K-means and K-medoids clustering algorithms are widely used for many practical applications. Original k-mean and kmedoids algorithms select initial centroids and medoids randomly that affect the quality of the resulting clusters and sometimes it generates unstable and empty clusters which are meaningless. The original k-means and k-mediods algorithm is computationally expensive and requires time proportional to the product of the number of data items, number of clusters and the number of iterations. The new approach for the k mean algorithm eliminates the deficiency of exiting k mean. It first calculates the initial centroids k as per requirements of users and then gives better, effective and stable cluster. It also takes less execution time because it elimi nates unnecessary distance computation by using previous iteration. The new approach for k-medoids selects initial k medoids systematically based on initial centroids. It generates stable clusters to improve accuracy.

In this paper, a deep reinforcement learning(DRL) based multimodal coaching model (DCM) for slot filling task in SLU is proposed. The DCM takes advantage of a DRL based model as a coach of the system to learn the wrong labeled slots with/without user's feedback, hence may further improve the performance of an SLU system. This DCM model is an improved model of the deep reinforcement learning based augmented tagging model as introduced in [1], by using a better DRL model with different rewards and adding in a user's feedback modal to achieve one-shot learning.The performance of DCM is evaluated on two datasets: one is the benchmark ATIS corpus dataset, another is our in-house dataset with three different domains. It shows that the new system gives a better performance than the current state-of-the-art results on ATIS by using DCM. Furthermore, we build a demo application to further explain how user's input can also be used as a real-time coach to improve model's performance even more.

We present a system, CRUISE, that guides ordinary software developers to build a high quality natural language understanding (NLU) engine from scratch. This is the fundamental step of building a new skill for personal assistants. Unlike existing solutions that require either developers or crowdsourcing to manually generate and annotate a large number of utterances, we design a hybrid rulebased and data-driven approach with the capability to iteratively generate more and more utterances. Our system only requires light human workload to iteratively prune incorrect utterances. CRUISE outputs a well trained NLU engine and a large scale annotated utterance corpus that third parties can use to develop their custom skills. Using both benchmark dataset and custom datasets we collected in realworld settings, we validate the high quality of CRUISE generated utterances via both competitive NLU performance and human evaluation. We also show the largely reduced human workload in terms of both cognitive load and human pruning time consumption.

In the western world,the prostate cancer is major cause of death in males. Magnetic Resonance Imaging (MRI) is widely used for the detection of prostate cancer due to which it is an open area of research. The proposed method uses deep learning framework for the detection of prostate cancer using the concept of Gleason grading of the historical images. A3D convolutional neural network has been used to observe the affected region and predicting the affected region with the help of Epithelial and the Gleason grading network. The proposed model has performed the state-of-art while detecting epithelial and the Gleason score simultaneously. The performance has been measured by considering all the slices of MRI, volumes of MRI with the test fold, and segmenting prostate cancer with help of Endorectal Coil for collecting the images of MRI of the prostate 3D CNN network. Experimentally, it was observed that the proposed deep learning approach has achieved overall specificity of 85% with an accuracy of 87% and sensitivity 89% over the patient-level for the different targeted MRI images of the challenge of the SPIE-AAPM-NCI Prostate dataset.

No abstract