arXiv subjects

Dinkar Sitaram

Publications and source records attributed to Dinkar Sitaram.

7 recordsLinked to original sources

arXiv2024

Initial Decoding with Minimally Augmented Language Model for Improved Lattice Rescoring in Low Resource ASR

This paper addresses the problem of improving speech recognition accuracy with lattice rescoring in low-resource languages where the baseline language model is insufficient for generating inclusive lattices. We minimally augment the baseline language model with word unigram counts that are present in a larger text corpus of the target language but absent in the baseline. The lattices generated after decoding with such an augmented baseline language model are more comprehensive. We obtain 21.8% (Telugu) and 41.8% (Kannada) relative word error reduction with our proposed method. This reduction in word error rate is comparable to 21.5% (Telugu) and 45.9% (Kannada) relative word error reduction obtained by decoding with full Wikipedia text augmented language mode while our approach consumes only 1/8th the memory. We demonstrate that our method is comparable with various text selection-based language model augmentation and also consistent for data sets of different sizes. Our approach is applicable for training speech recognition systems under low resource conditions where speech data and compute resources are insufficient, while there is a large text corpus that is available in the target language. Our research involves addressing the issue of out-of-vocabulary words of the baseline in general and does not focus on resolving the absence of named entities. Our proposed method is simple and yet computationally less expensive.

eess.AS↗

arXiv2022

Megha: Decentralized Global Fair Scheduling for Federated Clusters

Increasing scale and heterogeneity in data centers have led to the development of federated clusters such as KubeFed, Hydra, and Pigeon, that federate individual data center clusters. In our work, we introduce Megha, a novel decentralized resource management framework for such federated clusters. Megha employs flexible logical partitioning of clusters to distribute its scheduling load, ensuring that the requirements of the workload are satisfied with very low scheduling overheads. It uses a distributed global scheduler that does not rely on a centralized data store but, instead, works with eventual consistency, unlike other schedulers that use a tiered architecture or rely on centralized databases. Our experiments with Megha show that it can schedule tasks taking into account fairness and placement constraints with low resource allocation times - in the order of tens of milliseconds.

cs.DC↗

arXiv2022

A Hierarchical Approach to Conditional Random Fields for System Anomaly Detection

Anomaly detection to recognize unusual events in large scale systems in a time sensitive manner is critical in many industries, eg. bank fraud, enterprise systems, medical alerts, etc. Large-scale systems often grow in size and complexity over time, and anomaly detection algorithms need to adapt to changing structures. A hierarchical approach takes advantage of the implicit relationships in complex systems and localized context. The features in complex systems may vary drastically in data distribution, capturing different aspects from multiple data sources, and when put together provide a more complete view of the system. In this paper, two datasets are considered, the 1st comprising of system metrics from machines running on a cloud service, and the 2nd of application metrics from a large-scale distributed software system with inherent hierarchies and interconnections amongst its system nodes. Comparing algorithms, across the changepoint based PELT algorithm, cognitive learning-based Hierarchical Temporal Memory algorithms, Support Vector Machines and Conditional Random Fields provides a basis for proposing a Hierarchical Global-Local Conditional Random Field approach to accurately capture anomalies in complex systems across various features. Hierarchical algorithms can learn both the intricacies of specific features, and utilize these in a global abstracted representation to detect anomalous patterns robustly across multi-source feature data and distributed systems. A graphical network analysis on complex systems can further fine-tune datasets to mine relationships based on available features, which can benefit hierarchical models. Furthermore, hierarchical solutions can adapt well to changes at a localized level, learning on new data and changing environments when parts of a system are over-hauled, and translate these learnings to a global view of the system over time.

cs.LG↗

arXiv2022

Learning Algorithms in Static Analysis of Web Applications

Web applications are distributed applications, they are programs that run on more than one computer and communicate through a network or server. This very distributed nature of web applications, combined with the scale and sheer complexity of modern software systems complicate manual security auditing, while also creating a huge attack surface of potential hackers. These factors are making automated analysis a necessity. Static Application Security Testing (SAST) is a method devised to automatically analyze application source code of large code bases without compiling it, and design conditions that are indicative of security vulnerabilities. However, the problem lies in the fact that the most widely used Static Application Security Testing Tools often yield unreliable results, owing to the false positive classification of vulnerabilities grossly outnumbering the classification of true positive vulnerabilities. This is one of the biggest hindrances to the proliferation of SAST testing, which leaves the user to review hundreds, if not thousands, of potential warnings, and classify them as either actionable or spurious. We try to minimize the problem of false positives by introducing a technique to filter the output of SAST tools. The aim of the project is to apply learning algorithms to the output by analyzing the true and false positives classified by OWASP Benchmark, and eliminate, or reduce the number of false positives presented to the user of the SAST Tool.

cs.CR↗

arXiv2019

Flavour Enhanced Food Recommendation

We propose a mechanism to use the features of flavour to enhance the quality of food recommendations. An empirical method to determine the flavour of food is incorporated into a recommendation engine based on major gustatory nerves. Such a system has advantages of suggesting food items that the user is more likely to enjoy based upon matching with their flavour profile through use of the taste biological domain knowledge. This preliminary intends to spark more robust mechanisms by which flavour of food is taken into consideration as a major feature set into food recommendation systems. Our long term vision is to integrate this with health factors to recommend healthy and tasty food to users to enhance quality of life.

cs.SI↗

arXiv2014

Efficient Support of Big Data Storage Systems on the Cloud

Due to its advantages over traditional data centers, there has been a rapid growth in the usage of cloud infrastructures. These include public clouds (e.g., Amazon EC2), or private clouds, such as clouds deployed using OpenStack. A common factor in many of the well known infrastructures, for example OpenStack and CloudStack, is that networked storage is used for storage of persistent data. However, traditional Big Data systems, including Hadoop, store data in commodity local storage for reasons of high performance and low cost. We present an architecture for supporting Hadoop on Openstack using local storage. Subsequently, we use benchmarks on Openstack and Amazon to show that for supporting Hadoop, local storage has better performance and lower cost. We conclude that cloud systems should support local storage for persistent data (in addition to networked storage) so as to provide efficient support for Hadoop and other Big Data systems

cs.DC↗

arXiv2012

Combining Heterogeneous Classifiers for Relational Databases

Most enterprise data is distributed in multiple relational databases with expert-designed schema. Using traditional single-table machine learning techniques over such data not only incur a computational penalty for converting to a 'flat' form (mega-join), even the human-specified semantic information present in the relations is lost. In this paper, we present a practical, two-phase hierarchical meta-classification algorithm for relational databases with a semantic divide and conquer approach. We propose a recursive, prediction aggregation technique over heterogeneous classifiers applied on individual database tables. The proposed algorithm was evaluated on three diverse datasets, namely TPCH, PKDD and UCI benchmarks and showed considerable reduction in classification time without any loss of prediction accuracy.

cs.LG↗