Conference Papers

Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506

Browse

Author: search.filters.author.Tripathi, A.Has files: No

Search Results

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    Computationally efficient fault tolerant ANTS
    (Association for Computing Machinery acmhelp@acm.org, 2016) Tripathi, A.; Maheshwari, A.; Chandrasekaran, K.
    In this paper, we formulate a method to utilize n mobile agents to solve a variant of Ants Nearby Treasure Search problem (ANTS), where an adversary can place treasure at any cell at a distance D from the origin. We devise a method which finds the treasure with the time complex-ity of O(D + D2=n + Df) where D is the Manhattan dis-tance of the treasure from the source and f is the maximum number of failures such that f 2 o(n). The algorithm is specially designed to reduce computation complexity of the distributed system as a whole by efficiently handling fail-ures and also, introducing the elements of parallelism with respect to handling failures. Using our algorithm, we bring down the computation cost/complexity of the system by an order of n, when failures occur, where n is the total number of ants. ANTS problem utilizes the multi-Agent system with self-organization and steering based on a control mechanism which is analogous to the problem of discovering resources that are available to the distributed system. © 2016 ACM.
  • No Thumbnail Available
    Item
    Normalized videosnapping: A non-linear video synchronization approach
    (Institute of Electrical and Electronics Engineers Inc., 2017) Tripathi, A.; Changmai, B.; Habib, S.; Chittaragi, N.B.; Koolagudi, S.G.
    Video synchronization is the task of content-based alignment of two or more videos depicting the same event with spatial variations or in the same object with temporal changes. Video synchronization is one of the most fundamental tasks when it comes to manipulations with temporally or spatially multi-perspective video-shots. In this paper, a model is proposed to deal with the synchronization problem and efficiently tackles issues arising during synchronizing two videos. Here, videos are dealt, at the frame level with features from each frame forming the basis of alignment. Features are matched and mapped to generate a cost matrix of similarities among the frames of the videos in concern. A modified version of Djikstra's algorithm that yields an optimal path through the matrix is applied. Through an optimal path, events are grouped into adjacent regions following which temporal warpings are introduced into the videos to achieve the best possible alignment among them. The model has proven to be efficient and compatible with all classes of quality levels of videos. © 2017 IEEE.
  • No Thumbnail Available
    Item
    A reinforcement learning and recurrent neural network based dynamic user modeling system
    (Institute of Electrical and Electronics Engineers Inc., 2018) Tripathi, A.; Ashwin, T.S.; Guddeti, R.M.
    With the exponential growth in areas of machine intelligence, the world has witnessed promising solutions to the personalized content recommendation. The ability of interactive learning agents to take optimal decisions in dynamic environments has been very well conceptualized and proven by Reinforcement Learning (RL). The learning characteristics of Deep-Bidirectional Recurrent Neural Networks (DBRNN) in both positive and negative time directions has shown exceptional performance as generative models to generate sequential data in supervised learning tasks. In this paper, we harness the potential of the said two techniques and strive to create personalized video recommendation through emotional intelligence by presenting a novel context-Aware collaborative filtering approach where intensity of users' spontaneous non-verbal emotional response towards recommended video is captured through system-interactions and facial expression analysis for decision-making and video corpus evolution with real-Time data streams. We take into account a user's dynamic nature in the formulation of optimal policies, by framing up an RL-scenario with an off-policy (Q-Learning) algorithm for temporal-difference learning, which is used to train DBRNN to learn contextual patterns and generate new video sequences for the recommendation. Evaluation of our system with real users for a month shows that our approach outperforms state-of-The-Art methods and models a user's emotional preferences very well with stable convergence. © 2018 IEEE.