unlike other methods, MTCM can handle regionwise classification based on linguistic and visual features. We evaluated our approach that outperformed the state-of-the-art method on a standard data set. In addition, we extended MTCM with Generative Adversarial Nets (MTCM-GAN), and enabled simultaneous data augmentation and classification.