Department of Computer Science and Automation

Indian Institute of Science, Bangalore, India

Prakhar Ojha

MSc.Engg, Research Student
Intelligent Systems Group
Machine and Language Learning Lab

News and Activities

Research Interests

Machine Learning: Inference, Knowledge Graphs, Knowledge Representation
Crowdsourcing: Collaborative tasks, Team formation
Applications: Question Answering, Network analysis


MSc. Engg,   Dept. of Computer Science & Automation     Aug 2014 - Present
Indian Institute of Science, IISc Bangalore
Advisor: Prof. Partha Talukdar
Thesis: Utilizing Worker Groups And Task Dependencies In Crowdsourcing

B.Tech,   Dept. of Computer Science Engineering     Jul 2010 - May 2014
National Institute of Technology Karnataka, NITK Surathkal
Advisor: Prof. Vani M
Thesis: Reinforcement Learning based Schedulers


  1. Prakhar Ojha, and Partha Talukdar. Quality Estimation of Workers in Collaborative Crowdsourcing Using Group Testing. The Fourth AAAI Conference on Human Computation and Crowdsourcing (HCOMP) 2016, Austin, TX. [paper] [bibtex]
  2. Prakhar Ojha, Siddhartha R Thota, Vani M, Mohit P Tahilliani. Learning Scheduler Parameters for Adaptive Preemption. International Conference on Advanced Information Technologies and Applications (ICAITA) 2015, Dubai. [paper] [bibtex]

Under Review

  1. Prakhar Ojha, and Partha Talukdar. Estimating Accuracy of Automatically Constructed Knowledge Graphs. [arXiv pre-print v2] [v1] -- Accepted for Oral Presentation [rate 9%] at CODS-2017

Awards / Achievements

  • Awarded Best Poster Award at IBM Collaborative Academia Research Exchange Conference I-CARE 2016, IBM Research.
  • Awarded Microsoft Research Travel Grant to attend HCOMP-16, Thanks Micrsoft!.
  • Awarded Best Poster Award at Joint Research Students Symposium of Electrical Division EECS-2016, IISc.
  • Secured 449th All India Rank in Graduate Aptitude Test in Engineering, Computer Science (GATE-CS), 2014.



Room 303, CDS, SERC Building
Indian Institute of Science
IISc, Bangalore- 560012
Karnataka, India