-
Multi-Perspective Anomaly Detection
Authors:
Peter Jakob,
Manav Madan,
Tobias Schmid-Schirling,
Abhinav Valada
Abstract:
Anomaly detection is a critical problem in the manufacturing industry. In many applications, images of objects to be analyzed are captured from multiple perspectives which can be exploited to improve the robustness of anomaly detection. In this work, we build upon the deep support vector data description algorithm and address multi-perspective anomaly detection using three different fusion techniq…
▽ More
Anomaly detection is a critical problem in the manufacturing industry. In many applications, images of objects to be analyzed are captured from multiple perspectives which can be exploited to improve the robustness of anomaly detection. In this work, we build upon the deep support vector data description algorithm and address multi-perspective anomaly detection using three different fusion techniques, i.e., early fusion, late fusion, and late fusion with multiple decoders. We employ different augmentation techniques with a denoising process to deal with scarce one-class data, which further improves the performance (ROC AUC $= 80\%$). Furthermore, we introduce the dices dataset, which consists of over 2000 grayscale images of falling dices from multiple perspectives, with 5\% of the images containing rare anomalies (e.g., drill holes, sawing, or scratches). We evaluate our approach on the new dices dataset using images from two different perspectives and also benchmark on the standard MNIST dataset. Extensive experiments demonstrate that our proposed {multi-perspective} approach exceeds the state-of-the-art {single-perspective anomaly detection on both the MNIST and dices datasets}. To the best of our knowledge, this is the first work that focuses on addressing multi-perspective anomaly detection in images by jointly using different perspectives together with one single objective function for anomaly detection.
△ Less
Submitted 9 August, 2021; v1 submitted 20 May, 2021;
originally announced May 2021.
-
Early Bird Catches the Worm: Predicting Returns Even Before Purchase in Fashion E-commerce
Authors:
Sajan Kedia,
Manchit Madan,
Sumit Borar
Abstract:
With the rapid growth in fashion e-commerce and customer-friendly product return policies, the cost to handle returned products has become a significant challenge. E-tailers incur huge losses in terms of reverse logistics costs, liquidation cost due to damaged returns or fraudulent behavior. Accurate prediction of product returns prior to order placement can be critical for companies. It can facil…
▽ More
With the rapid growth in fashion e-commerce and customer-friendly product return policies, the cost to handle returned products has become a significant challenge. E-tailers incur huge losses in terms of reverse logistics costs, liquidation cost due to damaged returns or fraudulent behavior. Accurate prediction of product returns prior to order placement can be critical for companies. It can facilitate e-tailers to take preemptive measures even before the order is placed, hence reducing overall returns. Furthermore, finding return probability for millions of customers at the cart page in real-time can be difficult. To address this problem we propose a novel approach based on Deep Neural Network. Users' taste & products' latent hidden features were captured using product embeddings based on Bayesian Personalized Ranking (BPR). Another set of embeddings was used which captured users' body shape and size by using skip-gram based model. The deep neural network incorporates these embeddings along with the engineered features to predict return probability. Using this return probability, several live experiments were conducted on one of the major fashion e-commerce platform in order to reduce overall returns.
△ Less
Submitted 28 June, 2019;
originally announced June 2019.
-
Multi Agent Driven Data Mining For Knowledge Discovery in Cloud Computing
Authors:
Vishal Jain,
Mahesh Kumar Madan
Abstract:
Today, huge amount of data is available on the web. Now there is a need to convert that data in knowledge which can be useful for different purposes. This paper depicts the use of data mining process, OLAP with the combination of multi agent system to find the knowledge from data in cloud computing. For this, I am also trying to explain one case study of online shopping of one Bakery Shop. May be…
▽ More
Today, huge amount of data is available on the web. Now there is a need to convert that data in knowledge which can be useful for different purposes. This paper depicts the use of data mining process, OLAP with the combination of multi agent system to find the knowledge from data in cloud computing. For this, I am also trying to explain one case study of online shopping of one Bakery Shop. May be we can increase the sale of items by using the model, which I am trying to represent.
△ Less
Submitted 21 March, 2017;
originally announced March 2017.
-
Cloud Network Management Model A Novel Approach to Manage Cloud Traffic
Authors:
Mamta Madan,
Mohit Mathur
Abstract:
Cloud is in the air. More and More companies and personals are connecting to cloud with so many variety of offering provided by the companies. The cloud services are based on Internet i.e. TCP IP. The paper discusses limitations of one of the main existing network management protocol i.e. Simple Network Management Protocol (SNMP) with respect to the current network conditions. The network traffic…
▽ More
Cloud is in the air. More and More companies and personals are connecting to cloud with so many variety of offering provided by the companies. The cloud services are based on Internet i.e. TCP IP. The paper discusses limitations of one of the main existing network management protocol i.e. Simple Network Management Protocol (SNMP) with respect to the current network conditions. The network traffic is growing at a high speed. When we talk about the networked environment of cloud, the monitoring tool should be capable of handling the traffic tribulations efficiently and represent a correct scenario of the network condition. The proposed Model Cloud Network Management Model provides a comprehensive solution to manage the growing traffic in cloud and trying to improve communication of manager and agents as in SNMP (the traditional TCP IP network management protocol). Firstly CNMM concentrates on reduction of packet exchange between manager and agent. Secondly it eliminates the counter problems exist in SNMP by having periodic updates from agent without querying by the manager. For better management we are including managers using virtualized technology. CNMM is a proposed model with efficient communication, secure packet delivery and reduced traffic. Though the proposed model supposed to manage the cloud traffic in a better and efficient way, the model is still a theoretical study, its implementation and results are yet to discover. The model however is the first step towards development of supported algorithms and protocol. Our further study will concentrate on development of supported algorithms.
△ Less
Submitted 8 November, 2014;
originally announced November 2014.
