The study of cross-domain mapping without supervision has recently attracted much attention. Much of the recent progress was enabled by the use of adversarial training as well as cycle constraints. In a recent paper, it was shown that cross-domain mapping is possible without the use of cycles or GANs. Although promising, this approach suffers from several drawbacks including costly inference and an optimization variable for every training example preventing the method from using large training sets. We present an alternative approach which is able to achieve non-adversarial mapping using a novel form of Variational Auto-Encoder.
The goal of this study was to investigate the translate-ability of creative works into other domains. We tested whether people were able to recognize which works of art were inspired by which piece of music. Three expert painters created four paintings, each of which was the artist's interpretation of one of four different pieces of instrumental music. Participants were able to identify which paintings were inspired by which pieces of music at significantly above-chance levels. The findings support the hypothesis that creative ideas can exist in an at least somewhat domain-independent state of potentiality and become more well-defined as they are actualized in accordance with the constraints of a particular domain.
In this paper, we consider cross-domain imitation learning (CDIL) in which an agent in a target domain learns a policy to perform well in the target domain by observing expert demonstrations in a source domain without accessing any reward function. In order to overcome the domain difference for imitation learning, we propose a dual-structured learning method. The proposed learning method extracts two feature vectors from each input observation such that one vector contains domain information and the other vector contains policy expertness information, and then enhances feature vectors by synthesizing new feature vectors containing both target-domain and policy expertness information. The proposed CDIL method is tested on several MuJoCo tasks where the domain difference is determined by image angles or colors. Numerical results show that the proposed method shows superior performance in CDIL to other existing algorithms and achieves almost the same performance as imitation learning without domain difference.
Xu, Xiaoxi (University of Massachusetts Amherst) | Murray, Tom (University of Massachusetts Amherst) | Woolf, Beverly Park (University of Massachusetts Amherst) | Smith, David A. (Northeastern University)
In this paper we describe automatic systems for identifying whether participants demonstrate social deliberative behavior within their online conversations. We test 3 corpora containing 2617 annotated segments. With machine learning models using linguistic features, we identify social deliberative behavior with up to 68.09% in-domain accuracy (com- pared to 50% baseline), 62.17% in-domain precision, and 84% in-domain recall. In cross-domain identification tasks, we achieve up to 55.56% cross-domain accuracy, 59.84% cross-domain precision, and 86.58% cross-domain recall. We also discover linguistic characteristics of social deliberative behavior. In the context of identifying social deliberative be- havior, we offer insights into why certain machine learning models generalize well across domains and why certain domains pose great challenges to machine learning models.
Hu, Liang (Shanghai Jiaotong University) | Cao, Jian (Shanghai Jiaotong University) | Xu, Guandong (University of Technology Sydney) | Wang, Jie (Stanford University) | Gu, Zhiping (Shanghai Technical Institute of Electronics &) | Cao, Longbing (Information)
Cross-domain collaborative filtering (CDCF), which aims to leverage data from multiple domains to relieve the data sparsity issue, is becoming an emerging research topic in recent years. However, current CDCF methods that mainly consider user and item factors but largely neglect the heterogeneity of domains may lead to improper knowledge transfer issues. To address this problem, we propose a novel CDCF model, the Bilinear Multilevel Analysis (BLMA), which seamlessly introduces multilevel analysis theory to the most successful collaborative filtering method, matrix factorization (MF). Specifically, we employ BLMA to more efficiently address the determinants of ratings from a hierarchical view by jointly considering domain, community, and user effects so as to overcome the issues caused by traditional MF approaches. Moreover, a parallel Gibbs sampler is provided to learn these effects. Finally, experiments conducted on a real-world dataset demonstrate the superiority of the BLMA over other state-of-the-art methods.