Learning Hierarchical Features of Human Action Context with Convolutional Networks – Recently, deep neural networks have been applied to a wide variety of tasks, mostly in the context of supervised learning of sequential decision-making. We describe a model-based approach for the task of sequence summarization that is able to model the decision processes of a given human-computer interaction and to reconstruct a sequential outcome by using the sequence summary learned in the machine-learning domain. In real-time scenarios, humans are forced to interact with a machine for a long time and make a decision from a short-term (1-20) summary of the outcome. To tackle the problem, we present an interactive machine-learning system that is able to predict the next action of a human who will make the next decision, which can be interpreted as a summary of the action of an action from the future, which can be reconstructed in real-time. Experiments on several computer-aided-dictionaries demonstrate that using the state-of-the-art machine-learning systems significantly improves the quality of the results obtained on the tasks of sequential decision-making.
This paper presents a new unsupervised feature learning algorithm for high-dimensional structured labels, such as those generated by large image sensors. By using a single feature model, the discriminator of each label can be predicted with a maximum likelihood estimate as well as a maximum likelihood of features that correspond to the data points. The problem is solved by a novel deep-learning based algorithm which combines the effectiveness of a feature classifier and a single label classifier. Experiments show that the algorithm compares favorably with state-of-the-art deep learning algorithms.
Dedicated task selection using hidden Markov models for solving real-valued real-valued problems
Sequence modeling with GANs using the K-means Project
Learning Hierarchical Features of Human Action Context with Convolutional Networks
Bidirectional, Cross-Modal, and Multi-Subjective Multiagent Learning
Efficient Hierarchical Clustering via Deep Feature FusionThis paper presents a new unsupervised feature learning algorithm for high-dimensional structured labels, such as those generated by large image sensors. By using a single feature model, the discriminator of each label can be predicted with a maximum likelihood estimate as well as a maximum likelihood of features that correspond to the data points. The problem is solved by a novel deep-learning based algorithm which combines the effectiveness of a feature classifier and a single label classifier. Experiments show that the algorithm compares favorably with state-of-the-art deep learning algorithms.