arXiv subjects

Abhik Ray

Publications and source records attributed to Abhik Ray.

4 recordsLinked to original sources

arXiv2020

SIBRE: Self Improvement Based REwards for Adaptive Feedback in Reinforcement Learning

We propose a generic reward shaping approach for improving the rate of convergence in reinforcement learning (RL), called Self Improvement Based REwards, or SIBRE. The approach is designed for use in conjunction with any existing RL algorithm, and consists of rewarding improvement over the agent's own past performance. We prove that SIBRE converges in expectation under the same conditions as the original RL algorithm. The reshaped rewards help discriminate between policies when the original rewards are weakly discriminated or sparse. Experiments on several well-known benchmark environments with different RL algorithms show that SIBRE converges to the optimal policy faster and more stably. We also perform sensitivity analysis with respect to hyper-parameters, in comparison with baseline RL algorithms.

cs.LG↗

arXiv2016

Connecting Data Science and Qualitative Interview Insights through Sentiment Analysis to Assess Migrants' Emotion States Post-Settlement

Large-scale survey research by social scientists offers general understandings of migrants' challenges and provides assessments of post-migration benchmarks like employment, obtention of educational credentials, and home ownership. Minimal research, however, probes the realm of emotions or "feeling states" in migration and settlement processes, and it is often approached through closed-ended survey questions that superficially assess feeling states. The evaluation of emotions in migration and settlement has been largely left to qualitative researchers using in-depth, interpretive methods like semi-structured interviewing. This approach also has major limitations, namely small sample sizes that capture limited geographic contexts, heavy time burdens analyzing data, and limits to analytic consistency given the nuances of qualitative data coding. Information about migrant emotion states, however, would be valuable to governments and NGOs to enable policy and program development tailored to migrant challenges and frustrations, and would thereby stimulate economic development through thriving migrant populations. In this paper, we present an interdisciplinary pilot project that offers a way through the methodological impasse by subjecting exhaustive qualitative interviews of migrants to sentiment analysis using the Python NLTK toolkit. We propose that data scientists can efficiently and accurately produce large-scale assessments of migrant feeling states through collaboration with social scientists.

cs.CY↗

arXiv2013

StreamWorks - A system for Dynamic Graph Search

Acting on time-critical events by processing ever growing social media, news or cyber data streams is a major technical challenge. Many of these data sources can be modeled as multi-relational graphs. Mining and searching for subgraph patterns in a continuous setting requires an efficient approach to incremental graph search. The goal of our work is to enable real-time search capabilities for graph databases. This demonstration will present a dynamic graph query system that leverages the structural and semantic characteristics of the underlying multi-relational graph.

cs.DB↗

arXiv2012

Continuous Queries for Multi-Relational Graphs

Acting on time-critical events by processing ever growing social media or news streams is a major technical challenge. Many of these data sources can be modeled as multi-relational graphs. Continuous queries or techniques to search for rare events that typically arise in monitoring applications have been studied extensively for relational databases. This work is dedicated to answer the question that emerges naturally: how can we efficiently execute a continuous query on a dynamic graph? This paper presents an exact subgraph search algorithm that exploits the temporal characteristics of representative queries for online news or social media monitoring. The algorithm is based on a novel data structure called the Subgraph Join Tree (SJ-Tree) that leverages the structural and semantic characteristics of the underlying multi-relational graph. The paper concludes with extensive experimentation on several real-world datasets that demonstrates the validity of this approach.

cs.DB↗